Stouffer Foods Corporation

Text (OCR of the scan at left; may contain errors)

IN THE MATTER OF STOUFFER FOODS CORPORATION FINAL ORDER, OPINION, ETe. , IN REGARD TO ALLEGED VIOLATION OF SECS. 5 AND 12 OF THE FEDERAL TRADE COMMISSION ACT Docket 9250. Complaint, Oct. 1991- Final Order, Sept. 26, 1994 This final order prohibits Stouffer Foods Corporation, the manufacturer and advertiser for Lean Cuisine frozen entrees, from misrepresenting, in any manner, the existence or amount of sodium or any other nutrient or ingredient in any of its frozen-food products.

Appearances For the Commission: Theodore H. Hoppock and Nancy Warder.

For the respondent: Hugh Latimer, Wiley, Rein Fielding, Washington, D.

COMPLAINT The Federal Trade Commission, having reason to believe that Stouffer Foods Corporation, Inc. ("Stouffer" or "respondent ), a corporation, has violated provisions of the Federal Trade Commission Act, and it appearing to the Commission that a proceeding by it would be in the public interest, alleges:

PARAGRAPH 1. Stouffer is a Pennsylvania corporation with its offices and principal place of business at 5750 Harper Road, Solon, Ohio.

PAR. 2. Stouffer has advertised, offered for sale, sold, and distributed Stouffer s Lean Cuisine, a "food" within the meaning of Sections 12 and 15 of the Federal Trade Commission Act. PAR. 3. The acts or practices of Stouffer alleged in this complaint have been in or affecting commerce as "commerce" is defined in the Federal Trade Commission Act.

PAR. 4. Respondent has disseminated or caused to be disseminated advertisements for Stouffer s Lean Cuisine, including but not STOUFFER FOODS CORPORATION 747 746 Complaint necessarily limited to, the advertisement attached hereto as Exhibit A. The headline of Exhibit A contains the following statement: OF ALL THE THINGS WE MAKE, WE MAKE SENSE (Emphasis added.

The text of Exhibit A contains the following statements: Of all the things we at Stouffer s pack into our 34 Lean Cuisine entrees - the freshest ingredients, the ripest vegetables and the perfect blend of herbs and spices there are some things we skimp on: Calories. Fat. Sodium. With less than 300 calories, controlled fat and always less than I gram of sodium* per entree, we make good sense taste great.

In a footnote next to a second asterisk Exhibit A states in fine print as follows:

* All Lean Cuisine entrees have been reformulated to contain less than I gram (1000 mg.) of sodium.

PAR. S. Through the use of the statements contained in the advertisements referred to in paragraph four, including but not necessarly limited to the advertisement attached as Exhibit A, respondent has represented, directly or by implication, that Stouffer s Lean Cuisince entrees are low in sodium.

PAR. 6. In truth and in fact, in many cases, Stouffer s Lean Cuisine entrees are not low in sodium. Therefore, the representation set forth in paragraph five was and is false and misleading. PAR. 7. In its advertising for Stouffer s Lean Cuisine entrees respondent has represented, directly or by implication, that the entrees contain less than I gram of sodium. This advertising has failed to disclose adequately that I gram is equivalent to 1000 milligrams which is the commonly used unit of measurement for sodium. This fact would be material to consumers in their purchase or use decisions regarding the product. In light of the representation made the failure to disclose adequately this fact is likely to lead reasonable consumers to underestimate the level of sodium in the entrees and is a deceptive practice.

PAR. 8. The acts and practices alleged in this complaint constitute unfair or deceptive acts or practices and the making of false advertisements in or affecting commerce in violation of Sections Sea) and 12 of the Federal Trade Commission Act. :,, :; ,,(. : ./::. ,/ , j/. . . ._ ./, ./ . .. q:.. Cumplaint 118 FTC. EXHIBIT A OF. THE THIN S WE MAE.

WE MAKE SENSE! Of a/llhe Ihings we at Szouffer ' pack Into our 34 Lean Cuisine entrees-i.he freshest mgreaiems. the rides! vegetables and the oerfect blend of herbs and splces- Ihere are some things we skimp on: Calories Fat. Sodium With ,less than 300 cclorres, controlled fat and always less C'- C';"than gram of SOdlum' per entree, we make 900d \:-0 ii-sense taste great % 'V , i/ "v- "cea"Cu's:nee esnay (Oe I()W d'C. :han ,g'"''i1000rrg.:ols.')'u''e!c'mula'e- / , r" LEA ON LEA CUISIN STOUFFR FOODS CORPORATION 749 746 Initial Decision INITIAL DECISION BY JAMES P. TIMONY , ADMINISTRATIVE LAW JUDGE AUGUST 6, 1993 INTRODUCTION On October 28, 1991, the Federal Trade Commission issued an administrative complaint charging Stouffer Food Corporation with violations of Section 5 of the Federal Trade Commission Act in connection with claims made by Stouffer in the advertising and sale of its Lean Cuisine brand of frozen entrees.

After pleading and discovery, the case came on for evidentiary hearings commencing on February 8, 1993, and closing on March 8 1993. The transcript of the hearings consists of 1662 pages. About 580 exhibits, some of which were deposition transcripts, were admitted into evidence. Proposed findings were completed by June 21 1993, and indexes to the proposed findings were filed on July 14 1993.

SUMMARY OF COMPLAINT ALLEGATIONS The complaint alleged (1) that respondent s ads falsely represented that Lean Cuisine entrees are low in sodium through "statements contained in the advertisements, " including that they "skimp on: Calories. Fat. Sodium. With less than 300 calories, controlled fat and always less than 1 gram of sodium per entree, we make good sense taste great." The complaint quoted a footnote "in fine print from the ads: "All Lean Cuisine entrees have been formulated to contain less than I gram (1000 mg.) of sodium." (Paragraphs 4 and 5 of complaint.) The complaint also alleged (2) that the ads failed to disclose adequately the material fact that "I gram is equivalent to lOOO milligrams, which is the commonly used unit of measurement for sodium. " (Paragraph 7 of complaint.) FINDINGS OF FACT Respondent and Jurisdiction l. Stouffer Foods Corporation, Inc., (Stouffer) is a corporation organized, existing and doing business under and by virtue of the Initial Decision 118 F.T. laws of the State of Pennsylvania, with its offices and principal place of business located at 5750 Harper Road, Solon, Ohio. (Answer paragraph 1.) 2. Stouffer manufactures and sells frozen entrees consisting of two product lines: the Stouffer "Red Box" line and the Lean Cuisine line. (Annett, Tr. 875, 931.) 3. For the purposes of Section 12 of the Federal Trade Commssion Act, 15 U. c. 52, Lean Cuisine is a "food," as defined in Section 15 of the Act, 15 U. c. 55. (Camp!. paragraph 2; Answer paragraph 2.

4. During all times relevant, including the years 1990- , Stouffer has advertised, offered for sale, sold, and distributed Stouffer Lean Cuisine. (Answer paragraph 2.

5. At all times relevant to the complaint, the acts and practices of respondent alleged in the complaint have been in or affecting commerce. (Answer paragraph 3.

6. Stouffer is a subsidiary of Nestle U. A. which is owned by Nestle S.A. of Switzerland. (Annett, Tr. 925. Lean Cuisine and Frozen Entrees 7. Lean Cuisine is a line of frozen entrees. (Block, Tr. 775. 8. As an entree, Lean Cuisine is packaged in a tray as a single serving item. (Annett, Tr. 876.

9. During 1990- , the Lean Cuisine jine averaged 850 milligrams of sodium per entree. (CX-523- There were Lean Cuisine entrees that contained more than 1000 milligrams of sodium. (Annett, Tr. 909.

10. During 1990- 1991 , annual sales for the Lean Cuisine line were over two hundred milion dollars. (CX-523-Z- 1, Z- 11. Stouffer also manufactures and sells the "Red Box " line. (Annett, Tr. 875, 931; CX 84.

12. Beginning in October, 1989, Stouffer also manufactured and sold another line of frozen entrees, the Right Course line. (CX-382 at 21 (Audette Dep.J; Annett, Tr. 880. These products were promoted on their lower levels of fat, cholesterol, and sodium compared to the Stouffer Red Box line and the Lean Cuisine line. (Annett, Tr. 880, 890, 931; CX-96; CX-88.) The average sodium content for Right Course was under 600 milligrams. (Annett, Tr. 880. STOUFFER FOODS CORPORA non 751 746 Initia! Decision In the fall of 1990, the Right Course line was dropped. (Annett, Tr. 880-81.) 13. The Lean Cuisine line was introduced in 1981. (Block, Tr. 775. ) The brand featured calorie-control (under 300 calories per entree) and taste. (Jd. 14. In the mid- 1980' , new "healthy" frozen food products entered the market, including Weight Watcher, Budget Gourmet, and later, Conagra s Healthy Choice. (Annett, Tr. 874, 878. 15. Lean Cuisine began losing market share. (Jd. at 864; CX-84). In 1989, Lean Cuisine had 33% of the calorie-controlled entree market; that figure dropped to 25% in 1990. (CX-84. 16. During this time, consumers became concerned about nutrition, including the fat, cholesterol, and sodium in food. (Annett, Tr. at 864, 902, 914; Block, Tr 777; CX-84.

17. Consumers were confused about the Lean Cuisine line, particularly the sodium content. (Block Tr. 785.) Many consumers viewed Lean Cuisine s sodium content as high. (Annett, Tr. 917- 18; Block, Tr. 809; CX-58-G; CX- 65; CX- 139-62. 18. Responding to consumer s new nutritional awareness, Stouffer reformulated Lean Cuisine with new recipes and seasonings diminished the importance of low calories and reduced the fat and sodium. (Block, Tr. 781.) In order to counteract the perception that Lean Cuisine was high in sodium, and because it was becoming a health issue in the media, Stouffer asked Irene Block of Tatham/ RSCG (Tatham), Stouffer s advertising agency, to develop ads stating the facts on the sodium content of the product. (Block Tr. 785- 86.

19. In March of 1987, Richard B. Annett, the group marketing manager for Lean Cuisine, sent a letter to the National Advertising Division (NAD) of the Council of Better Business Bureaus concerning an ad disseminated by a competitor, Budget Gourmet, in the Miami, Florida area. (CX-24; Annett, Tr. 894-95. ) The ad claimed that the Budget Gourmet Slim Selects were: At Around $1.89, Under 300 Calories. And Under I Gram of Sodium, One of Man Lighter Creations.

(CX-24- 20. The letter to the NAD was about Budget Gourmet s sodium claim (CX-24):

, .

Initial Decision 118 FTC. Print advertising for Budget Gourmet's " Slim Select" entrees has come to our attention. which, as you wiB note, has prominently displayed the representation that the Slim Select entrees contain "Under 1 Gram of Sodium." We draw this matter to your attention as we view this statement as blatantly misleading to the consuming public and one which contravenes the industry-wide practice of utilizing the descriptor of sodium content in terms of milligrams and not grams. In essence the producers of Budget Gounnet Slim Select entrees have intentionally misrepresented the sodium content in this product by quantifying sodium content in grams. 21. The Budget Gourmet ad did not mention milligrams. (CX-24- 22. On April 8, 1987, NAD wrote to Mr. Annett that there was no basis to believe that the accurate statement ' Under I Gram of Sodium,' is misleading to consumers. " (RX- I2-A.) Mr. Annett had no consumer research showing that use of the phrase "under I gram of sodium" was misleading to consumers. (Annett, Tr. 870 926-27. 23. Sue Lally, manager of regulatory affairs for Stouffer, informed Mr. Annett that the U.S. Department of Agriculture permitted sodium disclosure statements on labels in terms of grams as well as milligrams. (Annett, Tr. 872, 927-28.

24. Stouffer then determined that it would be appropriate to use the I gram terminology in its new Lean Cuisine ads. (Annett, Tr. 872-73.

25. When the "Lean on Lean Cuisine" campaign was launched in late 1989 with "Lean on Lean Cuisine" and ''Taste Like A Million, there was no reference to sodium in the ads. (Block, Tr. 783-84. After Lean Cuisine had been reformulated, sodium content was included in the two ads. (Block, Tr. 784-85. 26. Mr. Annett informed Tatham-Laird personnel working on the campaign that the use of "lower" sodium or "controlled" sodium was acceptable for the advertising but that "low" was not. (RX- Block, Tr. 788-90; Annett, Tr. 887-89.

27. In the early 1990' s Conagra s Healthy Choice became the market leader on the low end of the nutritional spectrum for frozen entrees. (Annett, Tr. 878. ) Healthy Choice products competed successfully with low sodium, low cholesterol and low fat. (Annett, Tr. 878-79; RX-58.

28. Stouffer, in 1989-90, was marketing three lines of frozen food, each to different dietary needs. Lean Cuisine occupied middle ground. (CX-88; Annett, Tr. 878-92.) Stouffer marketed its Red Box frozen products to consumers who did not control their fat, sodium STOUFFER FOODS CORPORATION 753 746 Initia!Decision or cholesterol intake. (CX-88; Annett, Tr. 878-79, 890. ) Stouffer marketed its Right Course entrees, as a healthier product line than Lean Cuisine, with less than 600 miligrams of sodium and lower levels of cholesterol and fat. (CX- 88; Annett, Tr. 880, 889-93. 29. The Chairman and CEO of Nestle Enterprises, Inc., did not permit Lean Cuisine to use "health-oriented" advertising, since he felt it might interfere wjth the marketing of the Right Course line of products, (CX-45-A; Annett, Tr. 890-93, 928-30. 30. Stouffer reduced the cholesterol, fat and sodium in the Right Course line, but in late 1990 the Right Course line of products was discontinued. (Annett, Tr. 880-81.

31. Stouffer then embarked on a second reformulation of the Lean Cuisine line. The sodium was again reduced, to a maximum of 600 miligrams per entree, and the fat and cholesterol content also was reduced. (Block, Tr. 803; RX- 32. In July 1991 , Stouffer and Tatham-Laird ran a singing radio commercial known as "Anniversaryrrurkey Rev. " (CX-7; Block, Tr. 803.

The Ads 33. From January 1990 through August 1991 , Stouffer ads featured Lean Cuisine entrees. (CX-523- Q; CX-527; CX-528- 116.) This campaign cost three million dollars (CX-523-S; CX-527- , CX-528-G), and reached millions of consumers nationwide. (CX- 79.

34. The Lean on Lean Cuisine ad is a two-page magazine ad. (CX- 1.) The ad, at 64% of its size, is attached as Appendix A. 35. The Lean on Lean Cuisine ad ran in magazines from January through February, 1990. (CX-523-M-Q.) The magazines were Cosmopolitan, Redbook, Bon Appetit, Shape, New Woman Glamour, Working Mother, and Working Woman, all directed primarily to women. (Zinkhan, Tr. 486.

36. The 300 Like a Million ad (CX-2) is attached as Appendix B. 37. The 300 Like a Million ad ran in magazines from June, 1990 through January, J991. (CX-523- ) These magazines included Moxie, Eating Well, Glamour, Business Woman, Family Circle Newsweek Woman, Working Woman, Ladies' Home Journal and New Woman, directed primarily to women. (Zinkhan, Tr. 486. 38. The Make Sense ad (CX-4) is attached as Appendix C. Initial Decision 118 FTC. 39. The Make Sense ad ran in Good Housekeeping, Glamour Family Circle, Cosmopolitan, People, Shape, and New Woman, directed primarily to women. (CX-523-M-Q: Zinkhan, Tr. 486; Annett Tr. 919-20.) This ad ran from January through March, 1991. (CX- 523-M-Q.

40. A version of the Make Sense ad (CX-5) ran in Military Lifestyle, People, and Health (CX-523-N), with different text: 95% fat free. Never more than a gram of sodium. * Always less than 300 calories. Lean Cuisine makes great food and good sense. . (CX- ) This ad ran from February through April, 1991. (CX-523- 41. The Ole, O' lean ad is a two page ad promoting both Stouffer s "Red Box" and Lean Cuisine New Mexican entrees. (CX- The left-hand side of the ad prescnts claims for the "Red Box " line. The right- hand side promotes Lean Cuisine. (ld. The ad, at 64% size is attached as Appendix D.

42. The Ole, O' lean ad ran in People, Cosmopolitan, Working Mother, Redbook, and New Woman, directed primarily to women and also in Newsweek. (CX-527; Zinkhan, Tr. 486; Annett, Tr. 919- 20.) This ad ran from April through May, 1990. (CX-527. 43. The radio advertisement, Anniversary Turkey, was sixty seconds long. (CX- ) This ad stated:

new tenth anniversary entrees fromnyou guessed it--Stouffer s Lean Cuisine. These babies are healthier than ever. Lower in sodium, fat and cholesterol. Read those boxes, people, these numbers are low. The ad concluded with singers singing "Stouffer s Lean Cuisine. . . Taste you can love for life. (ld. 44. The Anniversary Turkey ad went over 230 radio stations from June through August, 1991. (CX-528-G to Z- 1l6. Facial Analysis of Ads 45. One message of the challenged print ads is healthy eating: Lean Cuisine has large quantities of healthy ingredients, and small amounts of undesirable nutrients. (CX- 46. The Make Sense ads ' headlines state " Of all the things we make, we make SENSE'" (CX- , CX- ) The ad describes all the STOUFFER FOODS CORPORA non 755 746 Initial Decision good ingredients in Lean Cuisine entrees in contrast to the undesirable nutrients that are present only in minimal amounts (CX-4): Of all the things we at Stouffer s pack into our 34 Lean Cuisine entrees- the freshest ingredients, the ripest vegetables and the perfect blend of herbs and spices-there are some things we skimp on: Calories. Fat. Sodium. With !ess than 300 calories, controlled fat and always less than 1 gram of sodium* per entree, we make good sense taste great.

47. CX-4 states that Stouffer "skimp(sJ" on sodium, a description virtually synonymous with a low amount of sodium. 48. CX-2 and CX-3 state in a footnote that "All Lean Cuisine entrees are currently being reformulated to contain less than I gram (1000 mg.) of sodium.

49. CX-4 and CX-5 state in a footnote that "All Lean Cuisine entrees have been reformulated to contain less than I gram (1000 mg. of sodium.

50. The radio spot, Anniversary Turkey, (CX-7) describes Lean Cuisine as follows:

These babies are healthier than ever. Lower in sodium, fat and cholesterol. Read those boxes, people, these numbers are low. 51. The first low sodium statement in the radio spot claims that the entrees are "healthier than ever" because, among other things they are now "(lJower in sodium." The ad then refers to the nutritional information on the packages and states, in absolute terms, that these numbers are low " for the undesirable nutrients including sodium. (Block, Tr. 823-24.

ZINKHA"i COpy TEST 52. U.S. Research Company ("USR") did a copy test of three of the print ads to determine whether they conveyed the low sodium claim. (CX-374.) USR is experienced in such copy tests. (Kloc. Tr. 304- 313- 14. ) The questionnaire USR used was designed by Dr. Zinkhan, a professor of marketing at the University of Houston. (CX-373: Zinkhan, Tr. 475: Kloc. Tr. 312.

53. Dr. Zinkhan ' s questionnaire used open-ended and close-ended questions. (CX-374- 29, Z-30. ) An open-ended question provides copy test participants with little context in order to obtain unprompt- Initia! Dccision 118 F.T. ed answers phrased in their own words. (Zinkhan, Tr. 478; Kloc, Tr. 306.) A structured, close-ended question asks about a specific issue and provides the answers. Consumers select one of the answers. (Zinkhan, Tr. 478; Kloc, Tr. 307; CX-522.

54. Dr. Zinkhan s copy test asked open-ended questions followed by close-ended questions. (Zinkhan, Tr. 499-508. ) It used a control question, regarding the sugar content of Lean Cuisine, to find any bias from the use of close-ended questions. (ld. at 513- I 4. 55. The three print ads tested were Lean on Lean Cuisine, 300 Like a Million and We Make Sense. (Kloc, Tr. 331-32; Zinkhan Tr. 522-24; CX- , CX- ) One hundred participants viewed these three ads at four shopping malls. (Kloc, Tr. 339-40; CX-374- Zinkhan, Tr. 539.

56. From 43 to 60% of participants answering open-ended questions stated that the ads claimed that Lean Cuisine frozen entrees are low in sodium and, after subtraction of the control question responses, from 78 to 86% gave that response to dose-ended questions. (Zinkhan, Tr. 523-26; CX-374- 11, Z-20-21; CX-526. 57. The copy test was conducted in four shopping malls located in Poughkecpsie, NY; Orlando, FL; Houston, TX; and Mission Viejo CA. (CX-374B; Kloc, Tr. 320.) The interviewing was done by USR. (Kloc, Tr. 308-09.) Dr. Zinkhan approved the mall sites. (Zinkhan Tr. 539.

58. The copy test consisted of a screener and the main questionnaire. (CX-374- 25 to Z-52.) USR employees screened consumers in the shopping malls. (KJoc, Tr. 323.

59. Qualified consumers were asked to view some ads. (CX-374- 28; Zinkhan, Tr. 497-98.) These participants read one of the three ads and were questioned by trained interviewers. (Kloc, Tr. 328-33; Zinkhan, Tr. 498-501; CX-374- 29.

60. The interviews were supervised by Mr. KJoc ofUSR. (Kloc Tr. 320.

61. Dr. Zinkhan observed the interviewer training and interviews at the Houston mall facility. (Zinkhan, Tr. 522, 535-36.) The training and interviews were conducted professionally. (ld. at 535-36. 62. A pretest of the main questionnaire was conducted prior to the copy test. (Kloc, Tr. 312.

63. As a result of the pretest, the wording of Question 3 of the main questionnaire was changed to eliminate the misinterpretation by STOUFFER FOODS CORPORA non 757 746 Initial Decision participants. (Kloc, Tr. 316- 18. ) Dr. Zinkhan gave his approval of this change. (Zinkhan, Tr. 534-35; Kloc, Tr. 318. 64. USR interviewed 300 participants, 100 for each of the three ads. (Kloc, Tr. 339; CX-374- 65. USR creates code categories into which responses are placed. (Kloc, Tr. 340-41.) Based on their review of one-third of the questionnaires, USR created a preliminary set of coding categories. (ld. at 341.) 66. Dr. Zinkhan suggested changes including a separate coding category for "low sodium " responses. (ld. Dr. Zinkhan s changes were used by the coders to categorize the responses to each of the three open- ended questions. (ld. at 538; Kloc, Tr. 344. 67. Two experienced coders, coded each of the 300 questionnaires. (KJoc, Tr. 344-45. ) The coders did not koow that the FTC was the client or that the issue of interest was whether the ad conveyed a low sodium claim. (ld. at 346. Universe 68. The universe of Dr. Zinkhan s copy test was comprised of the consumers Stouffer intended to persuade to purchase the product by disseminating the challenged ads. (Zinkhan, Tr. 475, 479, 481; Popper, Tr. 1509; Annett, Tr. 919.

69. The universe consisted of women who were the principal food shoppers for their household, between thc ages of 25 and 54, who had purchased a frozen entree in the last three months and who were not following a medically supervised diet. (CX-374- 27 to Z-29; Zinkhan, Tr. 481-97.) Participants who wore glasses to read necded to have those glasses to qualify. (CX- 374- 26; Zinkhan, Tr. 488. 70. In determining the universe, Dr. Zinkhan relied on Stouffer 7 to Z8), Stoufferdescription of its target audience (CX-523consumer surveys (CX-65- 3 to Z-25; CX-524) and his own judgment. (Zinkhan, Tr. 479-97.) !-e reviewed consumer research (CX- 69- W), consumer correspondence with Stouffcr (CX- 140; CX- 181; CX- 182; CX-221; CX-276) and an analysis of the magazines in which the ads appeared. (Zinkhan, Tr. 485-86, 490-93, 495-97. 71. Stouffer described the target audience for Lean Cuisine ads as primarily female although not exclusively, without specifying the 7 to Z- ) Dr.perccntage of men. (Zinkhan, Tr. 484: CX-523- Zinkhan did not include males in his sample. (Zinkhan, Tr. 484. Initial Decision 118 FTC. During 1990- , 15.5 to 17% of regular Lean Cuisine purchasers were men. (RX-37-B; Ross, Tr. 1101-03.) Stouffer also described the age of its target audience as "25- , with an opportunity in the under 25 segment." (CX-523- ) Of those who regularly bought Lean Cuisine in 1990- 9% were under 25; 25% were over 54. (RX-37- 72. Most of the magazines in which the ads appeared were women s magazines. (Zinkhan, Tr. 486. ) People, the magazine with the largest circulation, is read "primarily" by women. (Annett, Tr. 920.

Funneling Questions 73. Funneling of questions in a copy test refers to procecding from general questions to more narrow questions on specific issues. (Zinkhan, Tr. 476; Popper, Tr. 1505; Ross, Tr. 1251.) Funneling reveals the participants' unaided response to the ads. (Zinkhan, Tr. 476; Kloc, Tr. 307; Popper, Tr. 1505.

74. Funneling is the best way to ask questions on a copy test. (Zinkhan, Tr. 476; Popper, Tr. 1506; Ross, Tr. 1251-53. 75. Dr. Zinkhan s copy test used funneling. (Zinkhan, Tr. 499. It began with an open-ended question designed to get participants to state:

What point or points does the Lean Cuisine ad make about the product? What reason or reasons does the ad mention or suggest for you to buy Lean Cuisine? Is there anything else you can recall about the ad? (CX-374- 29 to Z-30.

76. The remaining questions in Dr. Zinkhan s copy test were close-ended questions. (Dr. Zinkhan, Tr. 500-01.) The test (CX- 374- 30) asks: "Does the ad say or suggest anything about the amount of calories (or sugar) lor sodiuml in Lean Cuisine, entrees If "yes," it asks: "Does the ad say or suggest that Lean Cuisine entrees are...

1. High in calories (or sligar) lor sodiuml 2. Low in calories (or sugars (or sodiumJ 3. Neither high nor low in calories (or sugars Lor sodiumJ. STOUFFER FOODS CORPORATION 759 746 Initial Dccision Open-ended Questions 77. In designing a copy test, the collection of data must occur as soon as possible after exposure to the ad. (Ross, Tr. 1233.) The first question in Dr. Zinkhan s copy test obtained data within seconds of when respondents read the ad. (ld. 78. Question I, the first open-ended question in Dr. Zinkhan copy test does not prompt participants for any specific response. (Zinkhan, Tr. 502; Kloc, Tr. 336; CX-536- 24. 79. Question I permits participants to give one answer, multiple answers, or no answer at all. (ld. at 501-02; Kloc, Ir. 336. ) It permits responses to be based upon the text or pictures in the ad and the visual depictions in the ad. (Zinkhan, Tr. 503-04; Kloc, Ir. 337. There is a reasonable likelihood that participants would answer Question I truthfully. (Zinkhan, Tr. 503: Kloc, Tr. 336-37. 80. Question 1 is an unbiased open-ended question. (Zinkhan, Tr. 501; Kloc, Tr. 335.

81. Questions 2 and 3 in Dr. Zinkhan s copy test arc also unbiased open-ended questions. (Zinkhan, Tr. 504-05; Kloc, Tr. 337- 38.) They do not prompt participants for any specific response nor give any context to answer the questions except the ad. (Zinkhan, Ir. 505-06.) They permit one answer, multiple answers, or no answer. (Zinkhan, Ir. 504-05; Kloc, Tr. 337-38.) There is reasonable likelihood that participants would answer these questions truthfully. (Zinkhan, Tr. 505; Kloc, Tr. 336-38.

82. A "control" in a copy test seeks bias in the question or in the participant. (CX-536- 33. ) A control "group" is a group of participants who see a different stimulus than the challenged ad. (ld. 83. Dr. Zinkhan did not use a control group for the open-ended questions in his copy test. (Zinkhan, Ir. 506-07. ) Open-ended questions do not prompt participants toward a particular attribute in the ad (F. 78-81), and a control group is not required to make the results reliable evidence. (Zinkhan, Tr. 507; Kloc, Tr. 368-70. 84. Both of Stouffer s expert witnesses in marketing research have in litigation based expert opinions on the results of open-ended questions for which there was no control group. (Popper, Ir. 1489- 91: Ross, Tr. 1297, 1303.

85. Dr. Popper designed for the Commission staff a copy test in which he did not LIse a control group for the open-ended questions. (ld. at 1491-92) 760 FEDERAL TRADE COMMISSIOI' DECISIOI'S Initial Decision 118 FTC. 86. Dr. Ross has given expert testimony based on the results of open-ended questions for which no control group existed. (Ross, Tr. 1288.

87. There is little evidence that consumers had a pre-existing belief that Lean Cuisine was low in sodium. Irene Block, of respondent s advertising agency, testified that the Lean Cuisine advertising campaign was directed at correcting consumers ' misconceptions about the amount of sodium in the product. She testified that many consumers thought Lean Cuisine had more sodium than it actually had, and that perception was exacerbated by the issue being played up in the media; she also testified that some consumers thought Lean Cuisine had less sodium than it actually had. (Block, Tr. 786-87. 88. Consumer research, done to assist Stouffer s advertising agency in the development of the challenged ads and not for litigation determined that consumers general perception was that the (sodium) levellofLean Cuisine entrees) was high." (CX-58-G; Block, Tr. 809- 10 ) 89. Most consumers believed that the sodium content of the entire frozen food category was high. (ld. At the time the challcnged ads were developed most consumer s pre-existing belief about the sodium content of Lean Cuisine and similar products was that sodium was high. (ld.

90. Sodium information was included in the challenged ads to inform consumers that Lean Cuisine s sodium content was lower than consumers believed it to be. (Block, Tr. 820-21.) The challenged ads were the first ads to mention the sodium content of Lean Cuisine. (ld. at 784- 787.

91. When consumers read ads, they use their beliefs in their interpretations of the ad. (Zinkhan, Tr. 725- 26; Shimp, Tr. 1563; Ross Tr. 1258: Popper, Tr. 1447.) They do not read ads in a vacuum, disregarding their experience and knowledge. (Shimp, Tr. 1563; Zinkhan, Tr. 726.

92. If an ad takes advantage of the reader s prior beliefs, the reader s perception of the ad may be attributed to the ad. (Ross, Tr. 1325-26: Popper, Tr. 1502-03.

STOUFFER FOODS CORPORATION 761 746 Initial Decision Close-ended Questions 93. The close-ended questions mentioned specific attributes. (Zinkhan, Tr. 500-01.) The purpose of such close-ended questions is to probe participants ' recollection of the ad. (ld. at 512. 94. Prior to answering the close-ended questions, participants were instructed to answer them "(b Jased on reading this ad. . . . (CX-374- 30; Kloc, Tr. 333-34. ) The close-ended questions sought responses based on what the ad suggested. (CX-374- 30; Zinkhan Tr. 507.

95. Close-ended questions asked if the ad suggested anything about the amount of sodium, calories, or sugar in Lean Cuisine entrees. (CX-374- 30.

96. Participants were asked whether the amount of the attribute in Lean Cuisine was "high low " or "neither high nor low " or don t know/don t remember. " (Kloc, Tr. 331-32; CX-374- 30. Response Categories 97. If participants thought the ad asked whether the attribute was reduced or lower but not "low" they would select "neither high nor low. " (Kloc, Tr. 417, 444; Popper, Tr. 1487- 88. ) If participants believed that none of the three responses were correct, they could respond "don t know. " (Kloc, Tr. 444.

Rotation of Close-ended Questions 98. The order in which copy test questions are asked can affect the results. (Zinkhan, Tr. 551-52; Ross, Tr. 1172. ) Rotating the order of close-ended questions controls order bias. (Z;nkhan, Tr. 552; Kloc, Tr. 323; Ross, Tr. 1173.

99. The close-ended questions in Dr. Zinkhan s copy test were rotated. (Kloc, Tr. 322- , 333.) Order bias was controlled in Dr. Zinkhan s copy test. (Kloc, Tr. 333; Ross, Tr. 1173, 1295-96; Zinkhan, Tr. 554.

Sugar Control 100. When a close-ended question calls for a yes or no answer some participants may answer by "yea saying," the tendency to give Initial Decision 118 FTC. the answer they think the interviewer is seeking. (Zinkhan, Tr. 513 642 744; Popper, Tr. 1411; RX-30-C.) Some participants may give an inattentive response. (Zinkhan Tr. 513 , 642, 744; RX-30-C.) 101. A close-ended question may also have a halo effect. (Zinkhan Tr. 513, 642 744. ) A participant with a favorable opinion of the product formed before taking the test may answer based on that opinion rather than what was in the ad. (ld. at 513- 14. ) Such responses to close-ended questions are based on "noise" factors. (RX-30-C.) 102. Because some close-ended questions may result from yea saying, inattention, or other noise factors, they require a control. (Zinkhan, Tr. at 641 671 742.) One control is the use of a control question. (Zinkhan, Tr. 513- 744; Ross, Tr. 968-69; CX-536to Z-36.

103. A control question asks about a product attribute reasonably associated with the advertised product, or product category, but not closely linked with explicit claims in the ad. (Zinkhan, Tr. 514- 744-45; Ross, Tr. 1198-99; Popper, Tr. 1470. 104. The control question measures the participants who answered based on yea saying, inattention, halo effect, or other noise factors. (Zinkhan, Tr. 513- 14; Ross, Tr. 969.) To eliminate the effect of such external factors, the results of the test question are reduced by the control question results. (Zinkhan, Tr. 514, 520- 526-26; CX- 536- 35 to Z-36; Ross, Tr. 969-70.

105. The control attribute must not be too closely linked with explicit claims in the ad. (Zinkhan, Tr. 514- , 744-45; Popper, Tr. 1470; Ross, Tr. 1198-99.) If the control attribute can be reasonably inferred from the ad, responses to the control question may be based on that inference. (Popper, Tr. 1472; Zinkhan, Tr. 744-45. 106. Dr. Zinkhan selected sugar as the attribute for the control question in his copy test. (Zinkhan, Tr. 514.) Participants were asked whether the ad suggested anything about the amount of sugar in Lean Cuisine. (CX-374- 30. ) The percentage who answered yes was subtracted from the percentage who said that the sodium content was low. (Zinkhan, Tr. 514, 520- , 524-26. ) This eliminated external factors from the final results. (ld. CX-526. 107. Dr. Zinkhan based the choice of sugar as the control because Lean Cuisine contained sugar and it is reasonably associated with Lean Cuisine, yet is not in the ads. (Zinkhan, Tr. 515- 19. 108. The choice of sugar as a control is supported by Stouffer data. (ld. at 517.

STOUFFER FOODS CORPORA non 763 746 Initial Decision 109. One study asked whether consumers controlled nutrients or ingredients in the food they buy. (CX-68-E; Zinkhan, Tr. 517. ) Fat calories, cholesterol, sodium, and sugar were "the five most frequent targets for dietary limitation or control." (CX-69- 17, 110. Some purchasers of Lean Cuisine entrees wrote letters to Stouffer raising concerns about the sugar content of the product. (CX-273; CX-301; CX-356-362; Zinkhan, Tr. 519. I I 1. Calories or fat could not be used in a control question because they were used in the ads. (Zinkhan, Tr. 514- 15; Popper, Tr. 1484; Ross, Tr. 969.) Because consumers link cholesterol in a product to its fat content, it should not be used in a control question. Implied cholesterol claims were created from the mention of fat content in one of the ads. (Zinkhan, Tr. 514- 15, 654; Popper, Tr. 1484.) Red meat should not be used in a control question because it is not contained in many Lean Cuisine products. (Zinkhan, Tr. 667. I 12. Of the attributes considered and avoided by purchasers of frozen entrees, the most frequently mentioned attributc suitable for use in a control question was sugar. (Zinkhan, Tr. 667. 113. Sugar is in all but one of Stouffer s Lean Cuisine entrees. (CX-409-506.) It is listed as an ingredient on the Lean Cuisine package. (Id. Zinkhan, Tr. 518. 114. Controlling sugar is important to Lean Cuisine consumers and it was a proper attribute for the control question. (CX-69- 18; Zinkhan, Tr. 517.

Results of Zinkhan Copy Test 115. Close-ended questions will generate higher response levels for an implied claim than open-ended ones. (Zinkhan, Tr. 533-34. Stouffer s expert witness testified that often a rescarcher must rely on open-ended responses of 8 to 10% as being meaningful. (Ross, Tr. 1299.) Open-ended responses of 16% constitute a substantial number of participants taking a claim from a tested ad. (Id. 116. The following percentage of participants in Dr. Zinkhan copy test responded to the open-ended questions that the ad communicated that Lean Cuisine entrees are low in sodium: Initial Dccision ! 18 FTC. Make Sense (CX- 60% 300 Like a Million (CX- 45% Lean on Lean Cuisine (CX- 43% (CX-374- ll; Zinkhan, Tr. 523.

117. The following percentage of participants in Dr. Zinkhan copy test gave the low sodium response to the close-ended questions: Make Sense (CX-4) 88% 300 Like a Million (CX- 90% Lean on Lean Cuisine (CX- 83% (CX-374- 21; CX-526; Zinkhan, Tr. 524).

118. The following percentage of participants in Dr. Zinkhan copy test answered the control question by stating that the ad said something about the sugar content of Lean Cuisine: Make Sense (CX-4) 300 Like a Million (CX- Lean on Lean Cuisine (CX- (CX-374- 21; CX-526; Zinkhan, Tr. 525).

119. The following percentage of participants stated that the ad communicated that Lean Cuisine was low in sodium in response to the sodium close-ended question after deducting the percentage who answered yes to the control question:

Make Sense (CX- 83% 300 Like a Million (CX- 86% Lean on Lean Cuisine (CX- 78% (CX-526; Zinkhan, Tr. 525-26).

Ross COpy TEST 120. Respondent Stouffer introduced a mall intercept copy test of the same three ads tested in Dr. Zinkhan s copy test. (RX-30. The test was designed by Dr. Ross, a professor of marketing at the Cniversity of Minnesota. (RX-31.) Due to methodological deft- , , STOUFFER FOODS CORPORA non 765 746 Initi:Jl Decision ciencies, the results of Stouffer s copy test do not rebut the findings of Dr. Zinkhan ' s copy test.

U ni verse 121. The universe for Stouffer s copy test consisted of potential purchasers of Lean Cuisine, regardless of whether they were in the target audience for the ads. (Ross, Tr. 998- 1000, 1094-96.) The universe was four-ninths women from 18 to 45 years, two-ninths women of any age over 45, two-ninths men from 18 to 45 and oneninth men of any age over 45. (Jd. at 1003. 122. The target audience for the challenged ads was limited by Stouffer to people from 25 to 54 years old with an opportunity in the under 25 segment." (Zinkhan, Tr. 541; CX-523- ) Yet a large percentage of the participants in the Ross test were older. About 25% of those who buy Lean Cuisine are 55 and older. (RX-37- ) The use of participants over the age limitations of the target audience makes the universe of the Stouffer copy test unduly broad. (Zinkhan Tr. 541-42.

123. The percentage of men in the copy test is twice as large as the percentage of male purchasers of Lean Cuisine. (RX-37-38; Zinkhan, Tr. 541-43.

124. Dr. Ross intended to include in the universe purchasers of aji frozen entrees with which Lean Cuisine competed. (Ross, Tr. 11 10- 11. ) The market in which Lean Cuisinc competed included Stouffer s own "Red Box" brand entrces. (Annett, Tr. 877-78. 125. Dr. Ross improperly excluded purchasers of Stouffer s "Red Box" from the universe in his copy test. (Ross, Tr. 1111- 12. Funneling Questions 126. The best method to determine consumer understanding of an ad is "to use a serics of increasingly focused, but starting out with open-end very unstructured questions about what consumers get as main ideas and then as other ideas from a commercial. . .. " (Ross Tr. 1249-50.) This describes the funneling approach of asking questions. (Ross, Tr. 1251; Popper, Tr. 1505; Zinkhan, Tr. 476. 127. The copy test Dr. Ross designed for Stouffer did not begin with open-ended questions. (Ross, Tr. 1232: Zinkhan, Tr. 543; RX- 30- ) Instead, it began with a close-ended question Did you get Initial Decision 118 F.TC. any understanding about the fat content of the product from the advertisement?" (RX-30- ) Each other attribute question (sodium calories, cholesterol and vitamins) in Stouffer s copy test also began with a close-ended question. (ld. at Z- 11.) These questions "run the risk of imparting ideas... or thoughts. " (Ross, Tr. 1250; Zinkhan Tr. 543-45; RX-30- 128. Ihe form of these questions prompts participants to think about the attribute rather than their uncoached reactions to the ad. (Zinkhan, Ir. 544.) It is not appropriate to start a copy test with such questions, (KIoc, Ir. 436-37; Ross, Ir. 1250, 1252-53: Zinkhan, Ir. 543-45.

Order Bias 129. Stouffcr s copy test asked five questions, each having four subparts. (RX-30- 7 to Z- ll. ) Each of the five questions asked about an attribute, fat, sodium, calories, cholesterol, or vitamins. (ld. The questions about fat were asked first, calories were third and cholesterol was fourth. (ld. In half of the questionnaires, the questions about sodium were second and the questions about vitamins were asked last. The other half reversed the ordcr of the sodium and vitamin questions. (Ross, Tr. 1179.

130. When asking close-ended questions, researchers rotate the order to minimize order bias. (Zinkhan, Tr. 552; Ross, Tr. 1295-96. I3 I. Order bias is espccially important in the first and last closeended questions. (Zinkhan, Tr. 553-54; Ross, Tr. 1038- 39, 1173. The first question sets up the survey. (Zinkhan Tr. 553; Ross Tr. 1038-39. ) The results of the last question may be affected by fatigue or boredom. (Zinkhan, Tr. 553-54.

132. The sodium question was asked last half of the time. (Ross Tr. 1179; RX-30- ) Because proper rotation of the questions would have placed this question in thc last position one-fifth of the time, this was not a proper control for order bias. (Zinkhan, Tr. 554. 133. The results of the sodium close-ended question for the Make Sense ad in the Stouffer copy test shows that when the question was asked in the second position (Question 2a), 22% answered " " but when it was asked in the last position (Question 5a) 42'70 responded no. " (CX-539-F; Ross, Tr. 1181- 82.) Since nearly twice as many participants answered " " to the sodium question when it was in the STOUFFER FOODS CORPORATION 767 746 Initial Decision last position, the low sodium results may be based on order bias rather than participants ' impressions of the ad. (Zinkhan, Tr. 553-54. Cleansed Ads 134. Stouffer s copy test used two controls. First participants were shown "cleansed" versions of the three challenged ads. (RX- 30- , 0, Q. ) A cleansed ad eliminates from the challenged ad all elements believed to convey the challenged claim. (Ross, Tr. 1009; Popper, Tr. 1430-33.) The theory is that any low sodium responses then obtained from the cleansed ad are the result of the participant prior beliefs that Lean Cuisine or products in its product category are low in sodium, rather than the result of any message conveyed by the ad. (Ross, Tr. 1016- 17.) The clcansed ad low sodium answers were subtracted from the low sodium results obtained from viewers of the challenged ad to control for these purported prior beliefs. (Jd. 135. Stouffer s copy test used cholesterol as a control question just as Dr. Zinkhan used sugar. (RX- 30-C.) 136. Dr. Ross testified that a "cleansed" ad is the only appropriate control ad. (Ross, Tr. 1008-09, 1014- 15, 1089-90; RX-30- C.) A cleansed ad can only function as a control ad if it does not convey the claim the tested ad is alleged to convey -- the low sodium claim in this case. (Zinkhan, Tr. 561-62; Ross, Tr. 961 1008-09, 1274; Popper, Tr. 1454.

137. In cleansing the ads, Dr. Ross changed the phrase "less than ) Withone gram" to "less than 1000 milligrams. " (RX-30- , 0, Q. regard to the Make Sense ad, cleansing removed the phrase "there are some things we skimp on: Calories. Fat. Sodium." (RX-30- 138. Dr. Ross assumed that the cleansed ads did not convey the low sodium claim. (Ross, Tr. 1274. ) Dr. Popper, Stouffer s other expert witness, stated that he would need empirical evidence to make that determination. (Popper, Tr. 1448.

139. Stouffer s "cleansed" ads contain elements likely to convey the low sodium claim. (Zinkhan, Tr. 563- . 569-73; Shimp, 'Ir. 1560- , 1567- 68, 1571- , 1577 , 1580-81.) Those ads fail as controls.

140. The challenged ads and the cleansed ads relate to sensible healthy eating. (Shimp, 'Ir. 1566-68; Zinkhan, Tr. 563-66, 569- 690-92.) The cleansed ads link the phrase "less than 300 calories with the phrase "less than 1000 milligrams of sodium." (Shimp, 'Ir. Initial Decision 118 FTC. 1572 1577 1580; RX-30- , 0, Q; Zinkan, Tr. 564, 571-73) These aspects of the cleansed ads contribute to conveying a low sodium claim to consumers. (Shimp, Tr. 1566- , 1571- , 1577, 1580-81; Zinkhan, Tr. 563- , 569-73.

14 I. The 1000 millgrams of sodium information is ambiguous information to consumers. (Shimp, Tr. 1567-68. Because the cleansed ads have made readers think about sensible, healthy eating, consumers relate the "less than 1000 milligrams of sodium" statement to the "less than 300 calories" statement. This results in the sodium information as part of the sensible, healthy eating. (Shimp, Tr. 1566- 68; Zinkhan Tr. 565. ) Thus, consumers interpret the cleansed ads to make the challenged low sodium claim. (Shimp, Tr. 1568. 142. Consumers understand that an entree with less than 300 calories is low in calories. (Shimp, Tr. 1600-02. Relating the phrases "less than 300 calories" and "less than 1000 milligrams of sodium " reasonable consumers therefore interpret "less than 1000 milligrams of sodium" as meaning Lean Cuisine is also low in sodium. (ld.

143. The phrase "less than" as a modifier of 1000 milligrams of sodium by itself contributcs to a low sodium claim. (Zinkhan, Tr. 564 570- 691.) 144. In CX- (300 Like a Million), the statement "less than 300 calories and most with less than I gram of sodium" is in bold print. The cleansed version of this ad changes " I gram" to " 1000 milligrams" but retains the bold print for the entire phrase. (RX-30- The accentuation of this information contributes to a low sodium claim. (Zinkhan, Tr. 572, 691; Shimp, Tr. 1578- 81.) 145. The bold print linking calories and sodium content of Lean Cuisine, and the headline, lead reasonable consumers to a low sodium claim in the ad. (Shimp, Tr. 1578-81.) 146. In creating a cleansed control ad, only the language causing the challenged claim should be removed. (Zinkhan, Tr. 566; Ross Tr. 1014; Popper, Tr. 1453. ) All other elements must be held constant. (Zinkhan, Tr. 566-67; Ross, Tr. 1014; Popper, Tr. 1453) 147. The cleansed Make Sense ad (CX-4) did not adhere to that principle. (Ross, Tr. 1286.) The cleansing of this ad did not "hold as much constant as possible. (ld. at 1285. 148. In the opinion of Stouffer s experrs, all that was required to create the cleansed version of CX- I (Lean on Lean Cuisine) was to change "I gram" to " 1000 milligrams" and to delete the footnote. STOUFFER FOODS CORPORA non 769 746 Initial Decision (Ross, Tr. 1274-75; Popper, Tr. 1469-70.) However, besides those changes, Dr. Ross deleted the first two lines, as well as some other phrases, in creating the cleansed version. (Ross, Tr. 1276; CX- RX-30- ) Dr. Ross could give no reason why these deletions were made. (Ross, Tr. 1276-77.

Cholesterol Control 149. The Stouffer copy test used cholesterol as a control question.(Ross, Tr. 1031; RX-30-C.) 150. Cholesterol is so closely related in consumers ' minds to fat that it is likely that consumers will take an implied cholesterol claim from the reference to fat in the tested ads. (Zinkhan, Tr. 557, 657). As a result, the cholesterol question in the Stouffer copy test is not valid. (ld. at 557-58, 745; Popper, Tr. 1470; Ross, Tr. 1199. 151. Consumers believe there is an association between fat and cholesterol. (Zinkhan, Tr. 559; Levy, Tr. 168; Ross, Tr. 1205-06. The 1990 Health and Diet Survey conducted for the FDA asked those who had heard of high blood cholesterol to state if certain actions would might " or "would not" help control high cholesterol. (CX-365-C.) One of the actions was "Eating less fat. (ld. Nearly 86% answered that eating less fat would help control high cholesterol. (Levy, Tr. 167-68; CX-39- 152. If consumers think a food is low in fat, they are likely to think it is low in cholesterol. (Levy, Tr. 169.) One of the tested ads made an express fat content claim for Lean Cuisine, while the others did so by implication. (Zinkhan, Tr. 559, 657. DECEPTION OF Low SODICM CLAIM Amount of Sodium 153. While the challenged ads ran, Lean Cuisine entrees averaged 850 milligrams of sodium. (F. 9; CX-409-506.) This exceeded regulatory and public health organizations' guidelines for low sodium. (21 CFR 101.3(a)(3) (1992); CX- 114; CX-520. 154. For eight years, the FDA has defined low sodium as 140 milligrams or less for "single serving foods" (a bowl of soup, a piece of pizza, a cup of macaroni and cheese). (21 CFR 101.3(a)(3) (1992).

Initial Decision 118 FTC. 155. In 1990 and 1991 , Lean Cuisine entrees contained about twice that amount of sodium. (CX-523- 156. The USDA has an informal policy of 140 milligrams per component for meal-type products such as frozen dinners and entrees. (Brewington, Tr. 265. ) Most frozen dinners and entrees have two three, or four components. (Jd. at 266.) For a three component food item, low sodium would be defined as 420 milligrams (3 times 140); for a two component food item, it would be 280 milligrams (2 times 140). (Jd.

157. By the USDA definition, Lean Cuisine entrees consist of two or three components. (Brewington, Tr. 285.) Low sodium for a two-component entree is 280 milligrams (2 times 140). (Jd. at 266. Recommended Maximum Daily Intake for Sodium 158. In 1989, the National Academy of Sciences recommended that Americans should limit their total daily intake of sodium to 2400 milligrams or less. (CX- 117-C.) 159. The Lean Cuisine line average of approximately 850 milligrams of sodium during the time in which the ads appeared represents over one-third of the recommended maximum daily intake. (FDA Food Regulations, 58 Fed. Reg. at 2227 (to be codified at 21 CFR 101.9(c)(9)); USDA Food Regulations, 58 Fed. Reg. at 645; CX- 117).

Consumer Perceptions of Low Sodium 160. Because of Stouffers' loss of sales of Lean Cuisine, Tatham conducted research in the spring of 1989. (Shimp at 1581; CX-58- ) The report stated as follows (CX-58-G):

The sodium content of Lean Cuisine products was frequently commented upon. Few respondents had a sense of what percentage of an average daily requirement of salt would be found in a Lean Cuisine entree, but the genera! perception was that the level was high.

161. Another report of four focus groups conducted in the fall of 1988 examined a proposcd line of frozen entrees similar to Lean Cuisine. (Shimp at 1583; CX- I02- ) That report stated (CX- 102- CX- 104):

STOUFFR FOODS CORPORATION 771 746 Initial Decision (Jlust laying out the levels (of cholesterol, fat, and sodiumJ adds confusion because many don t know how to evaluate them. Providing a comparison of the product levels along with the recorrended daily level... seemed to satisfy their desire for the facts and allows them to understand how the product could fit into an entire day s diet.

162. Stouffer knew in September 1988 that (CX- 102-K): Consumers are confused by the vast difference in acceptable levels of sodium vs. those of fat and cholesterol. Therefore, actual sodium levels should only be utilized when a reference to the recommended daily level is also shown. 163. Studies of food labels show that consumers have diffculty understanding sodium information stated numericaJly and would likely interpret 1 gram of sodium as being Jess than 1000 miligrams of sodium. (Levy, Tr. 155.

164. FDA label format studies show that consumers think that saturated fat levels are low because their numbers tend to be low (e. , 3 5, etc. grams of saturated fat); however, consumers tend to assess sodium levels as high because their numbers are high (e. 120 660, 910 milligrams). (Levy, Tr. 155.

165. One FDA labeling study had a food label with a nutrient claim of low sodium on the front panel and asked consumers whether the claim was true based on the nutritional information on the back panel. (Levy, Tr. 137.) Two claims involved low sodium: a cake with 115 miligrams of sodium and a frozen dessert with 20 milligrams of sodium, both true under FDA regulations. (CX-364- 166. For the frozen dessert with 20 milligrams of sodium, 75% of respondents perceived the "low sodium" claim as true; however this percentage dropped to 57% for the cake with 115 miligrams of sodium. (CX-364-A.) This supports the conclusion that consumers look at absolute numbers in assessing claims. (Levy, Tr. 139. 167. Dr. Levy of the Food and Drug Administration credibly testified that consumers perceive the actual sodium content of the Lean Cuisine line averaging 850 milligrams of sodium as high. (Levy, Tr. 149.) However, he stated that consumers viewing a less than I gram of sodium claim would view that claim as low. (ld. 156.

).) Initial Decision 118 F. Stouffer s Knowledge 168. Stouffer knew that its products were not low in sodium. (Block, Tr. 789; Annett, Tr. 888-89, 916- 17; CX-44. ) Mr. Annett Stouffer s manager in charge of the Lean Cuisine line at the time the ads ran, testified that a low sodium claim could not be used in Lean Cuisine advertising because "Lean Cuisine did not meet the FDA and USDA requirements for low sodium. " (Annett. Tr. 916- 17; CX-44- A; Block, Tr. 800.

169. Mr. Brewington of the Department of Agriculture, testified that he had been involved in the labeling approval process for Stouffer s Right Course line offrozen entrees during 1989. (Brewington Tr. 270. ) At that time, the Right Course product line averaged under 600 milligrams of sodium, less than the Lean Cuisine line average of 850 milligrams, and Stouffer was secking approval for a low sodium labeling claim for Right Course. (ld. That request was never granted, according to Mr. Brewington, because the sodium level (600 milligrams) was too high. (ld. at 272. 170. Stouffer knew that a low sodium claim was inappropriate for Lean Cuisine. (ld. Annett, Tr. 916- 17; Block. Tr. at 789. Materiality of Low Sodium Claims 171. The sodium claims chaHenged in this proceeding constitute health claims that are important to consumers. Based on medical evidence supporting a link between sodium consumption and high blood pressure, the National Academy of Sciences, the American Heart Association, and the Surgeon General of the United States recommend that people limit their daily sodium intake. (CX- 117 CX- 131 , and CX- 116.

172. Stouffer s copy test, to the extent that it is reliable, showed that 68% of the participants considered sodium to be important in making purchase decisions about frozen entrees. (Zinkhan, Tr. 584; CX-513.

173. Stouffer s consumer research in the spring of 1991 studied why people buy frozen dinners. (CX-65; CX-383 at 56-57 (DeVries Dep. ) The things considered were: brand name; cholesterol, fat calories from fat; price; vitamins and minerals; and sodium. (CX-65- R: CX-383 at 58-59 (DeVries Dep.

STOUFFER FOODS CORPORATION 773 746 Initia! Decision 174. The result reported was (CX-65-S) (emphasis in original): (The) analysis revealed that sodiuJI1 J.vel is the dominant factor. Respondents clearly favor products with the lowest level of sodium possible. The analysis found a strong negative reaction to products with 1000 milligrams of sodium (CX-65-S; CX-383 at 60- 61 (DeVries Dep. 175. Stouffer s consumer research in 1988 showed the importance of information about sodium to consumers. A report on focus groups conducted in the fall of 1988 stated as follows under the heading cholesterol, fat and sodium levels (CX- I02-I): These consumers are infonnation hungry. They are serious about their problem and therefore want to know the precise cholesterol, fat, and sodium levels. As a result of this research, Stouffer was also aware that a frozen entree containing 600 or more milligrams of sodium "could turn consumers off." (Jd. at h; CX- 382 at 48-49 (Audette Dep. 176. Stouffer began to develop a line of nutritionally-oriented entrees in the latter part of 1988. (CX-382 at 13 (Audette Dep. 177. Stouffer began a new line of frozen entrees called Right Course in the fall of 1989. (Jd. at 19.) The strategic positioning for Right Course emphasized its levels of sodium, fat, and cholesterol. (Jd. at 29.

178. The ad agency personnel assigned to the Lean Cuisine account were aware of the importance to consumers of claims about sodium. (Block, Tr. 774, 808; CX-379 at 48 (Wood Dep.); CX-381 at 47 (Blim Dep.j; CX-378 at 98-99 (Crain Dep. 179. Stouffer s ad agency documents regarding brand positioning for Lean Cuisine in June 1990, stated that science and the media have been evaluating the consequences of eating habits and contained a list of six nutritional issues, the first of which was "Sodium Awareness. (CX-77E.

180. In a presentation to Stouffer in November of 1990, the agency said that acceptable levels of sodium, fat, and cholesterol had become a "price of entry," to get consumers to try the product (CX- 378 at 67- , 98-99 (Crain Dep.l; CX-80-C), and that people want no bad stuff " that is, nutrients like sodium which are thought to be unhealthy, in the foods they eat. (CX-80-D; CX-378 at 96, 99 (Crain Dep.

Initia! Decision 118 F.T. 181. In January of 1990 Mr. Annett, the marketing manager for Lean Cuisine, sent a memorandum to Tatham instructing the agency to put health related executions "on a fast track." (CX26-A.) The memorandum also suggested using "hot buttons" or "strong 'buzz words '" about limiting sodium, fat, and cholesterol. (CX-26. 182. Health and Diet Surveys (for the National Heart, Lung, and Blood Institute, the National Cancer Institute, the Centers for Disease Control, and the USDA) evaluate consumer awareness of nutrition. (Levy, Tr. 106, 120-21.) 183. The Health and Diet Survey in the early fall of 1990, shows consumer awareness about sodium. (Levy, Tr. 122. 184. The survey asked if the participant had heard of anything that people eat or drink being related to high blood pressure. (CX- 365.

185. Of the participants, 44% answered sodium or salt (CX- 364), the most frequently given response. (Levy, Tr. 125-26. 186. Other questions show that 15% of the population were on a professionally recommended sodium reduction diet, 25. I % were on a self-prescribed sodium avoidance diet, and 40% of the adult population aged 18 years and over are on a sodium reduction diet, making it the most common diet restriction. (Levy, Tr. 13 I; CX-346- Disclosure of Milligrams 187. The print ads in this case state that Stouffer s Lean Cuisine entrees contain less than 1 gram of sodium, providing the metric equivalent in milligrams in a footnote. (CX- 6; CX-519 and CX- 525.

188. Dr. Muehling, a professor of marketing at Washington State University, tested five ads for a fictional camera. Some of the ads had fine print footnotes, others had large print footnotes. (CX-385- 89; Muehling, Tr. 27-28. ) Each of the ads contained information about attributes of the camera. (CX-385-89; Muehling, Tr. 28-29. 189. The survey was conducted on a "convenient sample of college students." (Muehling, Tr. 72.

190. The students were tested on their recall of statements made in the ad. (Muehling, Tr. 33-34.) According to Dr. Muehling, the results indicate that individuals were generally able to recall points that are made in the body of an ad much better than the points that are , STOUFFR FOODS CORPORATION 775 746 Initia! Decision made in the fine print or the footnote statements contained in the ad. (Muehling, Tr. 36, 44.

191. The footnotes Dr. Muehling tested in his study contained between 25 and 38 more words and more information than the footnotes in the Lean Cuisine ads. (Muehling, Tr. 67-68. 192. The text of the camera ad was more lengthy than the texts of the Lean Cuisine ads. After reading a lengthy ad, consumers may not pay attention to footnotes. (Muehling, Tr. 82-83. 193. To determine whether specific footnotes are comprehended conducting a test on those ads "would be a most effective way of answering that question. " (Muehling, Tr. 66-67. 194. Most consumers do not read or recall the footnotes in the Lean Cuisine ads. Responding to open-ended questions on the Zinkhan Copy Test, for the Made Sense ad (CX-4), none of the 100 participants recalled footnoted information. For the 300 Like a Million nine of 100 participants recalled footnoted information. And, for the Lean on Lean Cuisine ad, two of 100 participants recalled footnoted information. (Zinkhan, Tr. 532; CX-374- l2. 195. The preponderance of the credible evidence shows that the footnotes in the ads in this case did not adequately disclose that 1 gram equals 1000 milligrams. (ld. 196. The sodium content of food is commonly, although not uniformly, measured for consumers in milligrams. (F. 20, 23; RX- 24-E; RX-25- , J.) Although consumers are generally aware of the need to restrict sodium in their diet (F. 185-86), many are unaware of the precise recommended daily allowance for sodium (F. 160), in milligrams or grams. The failure to disclose adequately the sodium content in milligrams is, therefore, immaterial. DISCUSSION I. INTRODUCTION Respondent Stouffer, a subsidiary of the Swiss corporation Nestle SA, manufactures and markets frozen foods, primarily frozen entrees. Stouffer s frozen entree products consist of two product lines: a full calorie product Red Box " and a reduced, low calorie product line Lean Cuisine. Lean Cuisine sales were about two hundred milion dollars in 1990-91. (F. 10.

Initial Decision 118 FTC. During the late 1980' , Lean Cuisine s leadership of frozen entrees was challenged by Weight Watchers and new brands of Budget Gourmet. Despite a growing market, Lean Cuisine s business declined 24% in four years. (CX-58- During the fall of 1989, Stouffer started a new advertising campaign for Lean Cuisine. Stouffer s advertising agency, Tatham RSCG (Tatham), found that consumers worried less about calories but had an increasing interest in nutrition and the adverse health consequences of sodium and fat, and that consumers viewed Lean Cuisine and frozen entrees in general as high in sodium. (F. 17- 19, 88-89; CX-58-B, G.) Tatham created the ads at issue in this case. These ads included two, two-page print ads entitled "Lean Cuisine. (CX- , CX-5I9, CX-525) and "Ole! O' lean!" (CX-6); two, onepage print ads entitled "Who can make under 300 taste like a MIL- LION?" (CX- 3) and "Of all the things we make, we make SENSE!" (CX- 5); and a radio ad entitled "Anniversary/Turkey Rev." (CX- The complaint alleged that the ads falsely represented that Lean Cuisine entrees are low in sodium through "statements contained in advertisements." (Complaint, paragraphs 4, 5. ) The complaint also alleged that the ads failed to disclose adequately the material fact that I gram is equivalent to 1000 milligrams, which is the commonly used unit of measurement for sodium." (Complaint, paragraph 7. Respondent argues that its ad campaign stressed Lean Cuisine great taste and controlled fat, calories and sodium, and that the representations about sodium content were meant to be relative, showing a reduction in the amount of sodium but not implying low sodium, which consumers associate with bland taste. II. THE CHALLENGED ADS A. The Legal Standard The standard by which advertising is judged is whether it is likely to mislead reasonable consumers; proof of actual deception is not required. The issue is whether consumers, acting reasonably under the circumstances, would interpret the message of the advertisement to have made the alleged claims. Kraft, Inc., D. 9208, slip op. at 5- 21 (Jan. 30, 1991), aff'd 970 F. 2d 311 (7th Cir. 1992), cert. denied, Stouffer FOODS CORPORA TION 777 746 Initi'-l Decision 113 S. Ct. 1254 (1993). An ad can be deceptive even though other at n. 8.reasonable, truthful intcrpretations are just as possible. (Id. The Commission may rely on its own reasoned analysis to determine what "reasonably clear" implied claims are conveyed by examining the "overall net impression of an ad. Kraft 970 F.2d at 314 319. The analysis looks at the net impression created by the interaction of all of the different elements in the ad, rather than thc impact of each or a few elements. Thompson Medical Co. , Inc. 104 FTC 648 793 (1984), 791 F.2d 189 (D. C. Cir. 1986). ' The Commission does not have a license to go on a fishing expedition to pin liability , barely discernable claims. Id.on advertisers for barely imaginable at 319-20. But when implied claims are conspicuous, self-evident, or reasonably clear on the face of the ad, consumer surveys or other evidence beyond the ad are not required in reaching the decision. Id. at 320. If the implied claims may not be determined with confidence from the face of the ad, extrinsic evidence must be examined, including consumer surveys and expert testimony. Kraft 970 F.2d at 318. B. The Low Sodiwn Claim 1. Facial analysis of Stouffer s print ads The headline of the Make Sense ads (CX- , CX- 5) states "Of all the things we make, we make SENSE'" which evokes sensible eating. The ads describe the hcalthy ingredients in Lean Cuisine and note:

there are some things we skimp on: Calories. Fat. Sodium. With less than 300 calories, controlled fat and always less than 1 gram of sodium* per entree. we make good sense taste great.

A footnote states "All Lean Cuisine entrees have been reformulated If the footnote isto contain less than I gram (1000 mg.) of sodium." overlooked by a consumer, the ad explicitly describes the sodium content of Lean Cuisine as " 1" gram, a low number. The sodium is No first amendment concerns are raised when f8cially apparent implied claims are found without resort to extrinsic evidence. Zauderer v. Ohio, 471 C.s. 620, 652-53 (1985). A facial analysis involves the net impression conveyed by the ads and docs not involve the effect of individual words. phrases, or visual images. Thompsoll Mediml. 104 FTC at 793. Contr3.ry elements in the ads must be effective to dispel the net impression of the challenged claim- Kraft, slip op. at 10. Initial Decision 118 F. described as "less than" I gram, diminishing the quantity.' The ads state that Lean Cuisine "skimp(sJ on" sodium and other undesirable ingredients. The phrase "We make good sense taste good" reinforces the sensible eating message.

The net impression of all of the elements of the ads is that Lean Cuisine entrees are low in sodium. The ad contains nothing to give a contrary impression. Thompson Medical 104 FTC at 793. The footnote that a gram equals 1000 milligrams, assuming that consum- 3 is ambiguous unless consumers knew their recom- ers notice it mended daily allowance. The footnote in some of the ads stated that the product is being reformulated. This is consistent with the low sodium message. Thus, a facial analysis of the challenged ads shows that they convey the low sodium claim to reasonable consumers: 2. Radio ads The challenged radio ad described Lean Cuisine entrees (CX-77): These babies are healthier than ever. Lower in sodium, fat and cholesterol. Read those boxes. people, these numbers are low. Lower in sodium" is a comparative statement, but it is consistent with, and does not contradict, the flat, ahsolute statement that "these numbers are low. To prevent facial analysis and require extrinsic proof, a conflicting statement in the ad must be effective. Kraft, FTC slip opinion at 10. Here, the comparative statement does not conflict with "these numbers are low " and does not derogate from the net impression that the radio ad carries the message that Lean Cuisine entrees are low in sodium.

- The phrase "less than 300 calories and most less than I gram * of sodium" in the 300 Like a Million ad (CX- 1 and CX- 3) appears in bold print. (F. 144 3 The footnote did not adequately disclose that 1 gram equals 1000 milligrams (F. 195. 4 Thc other print ads are similar although they do not use the phrase "skimp on" as the Make Sense ad does. The above analysis applies to those ads as well. STOUFFER FOODS CORPORATION 779 746 Initial Decision C. Extrinsic Evidence 1. Zinkhan copy test Consumer surveys are the best extrinsic evidence of what words in an ad mean to consumers. Kraft, Inc., slip op. at p. 1 I n. 1 I , p. 13 n. 13. Copy tests must use a sound method, with a valid sample questions that minimize bias, and correct analysis. Thompson Medical 104 FTC at 790.

a. Universe The universe for the copy test is a valid sample from the "appropriate population.'" The target audience here is the group of people Stouffer tried to persuade to purchase its product with its advertising. (F. 70-71) Stouffer s target audience consisted of "primarily females, though not exclusively" who were "age 25 to 54 with an opportunity in the under 25 segment." (CX-523- 7 to Z- ) Based upon data from Stouffer, Dr. Zinkhan included women ages 25 to 54, and excluded 6 people who had not pur-women under 25 and over 54, and men chased frozen dinners or entrees within the last three months, people who were on medically supervised diets, and people who wore glasses but did not have them with them at the time. (F. 69, 71.) The "central anchor" of Lean Cuisine consists of purchases by women age 25 to 54. (Block, Tr. 792-93.) Dr. Zinkhan limited his sample to those women. (F. 71.) While omission of men and women under 25 and over 54 may diminish the certitude of the results, there is no evidence to show that the results would have differed if they would have been included, and there is no doubt that those surveyed were the bulJ' s eye of the target at which the ads were aimed.' The test results may therefore be relied on despite this defect. Thompson Medical Co. Inc. 104 FTC at 806-08.

5 Thompson Medical Co. , Inc. 104 FTC at 790. 6 Stouffer did not specify the percentage of mcn included in its target audience. 7 The Ross copy test, by contrast. included many who were offthc edge of the large!. (F. 121-22. Initia! Decision 118 FTC. b. Funneling The copy test began by asking three open-ended questions. (F. 75; CX-374- 29 to Z-30.) It then asked three close-ended questions each asking if the ad made any claim about one specific ingredient. (F. 76; CX-374- 30.) This pattern of questioning, called funneling, avoids suggesting answers that bias the results. National Football League Properties, Inc. v. New Jersey Giants, Inc., 637 F. Supp. 507 515 (D. N.J. 1986).

Dr. Zinkhan' s copy test asks "appropriate questions in ways that minimize bias. .. Thompson Medical I 04 FTC at 790. Funncling 9 provide unbiased evidence of questions, as used by Dr. Zinkhan 10 claims conveyed to consumers. Id. at 808. c. Open-ended questions Respondent argues that Dr. Zinkhan ' s copy test did not use a control ad to eliminate external factors affecting consumers. There is, however, no requircment of a control ad for open-ended questions. Thompson Medical, 104 FTC at 804-08. S SlOuffer s expert witness, Dr, Ross, endorsed the funneling approach (Ross, Tr. 172: F. 126. He designed a copy test for Stouffer, however. with a leading orening question askiog about specific ingredients. (RX 30- Those questions asked (CX-374. , Z-30). l. What point or points docs the Lean Cuisine ad make about the product? 2. What reason or reasons does the ad mention or suggesl for YOll to buy Lean Cuisine? 3. Is there anything eJse you can recall anout the ad? 10 Open-ended questions sometimes fail to elicit aU oflne claims an ad conveys. Kraft. Inc.. sjip op. at 13 n.13. Close ended questions with a control can also provide unbiased results, C\nd may probe Jeeper into consumers ' memories than open- ended questions. Thompsol1/Vledicol 104 FTC RQ4-06. 11 There is precedent to show that a control ad (nol a cleansed ad) may be helpful. In Thompson Medica!. the Commission approved two copy tests: the "FRC" copy test and the "AS! Theater Test. J 04 Fl' C at 804-08. The FRC copy test did not use a control ad. ld. at 804. It used control questions (regarding whether Ben-Gay or Menlholaturn contained aspirin) for the close-ended question "does the product in the commercial contain aspirin. /d. at 804. The Commission discounted responses lo the open question ("name the ingredient ) supporting Thompson. with only 3'7c recalling aspirin as an ingredient in Aspercreme; the Commission relied instead on the leading questions which showed 22% recalling Aspercreme cont,lining aspirin while the leading control questions showed that only 6'10 thought aspirin was an ingredient in Ben Gay and less than 5% perceived aspirin in mentholatum. at 804-05.

The AS! Theater" Test in Thompson Medical did include a control ad for a competing product. MobisyJ. fd. at 806. Responses to open-ended questions were that aspirin was all ingredient in Aspercreme (17(';') and MobisyI (J %). The lest also had control ingredients foi" the leading question Despite the yea saying bias indicated by the large percentage of pOirticipants who thought the control ingredients (hydrocortisone. letTlolin and menthol) were ingredients in Aspercn:mc and the control product Mobisyl the Commission relied on the result of the leading recall results indicLlting that the much larger rercentage of those who believed Aspercrerne contained aspirin than did those who saw the ). ), STOUFFER FOODS CORPORATION 781 746 Initial Dccision Marketing experts have found that credible evidence comes in response to open-ended questions, just as in trials where the unbiased testimony comes after direct, non-leading questions. The drawback of open-ended questions is that they are not as effective when the issue is the consumer s memory rather than the consumer s reaction. That is where close-end questions are effective, since they, like leading questions at trial, suggest the desired answer. They also tend to elicit bias.

Respondent argues that using a control ad for open-ended questions eliminates the influence of participants' preconceptions about the product. Even if some participants in the copy test had a prior belief that Lean Cuisine was low in sodium, that does not mean that the ads did not convey a low sodium claim. (Kloc, Tr. 442; Zinkhan Tr. 725, 729.) An ad that reinforces an inaccurate pre-existing notion is deceptive. (F, 91-92,) Not all consumers' pre-existing beliefs need to be removed from copy test results. Simeon Management Corp. FTC, 579 F,2d 1137, 1146 (9th Cir. 1978) (That the false belief " attributable in part to factors other than the advertisement itself does not preclude the advertisement from being deceptive There is no precedent mandating a control ad for open questions for a valid survey. Respondent s citations to the contrary are not persuasively A control ad was not needed for the open-ended questions in the Zinkhan copy test. (F, 83. ) There was no credible evidence that bias affected the results elicited by those questions. MobisyJ commercial. The Commission did not subtract the control responses in its analysis of the test ad. Jd. at 807-0R. The analysis dealt with responses to both leading and open questions, comparing the percentage of respondents who said Aspercreme contained aspirin (untrue) (J7o/) to those who said it contained salycin (true) (4%). Id. at 808. In Kraft, Inc. the Commission discussed the results of the C\V test done for Kraft. The question suggested that a comparison had been made ("was anything ' said or shown fin the adj that makes YOll think KRAIT Singles is different from other brands of individually wrapped cheese slices 'd. at 19 18. Refusing to rely on the results of that close-ended question, the Commission criticized the CWI copy test for not using any control. 'd. at n. 19. citing Thompson Medical where the ASI Theater Test used controls with close-ended and open-ended questions 12 :\os1 consumers ' pre- existing belicfabout the sodium content of Lean Cuisine was that it was higher than in fact was true. (F. 87- 89.) If the challenged ads changed this belief to a low sodium belief then they must have communicated that low sodium claim. (Ross, Tr. 1260-69. Respondent discerns the required use of a control ad by dlssectmg scattered statements In the 13 . footnotes of Thompson Medical and Kraft. Reply brief at pp. 23- 25. This inferred " new learning" is based, however, on misconception. Complaint counsel's reply brief at pp. 26- Initial Decision 118 F. d. Close-ended questions The close-ended questions in Dr. Zinkhan ' s copy test asked whether the ads suggested anything about the amount of three ingredients: sodium, calories, and sugar. The sugar question was a control question. (F. 106.) Close-ended questions direct participants to an aspect of the ad. Some may respond based on yea saying, inattention, or preconceptions. (F. 100-04. ) Close-ended questions require the use of a control. Thompson Medical 104 FTC at 804-06. The results of the control question are deducted from the results of the close- ended question to eliminate such bias. ld. Sugar is an appropriate control ingredient. (F. 106- 114.) It is not mentioned in the ad but is associated with Lcan Cuisine in consumers ' minds, and is an incorrect answer. (Zinkhan, Tr. 745.

The sequence of the close-ended questions was rotated. (F. 98- 99. ) Controlling for order bias is necessary to make the results of close-ended questions reliable evidence of ad communication. R.J Reynolds Tobacco Co. v. Loew s Theaters, Inc., 511 F. Supp. 867 872 (S. Y. 1980) at 872; CX-536- 25 to Z-26. The close-ended questions designed by Dr. Zinkhan minimized bias. e. Results of Zinkhan copy test Dr. Zinkhan s copy test shows that from 43 to 60% of the participants found the Jow sodium claim in response to open-ended questions. 14 (F. 116. ) Dr. Zinkhan s close-ended questions, after the control is dcducted, show from 78 to 83% of the participants took the low sodium claim from the challenged ads. (F. 119. 2. Ross copy test Stouffer s copy test uses two controls to show that the claim was not communicated. The issue is whether these control procedures biased the results of the copy test in Stouffer s favor. In Thompson t-'1dical the claims were conveyed to 16 to I !Vic . Id. at 805. Those results were derived from close-ended questions- 104 FTC at ROS. Smaller percentages ,He sufficient to eswblish that a claim is conveyed when based on open-ended results- The Gillerre Co. v. Wilkinson Sword. 111c, 89 CV J586 (KMW) (S. Y. 1991), slip op. at J7 (10%): Ross. Tr. 1299 (8- 10%) STOUFFR FOODS CORPORATION 783 746 Initial Decision a. Stouffer s cleansed control ad Stouffer used "cleansed" ads to control: participant s prior knowledge or beliefs (Ross, Tr. 961), the manner in which the close-ended question is written (Ross, Tr. 1171-72), yea saying (Popper, Tr. 1477- 79), and inattention (Popper, Tr. 1477-78). The theory is that the cleansed ad removes the elements that falsely affect the low sodium claim. Dr. Ross assumed that the cleansed ads did not convey the low sodium claim. (Ross, Tr. 1274; F. 138. The cleansed ads themselves, however, conveyed a low sodium claim. (F. 139-45.) The failure fully to cleanse the challenged ads, makes them invalid. The responses to those ads cannot properly be used to reduce the responses to open-ended or close-ended questions in the copy test. (Zinkhan, Tr. 573.) By using control ads that were likely to convey the challenged claim, Stouffer assured its "control over the study s outcome by the use of the control ads. Weight Watchers Int l v. Stouffer Corp., 744 F. Supp. 1259, 1275 (S. 1990).

Dr. Ross' removal of the sodium content modifier " less than 1 gram " and the phrase "skimp on " fails to consider that: (iJn evaluating advertising representations, we are required to look at the complete advertisements and fonnulate our opinions on them on the basis of the net general impression conveyed by them and not on isolated excerpts. Standard Oil Co. of Calif, 84 FTC 1401 , 1471 (1974), aff' d as modified, 577 F.2d 653 (9th Cir. 1978), cited in Deception Statement, 103 FTC at 179 n. 32. "The entire mosaic should be viewed, rather than v. Sterling Drug, Inc. 317 F.2d 669, 674each tile separately. FTC (2d Cir. 1963). Analysis of one or two isolated words or phrases does not result in a proper understanding of whether an implied claim is communicated. Deception Statement, 103 FTC 176 & n. 7 , 179 & 3I-32.

b. Cholesterol control Stouffer employed a control using cholesterol. The attribute in the control question must be relevant to the advertised product, but not closely enough linked with claims in the ad to convey an implied 15 Consumers "perceive the commercial in its (otality, " (CX-540- Initial Decision 118 FTC. claim about the attribute. (Zinkhan, Tr. 514- , 744-45. ) Cholesterol is linked to fat as used in the ads. (F. 150-52.) Readers infer cholesterol claims from fat claims. " (F. 150.

c. Results of the Ross copy test With cholesterol, as it did with the control ads, Stouffer selected a control that assured the outcome. Furthermore, the uni verse was defective (F. 122), and the questions were not properly rotated. (F. 129-33.) The Ross copy test is unreliable. II THE DECEPTION OF THE LOW SODIUM CLAIM During the period in which the challenged advertising ran, the Lean Cuisine line averaged 850 miligrams of sodium. (CX- CX-409-506.) This exceeds public health guidelines for "low sodium." 21 CFR 101.3(a)(3)(1992): CX- 114; CX- 520; F. 153-57; , 1230 (1976), aff' , 579Simeon Management Corp. 87 FTC 1184 F.2d 1137 (1978); Thompson Medical 104 FTC at 826. Stouffer knew that its products were not "low in sodium. (F.168-70. ) Stouffer s manager in charge of the Lean Cuisine line at the time the ads ran, testified that a "low sodium" claim was not possible because "Lean Cuisine did not meet the FDA and USDA requirements for low sodium. " (F. 168.

In the context of this market, these ads convey a low sodium message. Knowing that many consumers feel that Lean Cuisine frozen entrees contained high sodium (F. 89, 160), and that most do not know the recommended daily consumption for sodium (160-62), Stouffer took an unreasonable risk in using these ads. The healthy images and statements in the ads minimizing unhealthy ingredients and the ambiguous "less than 1 gram of sodium " and "skimp"" -- all lead to the impression of a low sodium message. The disclosure in a footnote, that "All Lean Cuisine entrees have been reformulated to contain less than I gram (1000 mg. ) of sodi- 16 The akc Sense ad expressly mentioned Lean Cuisine s fat content. Twenty-one percent of the participants (25 of 120). responded to a close-ended question that Lean Cuisine is low in cholesterol (RX- 30- 17.) Less than three percent gave that response for the 300 Like a Million ad which does not mention fat. (Zinkhan. Tr. 745-46.

17 The dcfllition of "skimp" is " scrimp," which is defined as ' to be sparing or restrictive of or in: limit severely... . '" Random House Dictionary of the English Language (2l! Ed. 1987). , g. STOUFFER FOODS CORPORA TION 785 746 Initial Decision "18 was not noticed by most consumers (F. 195), and would not be effective to dispel the net impression. Kraft slip opinion at p. 10. Reformulation and low sodium are consistent. It is not the clear contradictory element which would change the net impression of the ad. Thompson Medical 104 FTC at 799. The net message was not that sodium content was lower than it used to be, but, by clear implication, that the amount of sodium was healthfully low. IV. LOW SODIUM CLAIMS ARE MATERIAL Claims that "significantly involve health, safety, or other areas with which reasonable consumers would be concerned " are presumcd material. Kraft, Inc. v. FTC, 970 F.2d at 322-23. The calcium content claim for Kraft Singles was material because it was a health claim important to thc audience. Slip op. at 24-25. Because sodium consumption may cause high blood pressure public health organizations rccommend that Americans limit their sodium intake. The recommended daily intake of sodium is 2400 milligrams or less. (F. 158. ) The sodium in a Lean Cuisine entree has one-third of that amuunt. (F. 159. ) The low sodium claim is presumptively material to consumers of Lean Cuisine. Most consumers consider sodium important in buying frozen entrees. (F. 172. )" Sodium content is the dominant factor consumers consider in buying frozen entrces. (F. 173-75.) Consumers want precise information about negative nutritional attributes. including sodium. in frozen foods. " Over 40% of consumers are aware of the link between sodium and high blood pressure (F. 185) and they reducc their consumption of sodium. (F. 186.) Consumers relate low sodium claims to health. (F. 182-86.) A low sodium claim in a food is material to consumers and affects their purchase of frozen entrees. IS To some consumers who read the footnote 1000 milligrams" may connote high sodium. or. because they do not know the recommended daily Jllowance. it is ambiguous. (F 160. 167 . 174-75. To other consumers, who read the full context of the ad. it appar ently has a low sodium message. (F 140-45.

19 The other ads in the campaign also used the " less than I gram of sodium" language. and clements of some of the other ads may reinforce this claim of low sodium: bold type of the phrase Less than 300 calories and most with less than ! gram of sodium. " lCX-2; CX- 20 This is direct evidence of the impununce to cor.sumers ot' claims about sodium11 Kraft. Inc..

slip op. al ::3- 24; Thompso/j /Y/ediral 104 Frc at 21 Stouffer ueveioped a tine of frozen entrees that were promoted:Js having nutritionally appropriate levels of sodium. (F. 176-8 L) This evidence supports the conclusion that the low sodium claim is material Kmft. Inc.. slip op. ut::J- ::S Initial Decision 118 F. V. MILLIGRAMS DISCLOSURE The complaint charges as unfair and deceptive, and as a separate violation, the failure to disclose adequately the fact that a gram equals 1000 milligrams. The sodium content offood is commonly measured in milligrams. (F. 20 , 196.) The print ads in this case state that Stouffers' Lean Cuisine entrees contain less than I gram of sodium with footnotes explaining that I gram equals 1000 mg. of sodium. The type size of the footnotes is smaller than the rest of the ad. Fine print disclosures generally may not cure a misimpression created by the text of an advertisement. Ciant Food, Inc. 61 FTC 326, 348 (1962).

Dr. Darrel MuehJing s research tested the effect of print size on footnote information. (F. 188. ) The footnotes in the ads he tested were more complex and contained more information than the single footnotes in the challenged ads. (F. 191-92.) This survey was insufficient evidence to support the assertion that the footnotes in the challenged ads were ineffective in communicating the information that I gram equals 1000 milligrams. There was some evidence in the Zinkhan survey, however, that few consumers notice or read the footnotes in the Lean Cuisine ads. (F. 194. )23 Notwithstanding that finding, I do not believe that the failure to disclose adequately the sodium content in milligrams was unfair or deceptive. While the sodium content in milligrams is presumptively material information, the facts show that most consumers are unaware of the recommended daily allowance for sodium (F. 160-62), and knowing the precise miligrams" of sodium in an entree would be of little use. More sophisticated consumers, who are on a medically supervised diet and need precise information about sodium in milligrams, presumably read ads more carefully and would find the information in the footnote.

"- The manager fot Lean Cuisine complained to the Council of Better Business Bureaus about a competitor s ad stating sodium content in grams. That ad did not mention milligrams. (F 21.) The Lean Cuisine ads at least went a step in the right direction. 23 The Zinkhan survey excluded those on a medically supervised diet. (F 69. ) Those persons are more knowledgeable about the sodium content in food and would read the ad more carefully 24 There is evidence that many consumers do want precise information on mi1Jigrams of sodium. (1175.

, . .. STOUFFR FOODS CORPORATION 787 746 Initial Decision VI. SCOPE OF RELIEF Whether a broad fencing- " order bears a reasonable relationship to a violation depends on: "(1) the deliberateness and seriousness of the violation, (2) the degree of transferability of the violation to other products, and (3) any history of prior violations. Kraft, Inc. 970 F.2d 31 I at 326, Whether a violation is serious and deliberate depends on the cost, size, and duration of the advertising campaign and knowledge that the challenged ads were misleading. Kraft, 970 2d at 326; Bristol-Myers Co. v. FTC 738 F.2d 554, 561 (2nd Cir. 1984), cert. denied 469 U. S. I I 89 (1985). While the respondent knew that the low sodium claim was deceptive (F. 168-70), the Lean Cuisine ads were not part of a long-running television campaign. The print ads and one radio spot ran one to two times over seventeen months. The "Lean on Lean Cuisine" campaign cost $3 milion (CX-523, CX-527- , CX-528-G), far less than amounts in Bristol-Myers, American Homes Products, and Kraft. Stouffer only makes frozen food products and markets one other line -- the "Red Box" line -- for which nutritional claims are not made. "Transferability" of the violation by itself is not sufficient to justify a broad fencing- in order. Chrysler Corp. v. FTC 561 F. 357 (D. C. Cir. 1977); Fedders Corp. v. FTC 529 F.2d 1398 (2nd Cir. 1976), cert. denied, 429 U.S. 818 (1976). A broad fencing-in order is "reasonably related" to the violation when the respondent has a history of prior violations. American Home Products 695 F.2d at 707; Bristol-Myers 738 F.2d at 561-62; In re Sterling Drug, Inc. 102 FTC 395, 735 (1983). Stouffer has no history of prior violations.

This was a miscalculation rather than a blatant disregard for law. Therefore, a broad order need not issue in this case. Standard Oil Co. of Calif v. FTC 577 F.2d at 662-63.

In hompson Medical the company spent mllhon In five years advcrtlSlng Aspercreme. Thumpso/J Medical 104 FTC at 687. In Kraft the company spent $15 million ayear for two and onchalf years on mllional television of the challenged ads. Kraft. 970 F.2d aI325- 26. In !\mericut1 Home Products, the advertising cost $210 million over ten years. American Home Products 695 F.2d at 707- 08, 781.

Initial Decision 118 FTC CONCLUSIONS OF LA W 1. Thc Federal Trade Commission has jurisdiction over the advcrtising of Lean Cuisine entrees under Sections 5 and 12 of thc Federal Trade Commission Act.

2. Respondent s false, misleading, and deceptive statements as herein found were likely to mislead reasonable consumers into be- Jieving that such statements were true.

3. These acts and practices wcrc to thc injury of the public and constitute false and deccptive advertisements in or affecting commerce in violation of Sections 5 and 12 of the Federal Trade Commission Act.

4. While respondent failed to disclose adequately that 1 gram equals 1000 miligrams, that fact is immatcrial. ORDER It is ordered That respondent Stouffer Foods Corporation, a corporation, its successors and assigns, and it officers, representatives, agents, and employces, directly or through any corporation subsidiary, division or other device, in connection with the advertising, labcling, offering for sale, sale, or distribution of any frozen food product in or affecting commerce, as "commerce" is defined in the Federal Trade Commission Act, do forthwith cease and desist from:

A. Misrepresenting the sodium content of any such product. B. Describing the sodium content of any such product except by comparing it with "low sodium " and/or the rccommended daily allowance for sodium, as defined by the United States Dcpartment of Agriculture or the United States Food and Drug Administration. II.

It isfurtherordered That respondent Stouffcr Foods Corporation shall, for three years make available to the Federal Trade Commission all advertiscments covered by this order. STOUFFER FOODS CORPORA TION 789 746 Initia! Decision It isfurtherordered That respondent Stouffer Foods Corporation shall distribute a copy of this order to its operating divisions, and its officers, managers, agents, representatives, or employees engaged in advertising covered by this order and shall secure from each such pcrson a signed statement acknowledging receipt of this order. IV.

It is further ordered That respondent Stouffer Foods Corporation shall notify the Commission at least 30 days prior to any proposed change in the corporation such as the dissolution, assignment, or sale resulting in the emergence of a successor corporation, the creation or dissolution of subsidiaries or any other change in thc corporation which may affect compliance obligations arising out of this order. It isfurther ordered That respondent Stouffer Foods Corporation shall, wjthin 60 days after service upon it of this order and at such other times as the Commission may require, fie with the Commission a written report describing how it has complied with this order. . . .

Initial Decision 118 F.TC. APPENDIX A 3143468 5. .

App,Mij( II .y ,/ S7'OUPPICI? OODS COI?POl?A 7'1 IIJi/la/ D (Y'/SIOE) APPIC/YDI 79j 9250 131.1J468 /lor; "'o PElDEll(AL l'l( ADEl Co 'Al'lSS10/y Df:

'nfrial D CCIS/Of) C1Sl0/yS APPEllVD1J( 13 f 18J;t:c:

9.?$O "-4-3468 4f'1?":

/.; /, :\,,/ ,/ /. ..;yq./ :?: ;, :?, . , . ,: . . STOUFFER FOODS CORPORATION 793 746 !nitiaIDecision APPENDIX C OF AL THE THI S WE MAKE, WE MAE SENSE! ail the (,'rings we al Sroufter ' pack into our 34 Lean Cuisme ' er1rees-the freshest ingredients, the rrpest the perfect blend of herbs and vegetables ana' spices- there S " are some thmgs we skimp on: Calones. Fat Sodium With LV' ,, less than 300 calones, than gram of sodium. controlledper entr,fatweandmakealways9000'less '0 ; , C sense taste great / e'C /l1 Lean CUISine er\'e€s r. :Jn r tom1ulated ?' toOO!a,nles.TrJa 'g'arr, 800mglols.JUr / I""' LEA ON LEA - ,I II 11 111 /1....,, /'. ' 1', // ..... \ , \, : / ::.) (\.\ ..,/. /..;.,, / :/ \ / . , Initial Decision 118 ET.C. APPENDIX C THE THI S WE MAE. OF MAE SENSE! 95%fat fre. Never more than a gram of sodium' Always less than 300 calones. Lean Cuisine' ma1fes great foo and: 0 goo sese. And since all our 34 entrs are made with th frshest ingredients. Ripest etables. With the -;" c.:' pe blend of heros and Spices. Goo sense has 0:t-/ '0 .f' ner tasfed so great t / e,c- -cI./ AJll."C '""nel!r, trl!s"""" rl!tOrr, lati! IDCC'II 1gfl l!Ximgla' h, ",m / 95% FAT FREE Lean Cuisine:

n nn 1, f'Y '- oJ ,.(',JAJ, \(\\\\. . .. ; 19') ,,\\o COR"OR \o(\ ,\O\JI'I'tR foo\)'; nec.\ \)\'; C p.""t 1.6 '\7 7 .

Opinion 118 F.T. OPINION OF THE CO:vMISSION BY STEIGER Chairman:

Stouffer Foods Corporation, Inc. (Stouffer) appeals from the Administrative Law Judge (AU) James P. Timony s Initial Decision and Order holding Stouffer liable for misrepresentations regarding the sodium content of its Lean Cuisine entrees in violation of Sections 5 and 12 of the Federal Trade Commission Act (FTCA), 15 U.se. 45 , 52. Complaint counsel cross appeal the scope of the order s coverage. We affirm liability under Sections 5 and 12 of the FTCA and modify the AU' s order.

On October 28, 1991 , the Federal Trade Commission issued an administrative complaint charging Stouffer with violating Sections 5 and 12 of the FTCA by falsely representing in its ads the sodium content of its Lean Cuisine entrees. ' Specifically, the complaint alleged that certain of Stouffer s Lean Cuisine ads falsely represented, among other things, that Lean Cuisine entrees are low in sodium. Paragraph 4 of the complaint quoted language from an ad, attached to the complaint (CX-4),' which stated that Lean Cuisine " skimp(sJ on Calories. Fat. Sodium. With less than 300 calories, controlled fat and always less than 1 gram of sodium* per entree, we make good sense taste great." Paragraph 4 also quoted a footnote that appeared in the same ad which stated: "* All Lean Cuisine entrees have been formulated to contain less than 1 gram (1000 mg.) of sodium. " Paragraph 7 of the complaint alleged that Stouffer s advertising for Lean Cuisine entrees failed to disclose adequately the material fact that I The conduct chaJlengcd in this complaint occurred before the effective dale of the Nutrition Labeling and Education Act of J 990, Pub. L. No. 10 1-535 , 104 Stat. 2353 (codified in part at 21 U. 343(i), (q) and (r)).

- References to the record are abbreviated as follows IDF Initial Decision Finding Initial Decision Tc. Tnmscript of Testimony Complaint Counsel's Exhibit Respondent s Exhibit RAB Respondent's Appeal Brid CAB Complaint Cuunse1's Answering and Cross- appeal Brief RRAB Respondent s Reply and Answering Brief CRB Complaint Coullsel's Reply Brief STOUFFER FOODS CORPORA non 797 746 Opinion I gram is equivalent to 1000 miligrams, which is the commonly used unit of measurement for sodium.

The evidentiary hearings before All Timony began on February , 1993, and ended on March 8 , 1993. Proposed findings were completed on June 21 , 1993, and the Initial Decision and Order were filed on August 6, 1993. The All found that the net impression of all the elements in each of the ads is that Lean Cuisine entrees are low in sodium and that the low sodium claims are presumptively material to consumers because they involve health, safety, or other areas with which reasonable consumers would be concerned. ID at 37, citing Thompson Medical Co. 104 FTC 648, 788-89 (1984), aff' d, 791 F.2d 189 (D. C. Cir. 1986), cert. denied 479 U.S. 1086 (1987). The All concluded that Stouffer failed to disclose adequately that I gram of sodium equals 1000 miligrams, but found that fact to be immaterial. ID at 39.

The All' s order prohibits Stouffer from misrepresenting the sodium content of any frozen food product and from describing the sodium content of any frozen food product except by comparing it with "low sodium " and/or the recommended daily allowance for sodium, as defined by the Food and Drug Administration or the United States Department of Agriculture. The All declined to extend the scope of the order beyond sodium to all ingredients and nutrients because he concluded that Stouffer did not blatantly disregard the law and had no history of prior violations. ID at 39. Stouffer s principal argument on appeal is that the All erred relying on his own analysis of the challenged ads and complaint counsel's consumer survey to conclude that the ads conveyed a low sodium message. Complaint counsel do not appeal the dismissal of the miligram disclosure allegation, but cross appeal the scope of the order s coverage.

We affirm liability under Sections 5 and 12 of the FTCA. We agree with the All' s findings and conclusions to the extent that they are consistent with those set forth in this opinion, and, except as noted herein, adopt them as our own. Based on our consideration of the record in this case and the arguments of counsel for both parties, we deny Stouffer s appeal and grant complaint counsel' s cross appeal. 3 The All' s findings of fact described five print ads (CX- L CX- , eX-3, eX-5. CX-6) in addition to the one att:lChed to the complaint (CX-4) and one radio ad (CX-7). JDF 33-5 J. The AU analyzed the ad attached to the complaint (CX-4) and statements that appeared in other ads (CX- . CX- . eX- . CX-7). JD at 28- , , Opinion 118 F.T.e. The order we adopt includes a provision for coverage of all nutrients and ingredients in Stopffer s frozen food products. I. FACTUAL BACKGROUND Stouffer is a subsidiary of the Swiss corporation Nestle SA and manufactures and markets frozen foods. There are two product lines for Stouffer s frozen entree products: a full calorie product Red Box " and a reduced, low calorie product Lean Cuisine." Responding to consumers' nutritional awareness, Stouffer twice reformulated Lean Cuisine in the late 1980' s and early 1990' s with new recipes and seasonings and reduced the sodium and fat content of the products. IDF 18, 31. In order to counteract the perception that Lean Cuisine was high in sodium, and because sodium was becoming a health issue in the media, Stouffer asked its advertising agency, TathamlRSCG (Tatham), to develop ads stating the facts of the sodium content of the product. IOF 18.

II. THE CHALLENGED REPRESENT A TIONS A. Legal Framework The Commission will find deception if there is a representation or omission of fact that is likely to mislead consumers acting reasonably under the circumstances, and that representation or omission is material.' The first step in a deception analysis is to identify the claims made by looking at the ad itself.5 If, after examining the interaction of all the different elements in the ad, the Commission can conclude with confidence that an ad can reasonably be read to contain a particular claim, a facial analysis is sufficient basis to conclude that the ad conveys the claim. See Kraft, 114 FTC at 121; Thompson Medical 104 ftc at 789.

, after a facial analysis, the Commission cannot conclude with confidence that a particular ad can reasonably be read to contain a 4 See Kmji , Il1c.. 114 FTC 40, 1:20 (1991). affd. 970 F 2d 31 j (7th Cir. 1992). cert. denied. 113 S. Ct. J254 (1993): Cllffda/e Associates. IlIc. 103 FTC 110. 164- 66. 175- 76 (1984). 5 Advertising claims alt generally categorized as either- express or imrlied. Express cl:ims direclly state the representation ill issue . while i:-nplied claims, which encompass all claims that are not express. can range from those that are virtually synonymous wi h express claims to very subtle language where only relatively few consumers discern that particulm cl;1irn. Krafl j 14 FTC a! 120; Thompson Medical, 104 Frc fit 788- STOUFFER FOODS CORPORATION 799 746 Opinion particular implied message, we wil not find the ad to have made the claim unless extrinsic evidence allows us to conclude that such a reading of the ad is reasonable. Kraft 114 FTC at 121; Thompson Medical 104 FTC at 789. The Commission wil carefully consider any extrinsic evidence that is introduced, taking into account the quality and reliability of the evidence. See Kraft 114 FTC at 122. Extrinsic evidence includes, but is not limited to, reliable results from methodologically sound consumer surveys. Kraft 114 FTC at 121; Clifdale, 103 FTC at 164-66. In detcrmining whether a consumer survey is methodologically sound, the Commission will look to whether it "draws(sJ valid samples from the appropriate population ask(sJ appropriate questions in ways that minimize bias, and ana- Iyze(s) results correctly. Thompson Medical 104 FTC at 790. The Commission does not require methodological perfection before it will rely on a copy test or other type of consumer survey, but looks to whether such evidence is reasonably reliable and probative. See Bristol-Myers Co. 85 FTC 688, 743-44 (1975). Flaws in the methodology may affect the weight that is given to the results of the copy test or other consumer survey.

Whether examining the ad itself, extrinsic evidence, or both, the Commission considers the overall, net impression made by the ad in determining what claims may reasonably be ascribcd to it. Kraft 114 FTC at 122; Thompson Medical 104 FTC at 790. To be considered reasonable, however, an interpretation nccd not be the only interpretation as long as the subset of consumers making it is representative of the group of consumers to whom the ad is addressed. Kraft, 114 FTC at 120-21 n. 8; Thompson Medical 104 FTC at 789 n. The second step in a deception analysis is to determine if the claim is likely to mislead a consumer acting reasonably under the circumstances. Cliffdale 103 FTC at 164- 65, 175-76. Where more than one meaning is conveyed by an ad, one of which is false, thc seller is liable for the false claim. Kraft 114 FTC at 120-21 n. Thompson Medical 104 FTC at 789 n.

The final step in a deception analysis is to determine whether the claim is material. Cliffdale 103 FTC at 164- 65. Information is material if it is likely to affect a consumer s choice of or conduct regarding a product. Id. at 165; Kraft 114 FTC at 134. There are several types of claims that the Commission presumes to be material: express claims; implied claims where there is evidence that thc seller intended to make the claim; and claims or omissions involving health Opinion J 18 F.TC. safety, or other areas with which reasonable consumers would be concerned. Kraft 114 FTC at 134; Thompson Medical 104 FTC at 816- 17; Clifdale 103 FTC at 182-83.

B. Respondent s Advertising From January 1990 through August 1991 , Stouffer ran a series of ads, including the Make Sense ad attached to the complaint (CX-4), as well as five other print ads: Lean on Lean Cuisine (CX- I), 300 Like a Million (CX-2), another version of 300 Like a Milion (CX-3), another version of Make Sense (CX-5), and Ole O' Lean (CX- 6). In addition, Stouffer ran a radio ad, Anniversary Turkey (CX-7). The Make Sense ads (CX-4 and CX-5) show a plate of chicken vegetables, and pasta, and a man and a woman on a bicycle. The headlines state:

OF ALL THE THINGS WE MAKE, WE MAKE SENSE' In the first Make Sense Ad (CX-4), smaller print follows this headline which states:

Of all the things we at Stouffcrs pack into our 34 Lean Cuisine entrees -- the freshest ingredients, the ripest vegetables and the perfect blend of herbs and spices -- there are some things we skimp on: Calories. Fat. Sodium. With less than 300 calories, controlled fat and always less than 1 gram of sodium* per entree, we make good sense taste great.

The footnote in both of the Make Sense ads (CX-4 and CX- 5) states in even smaller print lower in the page:

All Lean Cuisine entrees have been reformulated to contain Jess than 1 gram (l000 mg. ) of sodium.

The radio ad, Anniversary Turkey (CX-7), contains explicit language regarding low sodium: "These babies are healthier than ever. Lower in sodium, fat and cholesterol. Read those boxes, people these numbers are low.

The text in the other version of the Make Sense ad (CX S) states: 95% fat free. :"ever more than a gram of sodium* Always less than 300 calories. Lean Cuisine makes great food and good sense. And since all our 34 entrees are made with the freshest ingredients. Ripest vegetables. With the perfect hlend of herbs and spices. Good sense has never tasted so great. STOUFFER FOODS CORPORATION 801 '46 Opinion We agree with the AU that a facial analysis of the ads (CXthrough CX-7) permits us to conclude with confidence that the ads can reasonably be read to convey a low sodium message ' Several elements of the ads communicate this message, including the headlines, the language used, and the footnotes. One message from the print ads (CX- l through CX-6) is that Lean Cuisine has large quantities of healthy ingredients and small quantities of undcsirable nutrients. IDF 45. The AU concluded, and we agree, that the words "We Make Sense" in the headline (CX-4 and CX-5) condition the reader to think that Lean Cuisine is a healthy producl." The text in the body of the Make Sense ads (CX-4 and CX- 5), for example, further emphasizes sensible eating with language such as: "the freshest ingredients, the ripest vegetables and the perfect blend of herbs and spices" which are "packed" into the Lean Cuisine entrees. The text in the body of the Lean on Lean Cuisine and 300 Like a Million ads (CX- l through CX-3) also cxpresses the sensible eating message with such language as: "Stouffer s recipes use only the finest ingredients at their natural pcak of perfection combined in exciting and imaginative ways. The ads also represent that there are Jow levels of undesirable nutrients. IDF 46. For example, the Make Sense ad (CX-4) represents that the negative attributes, such as "Calories. Fat. Sodium. are "skimp(edJ on." The additional language in the Make Sensc ad (CX-4) "With less than 300 calories, controlled fat and always less than 1 gram of sodium* per cntree..." also reinforces the low sodium message. These representations communicatc that the negative attributes have been reduced to meager quantities. ID at 28. Similarly, the text in another version of the Make Sense ad (CX-5) provides Never more than a gram of sodium. *"9 In addition, we agree with the AU that describing sodium as Jess than" 1 gram reinforces the impression that sodium is prescnt The AU concluded that all of the print ads are similar and that the same analysis applied to them. ID at 29 n.4. The also found that the radio ad, Annivcrs.'ry Turkey (CX-7). comained a low sodium message . JD at 29.

8 Similarly, the headline "Lean. (3 smart. intelligent. sensible ",'ay to eat, )" in the Lean on Lean Cuisine ad (CX- l) evokes sensible eating I, The Lean on Lean Cuisine ad (ex" I) represents that " (10 Jach of our 30 entrees has less than 300 calories and most have less than a gram * of sodium:' The OO Like a Million ads (CX-2 and eX. claim that " fnJobody else knows how to crCclte such great lasting entrees, ail with less than 300 calories and most with less than I gram* of sodium. " The Ole C)' Lean ad (CX-6) states that "(sJo each has less than 300 caiories and less than one gram" of sodium " Opinion I J 8 F.T.c. in only a minimal quantity. 'o The language "less than" (CXthrough CX-4 and CX-6) minimizes the sodium content, and the number " I" appears in context to be a low number. Indeed, as the All noted, in the 300 Like a Milion ads (CX-2 and CX-3) the phrase and most with less than I gram* of sodium" was emphasized in bold print. IDF 144.

Accordingly, we find, as the All did, that the net impression of the elements in each of the print ads is that Lean Cuisine products are low in sodium. ID at 29. Moreover, we find that the radio ad, Anniversary Turkey, (CX-7) also communicates that Lean Cuisine sodium content is low. The ad (CX-7) expressly states that Lean Cuisine products are "(IJower in sodium, fat and cholesterol. . . . these numbers are low.

On appeal Stouffer argues that the All ignored elements in the challenged ads (I) which are contrary to a "low" sodium message, and (2) which reasonably convey that the reformulated Lean Cuisine products have a "reduced" or "lower" quantity of sodium, rather than an absolute "low" amount. " RAB at 17. We have carefully reviewed the ads in their entirety, including the elements referred to by Stouffer. We conclude that the low sodium claim is made. Stouffer argues that since great taste was a key element in the campaign and the perceptions associated with low sodium are those of poor taste, then the taste component of the ads contradicts the low sodium message. RAB at 21. We do not disagree with respondent that the ads convey a superior taste message. Where we disagree is over respondent s unsupported contention that such a message necessarily contradicts a low sodium claim or that the existence of a nondeceptive message precludes our finding an implied deceptive claim. RRAB at 1-3. Stouffer relies on the testimony of Irene Block a partner at Tatham, its advertising agency, who provided conclusory testimony that the perceptions associated with low sodium are those of poor taste and that this would contradict any low sodium message. Tr. 787. Ms. Block offered no empirical support for her conclusion and her testimony. As stated above, it is well settled that an ad can The AU concluded that the preponderance of the evidence shows that the footnotes in the ads did not adequately disclose rhat one gram equals 1000 mliligrams. IDF 195: ID at 29. We agree with thiscondusion.

J j Stouffer does not appeal the All' s findings that a low sodium claim was false and misleading. lD at 39. Stouffer also does not challenge the materiality of a low sodium claim . TO at 39 g., , . . . . , . STOUFFER FOODS CORPORA non 803 746 Opinion convey more than one claim and that not all of the claims need be deceptive in order for the ad itself to be deceptive. Cliffdale, 103 12 Therefore, we see nothing inherently inconsistentFTC at 178. between a low sodium message and a superior taste message. For those concerned about sodium consumption, a product with low sodium and great taste would be attractive.

In addition, Stouffer argues that the relative nature of the terminology used in the ads conveys a reduced, rather than low sodium claim. RAE at 23. Even putting aside the fact that much of the terminology used in the ads is absolute, not relative, 13 we see no basis for concluding that reduced and low sodium claims are mutually exclusive. Indeed, reducing the amount of an element will often result in diminishing that element to a low level.'4 Stouffer also argues that the alleged low sodium claim is neither conspicuous" nor "self-evident" from the face of the challenged ads " and that to find a violation premised on a facial analysis of the ads would unreasonably chill Stouffer s exercise of its commercial speech rights. Accordingly, Stouffer argues that a methodologically valid consumer survey demonstrating that a substantial number of consumers take away the alleged low sodium claim is constitutionally required under the First Amendment. RAE at 28-29. We hold that there are no First Amendment concerns raised where, as here, facially apparent deceptive implied claims can be found without resort to extrinsic evidence Further. the Commission has held that an ad can make a deceptive lmpl1ed claim even if thead contains12contrary elements, as long as those contrary elements do not effectively ncgO!tc or qualify the implied claim. See Kraft I J4 i-iC at 124; Rell()\'I/"l1lnl'! Corp. ! II FTC 206, 294 (1988). affd. 884 F.2d 1489 (J 51 Cit. 1989).

13 See. e. CX- through CX- , CX- 6: (" less than 1 gram of sodium ); CX-5 ("never more than a gram of sodium ); CX-7 ("these numbers are low This canciuslOn \s confirmed by the copy test results mtroduced by complaint counsel which spccificaJly indicate that only a relative minority of consumers stated that the ads conveyed a reduced sodium claim. There 'I..as a specific coding category for open-ended questions in the copy test for less/lower/reduced sodium. Between percent and 14 percent of the respondents to the open-ended questions stated that the ad conveyed a lessllower/reduced sodium message. As noted infra 43 percent (060 percent of respondents responded that the ad conveyed a low sodium message. CX-374z- J I. 15 Stouffer appears to argue. citing Kraft and Thompson. that the absence of a visual image coupled with a written or verbal message prevents the Commission from finding an implied message in ads. RAB at 25-27. Although the Commission determines ad meaning from the ad taken as a whole the Commission has never required a visual image before making such a determination. In addition. Stouffer argues that the All erred in finding a low sodium claim on a facial analysis of the radio ad Anniversary Turkey (CX-7). RAB at 30- . In light of the express nature of the language in this ad and bccausc of its similarity to thc other ads C reject this argument See 20uderer v. Ohio, 471 S. 626 652- 53 (1985); Kraft 970 F.2d at 321 Opinion j I R F.T.C. From a facial analysis of the ads themselves we conclude that they convey a low sodium message. It would be, therefore, unnecessary to resort to extrinsic evidence. Nevertheless, consistent with our practice we have examined the extrinsic evidence offered on this issue by complaint counsel and find that it corroborates our conclusions regarding ad meaning. We turn next to a consideration of the extrinsic evidence.

C. Extrinsic Evidence Both complaint counsel and Stouffer proffered the results of copy tests conducted for this adjudication. The AU found that Stouffer copy test was unreliable. ID at 35. Stouffer does not appeal the AU' s rejection of its copy test evidence. Instead, Stouffer appeals the attribution of probative value to the methodologies employed in complaint counsel's copy test.

Complaint counsel engaged U.S. Research Company to conduct a copy test of three of the print ads 17 to determine if they conveyed the low sodium claim. The questionnaire used was designed by Dr. George Zinkhan, a professor of marketing at the Cniversity of Houston. IDF 52. Dr. Zinkhan determined an appropriate universe for the copy test " by relying on Stouffer s description of its target audience. IDF 70. The questionnaire contained six questions using both open-ended and closed-ended formats. An open-ended question provides copy test participants with an opportunity to provide answers phrased in their own words. " A closed-ended question asks about a specific issue and provides a choice of answcrs from which the consumer selects." lDF 53. Here, the questionnaire used a funneling approach which began with general, open-ended questions and led to more narrow, closed-ended questions on specific issues. IDF 73. The experts for both Stouffer and complaint counsel agree that funneling is the best way to ask questions on a copy test. IDF 74. 17 The three print ads tested were L(Can on Lean Cuisine (CX- l). 300 Like a Million (CX- 3). and Make Sense (CX-4). One hundred participants viewed each of the three ads at tour shopping malls across the country. IDF 55. 57 64: CX-374b. 18 That universe consisted of women \vho ""'ere the principal food shoppers for their household. were between the ages of25 and 54. had purchased a frozen entl ee in the last three months ane. were not following a medically supervised diet. TOF 69: CX- 374c Zinkhan. Tr. at 478 10 Zinkhnn, Tr. at 478.

STOUFFER FOODS CORPORA non 805 746 Opinion asked consumers what general point The open-ended questions'l or points the ads made. These questions were designed so as not to prompt participants for any particular response, or to give any context in which to answer the questions. The questions permitted participants to give one answer, multiple answers or no answer. IDF 79. No control group or control question was included in this portion of the survey. !DF 77-83; CX-374.

A majority of the copy test participants responded to the openended questions in a way that indicated that they received a low sodium message from the ad. Specifically, 43 percent to 60 percent stated that the ad communicated that the Lean Cuisine entrees are low in sodium. !DF 116; CX-374z- 11. This response rate is quite high. The COrmssion has found far lower response rates from open-ended questions to be significant. In Thompson Medical, for example, the Commission found that the ASI Theater test results, in which 17 percent of the Aspercreme ad viewers respondcd that the product contained aspirin in response to the unaided recall questions, was a sizablc percentage of participants who did not perceive or remember the disclosure that Aspercreme does not contain aspirin. 104 FTC at 808. We note that even Dr. Ivan Ross, one of Stouffer s experts, testified that often a rescarcher must rely on open-endcd responses in the magnitude of 8 percent to 10 percent as being meaningful. Tr. 1299.

The closed-endcd questions" dcsigned by Dr. Zinkhan asked if the ad suggested anything about the amount of sodium, calories or sugar in Lean Cuisine entrees. Participants were asked specifically whether the amount of the attribute was high, low, neither high nor low, or whether they did not know or remcmber. The order of thc closed-ended questions was rotated to minimize bias. IDF 99. - The following three open-ended questions were asked" I. What point or points does the Lean Cuisine ad make ablaut the rroduct? 2. \\'hat reason or reasons does the ad mention or suggest for you to buy Lean Cuisine 3. Is there anything else you can recall about the ad? C:X-374z- 29 - 3() -- The following closed-ended questions were asked in rotating order Does the ad SCly or suggest anything about the amount of calories (or sugars lor sodiumJ in Lean 2. CuisineDoes theentreesad 53)')orHycssuggest that Lean Cuisine entrees me: - high in caiorics for sugars (or sodiumj - low in calories for sugars lor sodiumJ - neither high nor low in cJ.lories for sugar) for sodiumj - don t kno\" " don ll"emembeT"

See CX-J74z- Opinion 118 FTC. Dr. Zinkhan incorporated several mechanisms in the design of the closed-ended question section of the copy test to minimize bias. One form of bias is "order bias," meaning that the sequence in which the IDF 98. The questions are asked can affect the results.'J Zinkhan copy test minimized "order bias' by rotating the order of the closedended questions. IDF 99. Other forms of bias include "yea saying, which is the tendency to give the answer the participant believes the interviewer is seeking, and "halo effect " where the participant has a favorable opinion of the product before taking the test and therefore answers questions with that favorable impression in mind. IDF 100- 101. In an effort to control for yea saying, inattention, halo effect or other noise factors, Dr. Zinkhan used a control question by repeating the questions relating to sodium, and substituting the word sugar for the word sodium.'" IDF 104 , 106. In using a control question, the percentagc of participants who responded affrmatively to the control question is deducted from the percentage of participants who responded affirmatively to the tested claim. IDF 106. In the Zinkhan copy test, after deducting the percentage of respondents who answered yes to the control question, 78 percent to 83 percent of the respondents found a low sodium claim from the ad. IDF 1 19. These results are unusually high and consistent with the responses to the open-ended questions.

We find that the Zinkhan copy test provides reliable and probative evidence and is methodologically sound. The results appear to be strikingly high for both the open and closed-ended questions and confirm the conclusion that we rcached based on our facial examination of the ads. Indeed, Stouffer s experts have previously relied on copy test results with much lower response rates. See supra at 11. Further, the Commission has likewise relicd on copy test results with lower response rates. See Thompson Medical 104 FTC at 808. Stouffer contends that the methodology employed in the Zinkhan copy test is so fundamentally flawed that the AU crred in relying upon the test results. RAE at 33. Specifically, Stouffer argues that some number of survey respondents may have come to thc test with the preexisting belief that Lean Cuisine frozen entrees are low in sodium. According to Stouffer, these "biased" participants may have 23 -1 he first question sets up the survey. and the results ot the last question may he affected by fatigue or boredom. Zinkhan, Tr. SS3- SS-i An pprorriatc control question asks ClDout a product at!ribule that is rein' an! to and reasonably :lssociated with the product. hui is not too closely linked to a claim in the ad, IDF 105 , ): . STOUFFER FOODS CORPORA non 807 746 Opinion responded to the various survey questions on the basis of their preexisting opinions, and without regard to the actual content of the advertisements. Specifically, Stouffer made this argument with regard to open-ended questions (e. What point or points does the Lean Cuisine ad make about the product ). RAB at 33 , RRAB at 18. Stouffer contends that the copy test could have employed a control group exposed to a control ad in order to quantify and eliminate the effects of participants ' preexisting bias. " RAB at 36. Stouffer also argues that the use of the control question in the closed-ended questions was inadequate. RAB at 41. Since there were no adequate controls, Stouffer concludes, the Zinkhan copy test may not be given any weight whatsoever. To support its arguments, Stouffer relies on Thompson Medical and Kraft.

Perfection is not the prevailing standard for determining whether a copy test may be given any weight. The appropriate standard is whether the evidence is reliable and probative. See Bristol-Myers, 85 FTC at 744. The Kraft decision instructs that, in al1 cases involving contested issues of ad interpretation the Commission will carefully consider any extrinsic evidence that is introduced, taking into account the quality and reliability of the evidence. " 114 FTC at 122. "The quality of any consumer research offered as evidence will be evaluated in the totality of the circumstances. . . Id. at 127 n. 13. A study may be flawed, that is, harbor one or more sources of potential error or bias, and still be probative." Thc nature and seriousness of any deficiencies will affect the weight that the Commission assigns to that piece of evidence. On the other hand, if the methodology of a consumer survey is fundamentally unsound, then that survey cannot assist the Commission in deciding whether an advertisement communicates a particular claim to consumers. Thompson Medical, 104 FTC at 794- 95; Sterling Drug, 102 FTC 395 , 754 (1983), aiid, 741 F.2d 1146 (9th Cir. 1984). The Commission s practice is, in this A control group IS a group of participants who see a stimulus different from the challenged ad - i. e., 25a "cleansed" Lean Cuisine ad that does not convey the hypothesized low sodium claim. The control group is then asked the same series of questions as the lest group. The control group s low sodium answers are subtracted from the low sodium results obtained from viewers of the chdllcnged ad to control for the purported preexisting belief. See Kraft 114 i-iC at 126- 27 n. 13 (';Although we agree with respondent that the design of the MOR survey qucstionnaire is nol without /laws. and that alternative or additional means could have been used to better minimize the potential for yea-saying bias inherent in using a closed-ended question format, on balance, we find the MOR survey results to he of some proh3tive v3.lue. Thomp.\o/J /vtedicaf 104 FTC at 796-97 (survey that hets "several potential sources of bias" nonetheless deemed to be "reasonably reliable extrinsic evidence Opinion 118 F. regard, consistent with that of most federal courts when evaluating surveys purporting to assess the meaning that consumers take from ads.

As discussed below, we find that the survey offered by complaint counsel was reliable and probative. Accordingly, it was proper for the AU to rely upon this extrinsic evidence, together with the facial analysis, in concluding that a low sodium claim is present in the Stouffer Lean Cuisine ads.

The AU found that there is nothing in Commission precedent mandating a control ad for open-ended questions, that Stouffer reliance on Thompson and Kraft is misplaced, and that there was no credible evidence that bias affected the responses elicited by those questions. ID at 33.

We agree with the ALl There is nothing in Commission precedent that requires the use of a control ad for open-ended questions. The Zinkhan open-ended questions properly attempted to elicit unprompted responses in a consumer s own words describing what he or she took away from the ad." In addition, the Zinkhan openended questions properly continued to probe for more responses. We therefore reject Stouffer s argument that the responses to the openended questions are fatally flawed because of the absence of a control ad.

We also agree with the AU with regard to Stouffer s argument concerning the requirement of a control ad in closed-ended questions. The Commission has long recognized that a control of some kind is necessary for closed-ended questions, and has noted, for example that there is a potential for yea-saying inherent in the closed-ended question fonnat. Kraft 114 FTC at 126 n. 13; see Thompson Medical 104 FTC at 804-08. The Commission, however, has never dictated the type of control necessary in a copy test. There is nothing in Commission precedent that requires the use of a control ad for closed- 27 See, e. McCarthy, Trademarks and Unfair Competition, Section 32.50 (3d. ed. 1992) ("n an extreme case. an improperly conducted survey with slanted questions or serious methodological defects may be excludable as ' irrelevant ' of the true state of mind of potential purchasers. But the majority rule is that while technical deficiencies can reduce a survey s weight. they will nol prevent the survey from being admitted into evidence. As one court correctly observed. ' No survey is perfect' and flaws in questions and methodology should only affect the weight accorded survey results. ) (footnotcs omitted) (quoting Selchow Righter Co. v. Decipher, Il1c.. 598 F. Supp. 1489 (E.D. Va. ! 984) 2H The claim at issue, low sodium. is both a simple cluim ami a primary one, making it particularly well SUill:d to !hl: open-ended format. On the other hand, open-ended questions are likely to understate secondary implied claims, particularly where, as in Kraft, those claims are also rather complex by virtue of being hath compound and comparative STOUFFR FOODS CORPORA non 809 746 Opinion ended questions. Dr. Zinkhan' s closed-ended questions were designed in a way that minimized bias through the use of a control question and by rotating the sequence of the questions. We find that the use of sugar in a control question was appropriate. Stouffer argues that the failure to control for preexisting beliefs is necessarily such an extreme error that a copy test that is flawed in this respect in entitled to no weight. However, the expert testimony cited by Stouffer is unconvincing, and the case law is to the contrary. Stouffer s two expert witnesses, Edward T. Popper and Ivan Ross did opine that a control ad is needed in order to account for and eliminate the effects of preexisting beliefs. Yet, the basis for this conclusion is unclear. Neither witness cited evidence that this sort of bias is common or significant in advertising copy tests. Both admitted that they had previously designed copy tests for litigation purposes that did not include a control ad group. Both further acknowledged that they had given sworn testimony regarding ad claims based upon the results of tests that did not employ a control group. IDF 83-86. Finally, there is no record evidence that, among experts in advertising or consumer research, the use of a control group is considered a since qua non of a valid copy test. In this regard, we note that complaint counsel's expert witnesses testified that the Zinkhan copy test is valid and reliable evidence of what claims the Stouffer ads communicated without the need for a control group. IDF 81. Copy tests are frequently evaluated by federal courts in the context of Lanham Act cases and other litigation. Stouffer has cited no case concluding that a study will not be deemed reasonably reliable unless it controls for preexisting bias. In fact, there are numerous cases relying on copy tests without any discussion of the use of Simi-a control group or the need to factor out pre-existing beliefs.'o larly' the Commission has often relied on copy tests that did not employ a control group. E.g., Thompson Medical 104 FTC at 796-97; American Home Products Corp. 98 FTC 136 , 394 (1981), enforced Stol1ffer also challenges the use of a control question as insufficient to correct for tho e whobase their29responses on pre-existing belief. RAE at 41. RRAB at 5. Complaint counsel has not argued that the use of a control question is appropriate where it is necessary to control for pre-existing beliefs. Further, as noted infra the record fails to establish that pre-existing beliefs affected the Zinkhan copy test results.

30 E.g. , Mc/Ilei/ah, lnc. Ii. American Home Prodllets Corp. 675 F. Supp. R19, 825 (S. 1987), affd, 848 F.2d 34 (2d Cir. 1988): Sliffel Co. v, Weslwood Lighting Group. 658 F . Supp, 1103, 1112- 14 (D.KJ. 1987); American HOIlf' Products Corp. v. Johnson Johl1\u/J 654 F. Supp. 568, 581\- 89 (S. NY 1987) Opinion 118 FTC. as modifed 695 F.2d 681 (3d Cir. 1982); Bristol-Myers Co. 85 FTC at 744 (1975).

The only Commission decision that directly addresses the issue of pre-existing beliefs is Kraft, and it is on this case that Stouffer principally relies. RAB at 42-43. The record in Kraft included reason to be concerned about the possible influence of pre-existing bias upon copy tests. The Commission evaluated two series of ads for Kraft Singles processed cheese. The "Skimp" ads were the first to be disseminated, and contained the explicit (but deceptive) representation that Kraft Singles contain more calcium than do most imitation cheese slices (a superiority claim). The "Class Picture/5 ounce" series of ads was introduced 15 months later, and "contained no explicit comparison between Kraft Singles and non-dairy slices. 114 FTC at 130. As evidence that the "Class Picture/5 ounce" ads contained an implied superiority claim, complaint counsel offered a copy test that did not control for consumers ' preexisting beliefs regarding the relative calcium content of Kraft Singles. The Commission concluded that this test was not reasonably reliable, explaining that "(tJhe apparent 45 percent response rate suggesting that an imitation superiority message was taken by survey participants may well be attributable to consumers' prior exposure to the ' Skimp' ads which did contain an explicit comparison to imitation slices, and which were disseminated extensively prior to the ' Class Picture/5 ounce' ads. Kraft, 114 FTC at 131 n. 19. This passage must be read in light of the Commission s other pronouncements on copy testing, and in particular the admonition to evaluate the "totality of the circumstances" bearing on the reliability of any consumer research. The case does not hold that consumer Insteadsurveys must invariably control for preexisting beliefs.31 Kraft teaches that the failure of a consumer survey to control for preexisting beliefs about the alleged advertising claim introduces a potential for bias, and indeed that this may be a critical defect. In any event, there must be evidence of preexisting bias to find that failure to control for such bias is a critical defect. In Kraft, there was evidence that (i) a large portion of consumers had a preexisting belief with regard to the superiority claim, and (ii) this preexisting belief had likely biased the consumer survey results relied upon by 3\ Indeed. it is established that respondents may be held liable I' Of dissemination of ads that capitalize on preexisting consumer beliefs. Simeon Management Corp. v. FTC 579 F.2d ! 137, ! 146 (9th Cir. 1978).

STOUFFR FOODS CORPORATION 811 746 Opinion complaint counsel. In the present case, the preponderance of the evidence indicates that, to the extent that consumers have any preexisting beliefs about the sodium content of Lean Cuisine entrees they likely believe that such products are high in sodium, not low. IDF 87- 89. Further, Stouffer cites no evidence that preexisting beliefs affected the survey results attained by Dr. Zinkhan; respondent s objections to the study are wholly theoretical. On the present record, it appears that the Zinkhan test was sufficiently reliable to constitute probative evidence on the issue of ad meaning. We therefore find that reliable and probative extrinsic evidence corroborates our conclusion, based on our facial analysis of the ads, that the Stouffer Lean Cuisine ads communicate a low sodium message.

II ORDER COVERAGE It is well settled that the Commission can issue orders containing fencing- in requirements. See, e. g., FTC v. Ruberoid Co. 343 U. 470, 473 (1952). This discretion is limited by two constraints. First the order must be sufficiently clear and precise to be understood. See, e. , FTC v. Colgate-Palmolive Co. 380 U. S. 374 , 394- (1965). Second, the order must bear a reasonable relationship to the unlawful practices. See, e. , Jacob Siegel Co. v. FTC 327 U.S. 608 612- 13 (1946).

Complaint counsel argue in their cross appeal that the AU erred in narrowing the notice order s claim coverage because he improperly weighed and evaluated the evidence of the seriousness and deliberateness of the violations and failed to consider the transferability of the type of claims made. CAB at 73-74.

The three criteria used by the Commission to detennine whether order coverage bears a reasonable relationship to a particular violation of Section 5 include: (I) the seriousness and deliberateness of the violation; (2) the ease with which the violative claim may be transferred to other products; and (3) whether the respondent has a J2 All of the three elements need not behistory of prior violations. present to warrant fencing- in relief. See, e. , Kraft 114 FTC at 142 (lack of history of prior violations did not make fencing-in improper). In considering these three elements, the Commission looks both to the presence or absence of a particular element and to the circum- 32 See Kmfr 114 FfC at 139 970 F.2d 3!! at 326: Thompson Medica! 104 FTC at 833. ), , Opinion 118 FTC. Consideration of the three elements leads us to stances as a whole.3J conclude that two of them -- (I) the deliberateness and seriousness of the violations and (2) the transferability of the unlawful practices to other products -- combined with the overall circumstances justify extending the order beyond the products for which the challenged claims were made.

The All articulated the proper standard for deciding whether fencing-in relief was appropriate, listing the criteria identified above. The All determined that Stouffer knew that its low sodium claim was deceptive. He appears to have found more compelling, however his assessment of the campaign as one of not long duration that cost far less than the amounts spent on other campaigns where the Commission has found serious violations, such as those in Kraft and Bristol-Myers. JS Finding, in addition, that "Stouffer only makes frozen food products and markets one other Jine . . . for which nutritional claims are not made, . . " (10 at 39) and that "(tJransferability of the violation by itself is not sufficient to justify a broad fencingorder (Jd. the All concluded ltJhis was a miscalculation rather than a blatant disrcgard for law. Therefore, a broad order need not issue in this case. Id. We disagree. The seriousness of the claim stems from the overall health ramifications of any sodium claim and, particularly, of a claim that a product is low in sodium when it is in actuality relatively high in that ingredient. The seriousness of the violations here is enhanced by the fact that consumers cannot readily judge for themselves thc truth or falsity of a low sodium claim. See Kraft 114 FfC at 140. The seriousness of the violation is further increased by the health-related nature of thc low sodium claim. There is medical evidence supporting a link between sodium consumption and high blood pressure, for some peoplc, on which basis such organizations as the National Academy of Sciences, the American Heart Association, and the Surgcon General of the United States recommend that consumers limit their daily sodium intake. See IOF 171. The cost and extensiveness of the ad campaign are not determinative, but they too may be relevant in assessing the seriousness and 33 Smrs, Roebuck Cu.. 95 FTC 406 (191\0), ajTd 676 F.:!d 385, 392 (9th Cir. 1982) 3.j Stouffer dotto not have a history of prior viol::ions. Brisro/-Mvers Co.. 102 FTC 21 (J981), atrd. 738 F.2d 554 Cd Cir. IlJR4). eer. denied. 469 S. 1189 (1985) STOUFFER FOODS CORPORATION 813 746 Opinion deliberateness of a violation. The Stouffer ad campaign that the Commission finds to be deceptive ran from January 1990 through August 1991. This campaign cost three million dollars and reached millions of consumers nationwide. IDF 33. The print ads at issue (CX- J through CX-6) appeared in eighteen different magazines from January 1990 through the first four months of 1991 , including such nationally distributed periodicals as People, Newsweek, Good Housekeeping, and Ladies Home Journal. IDF 37 39, 42. The radio advertisement, Anniversary Turkey (CX-7), was played on over 230 radio stations from June through August 1991. IDF 44. Such a distribution scheme would have reached approximately 70 percent of the population of the United States. Block, Tr. 797- 798. While not necessarily expensive when compared to campaigns that included television advertising, the campaign was far-reaching. Both the cost and the length of an ad campaign are measures of how widely the ads were disseminated, but they are not the only such measures. Here the publication of print ads in magazines of nationwide distribution and the broad distribution of the radio ad brought the objectionable J6 We believe that the adsads to large numbers of consumers. exposure contributed significantly to the seriousness of the violations before us. The evidence as to the success of the campaign in reaching consumers, therefore, weighs in favor of a broader order. As the AU found, the record also shows that Stouffer was aware of the potential risks and benefits of focusing on sodium in its ads. As the campaign began in 1990, a Tatham memorandum reporting on a telephone conference with Richard B. Annett, Stouffer s Group Marketing Manager for Lean Cuisine, noted that Stouffer "informed fTatham) that ' lower' sodium or ' controlled' sodium were acceptable terms but ' low sodium' was not possible. " CX-44a. It appears, therefore, that Stouffer was well aware that a low sodium claim was inappropriate for Lean Cuisine (IDF 169- 170) and that the characterization of sodium was a delicate matter.

Despite the delicate nature of the sodium message, however, the message projected consistently throughout the ad campaign stressed what Mr. Annett described to Tatham in a memorandum of January , 1990, as the "buzz words" used by competitors, such as "health" and ingredients with negative connotations like sodium, fat and cholesterol CX-26. Mr. Annett instructed Tatham: 36 See Tl riJ1/ :'vedicol. 104 ftc at 833-34 (analyzing dissemination of cen,lin claims th,ll ran only in print ads in finding lh3.t broad fencing-in was wamlntcd). Opinion 118 F. !'OTE THE STROI'G ' BUZZ WORDS' Iour competitor) USES IN THEIR PRINT:

HELPS YOU LIMIT CHOLESTEROL, SODIUM , AND FA T. NO MORE THA;\IOG OF FAT.

FOR SODIUM WATCHERS.

THESE ARE THE TYPES OF HOT BUTTONS WE MUST USE. THEY (competitors) HA VE TAKEN NEGATIVES AND TURNED THEM INTO POSITVES.

Id. Mr. Annett s memorandum also critiques one of Tatham s suggested ads for the Lean Cuisine campaign, saying: IT DOESN' T SCREAM HEALTH ENOUGH. .. . THE USE OF 'LEAN' IS EXCELLENT LEAN ON CALORIES, FAT, AND CHOLESTEROL BUT SODIUM SHOULD ALSO BE INCLUDED.

Id. Mr. Annett s directions to Tatham provide context for the implied low sodium claims we have found deccptive and, in doing so, they enhance the seriousness of the claims by reinforcing their relationship to good health.

For these reasons, we find Stouffer knew or should have known that the ads were likely, through their words and images, to communicate a false low sodium claim. ID at 36, 39. We find that under these circumstances, Stouffer s action was deliberate. See Thompson Medical, 104 FTC at 835.

We also find that the risk of transferability of the violation justifies broader order coverage. False nutrient content claims regarding the amount of sodium in frozen food appear to be readily transferable to claims for other nutrients and ingredients. In Kraft 114 FTC at 141 , the Commission noted that "(T)he violations in this case are readily transferable to other Kraft cheese products " citing Thompson Medical 104 FTC at 837 , and American Home Products, 98 FTC 136, (1981), aff'd 695 F.2d 681 (3rd Cir. 1982), where the Commission noted that "The effort to misreprcsent the nature of ... (an) ingredient is a technique that could easily be applied to advertising "37 The same could beof OTC drug products other than r this one J. said in the present matter. Stouffer s false sodium claims could 37 Thompson Medicn/, 104 FTC at 8:17; Amt';call Home Prodl/(/s Corp" 98 FTC ar 405. STOUFFER FOODS CORPORATION 815 746 Concurring Statement easily be transferred to any nutrient or ingredient in its frozen food products.

Although Stouffer has no history of prior violations before the Commission, that factor alone is insufficient to overcome the factors discussed above. On balance, therefore, we believe that broader order coverage is warranted and that the order should apply to all nutrients and ingredients in Stouffer s frozen food products. CONCURRING STATEMENT OF COMMISSIONER MARY L. AZCUENAGA The Commission today issues a final order and opinion holding that Stouffer Foods Corporation ("Stouffer ) violated Sections 5 and 12 of the Federal Trade Commission Act, 15 use. 45 and 52 ("FTC Act ), by making false and deccptive advertising claims concerning the sodium content of its Lean Cuisine frozen entrees. I concur in the order and, as far as it goes, in the opinion. As the majority properly states, a decision to impose fencingrelief ordinarily rests on consideration of three criteria, although not all three need be present to warrant fencing- in relief. Slip op. at 28. These criteria are: (I) the seriousness and deliberateness of the violation; (2) the transferability of the unlawful conduct to other products; and (3) any history of past violations. Thompson Medical Co. 104 FTC 648 833 and n. 78 (1984), aiid 791 F. 2d 189 (D. Cir. 1986), cert. denied 479 U. S. 1086 (J 987); see also, Kraft, Inc. 114 FTC 40 (1991), aff' 970 F.2d 311 (7th Cir. 1992), cert. denied 113 S.Ct. 1254 (1993) (absence of history of past violations did not make fencing-in improper). The majority has concluded, and 1 agree that Stouffer s violations of the FTC Act are readily transferable and that they are serious and deliberate. Based on these findings, the Commission imposes broad fencing-in relief. I write separately because I believe that it is necessary to address an issue not addressed by my colleagues before finding that the violation was deliberate and, therefore, before imposing fencingrelief. In particular, I think it necessary to weigh the evidence surrounding Stouffer s complaint to an industry self-regulatory organization about advertisements similar to those of Stouffer that were run by one of its competitors and the organization s response to that complaint. I also rely on additional documentary evidence Concurring Statement 118 F.T.c. reflecting Stouffer s intentions regarding sodium claims in its advertising campaign. I The facts cited in the majority opinion (Id. at 30-33) provide tenuous support for the conclusion that Stouffer s violations were deliberate. I need not decide, however, whether the evidence cited by the majority is sufficient, because additional facts in the record persuade me that Stouffer s violations were deliberate. The majority asserts, with little explanation, that Stouffer understood the "delicate nature of the sodium message" and that, despite this delicacy, the company strongly urged TathamlSCG ("Tatham its advertising agency, to "note the strong buzz words" used by competitors, make liberal use of "hot buttons" like "HELPS YOU LIMIT. . . SODIUM" and "FOR SODIUM WATCHERS" and make sure that advertisements for Lean Cuisine "SCREAM HEALTH" by including references to the products ' being "LEAN" on sodium. Slip op. at 32-33 and CX-26. A mere direction to use words and phrases likely to capture a consumer s attention, even in a sensitive context however, does not necessarily warrant a conclusion that any misleading impressions those words and phrases might convey are deliberate. Identifying catchy language to attract the attention of consumers is fundamental to the development of an effective advertisement.

The record contains additional facts not discussed by the majority that support the Commission s finding that Stouffer s violations were deliberate. It is important to address these facts both for the purpose of supporting the Commission s decision to impose fencing-in relief and because Stouffer argues that these same facts show a lack of intention rather than a deliberate effort to mislead. Reply and Answering Br. at 5- I I aso would reverse the conclusion of the Administrative Law Judge that "the failure to disclose adequately the sodium content in milligrams" was not "unfair or deceptive" and that "while respondent failed to disclose adequately that 1 gram equals 1000 milligrams. tl1m fact is immaterial" (ID at 38-39), that is, important to consumers in making their decisions to purchase Stouffer s product Judge Timony found that Stouffer s consumer research in 1988 showed the imponance of information about sodium to consumers, showing them to he .'information hungry" and interested in knowing "the precise cholesterol, fat, and sodium levels." IDF 175. He also found that Stouffer knew that "a frozen entree containing 600 or more milligrams of sodium could turn consumers off." Id. and exhibits cited therein. In my view, this evidence demonstrates that lIlthough some consumers might not know at what level sodium consumption might be harmful, they consider information about sodium content material to their purchasing decisions and some were likely to consider levels of over 600 milligrams unhealthy - For example, Stouffer argues that "finn a transparent distortion of the record, complaint counsel omit the fact that the NAD expressly responded to Stouffer by advising that there was no basis to believe that the use of grams rather than milligrams of sodium was misleading. Id. STOUFFER FOODS CORPORATION 817 746 Concurring Statement Stouffer s directions to its advertising agency did not occur in a vacuum. They followed the rejection, in April 1987, by the National Advertising Division CN AD")' of a complaint drafted by Stouffer in-house counsel and submitted by Stouffer s Marketing Manager Richard Annett, about a competitor s similar claim. Stouffer argued to NAD that a competitor s advertisements for a frozen entree were blatantly misleading to the consuming public" because (like Stouffer s later Lean Cuisine advertisements) the competitor s advertisements stated the product s sodium content in grams rather than milligrams. Stouffer further complained that the rival firm had "intcntionally misrepresented the sodium content in this product." CX-24. NAD declined to act on Stouffer s complaint, finding "no basis to believe" that thc claim "is mislcading to consumers." RX- 12A. NAD asked Stouffer to submit any consumer rcscarch that would support the complaint. IDF 22. Stouffer denied having any such empirical support for its complaint, and it produced none in response to NAD' s invitation. It seems reasonable to assume, however, that Stouffer and its counsel would not have filed such a strongly worded complaint with NAD as a frivolous cxcrcise and that with or without empirical basis, Stouffer must have bccn seriously concerned about the potential effects of the challenged claim. Despite the concern, Mr. Annctt subsequently sent thc mcmorandum to Tatham, instructing it to emphasize the healthful aspccts of Stouffer s product, particularly its "lean" sodium content. CX-26. Stouffer appears to have decided, in light of NAD' s rejection of its complaint, to meet the competition and to use and capitalize on the phrase "Jess than 1 gram of sodium" that the company previously had argued was mislcading. The advertisements at issue here were crcated after the memorandum was conveyed to Tatham and the instructions in thc mcmorandum had been reinforced by discussions between Stouffer and Tatham during the devclopment of the campaign. See, e. CX-40 3 The National Advertising Division of the Council 01 Better Business Bureaus examines and issues decisions on complaints made to it by industry members uguins! their competitors. NAD often is successful in getting advertisers to I'' withdraw or modify claims that it has found unsuhstantiated or otherwise misleading 4 The record shows thelt in the new Lean Cuisine mJn:nisements. Stouffer intended the n less lhdn I granl of sodium" claim, which it had ;lrgucd to )\'AD "'' JS misleading. to have a " disclaimer"" (CX-40J with respect to the health-related sodium claim made in the body of the adveniserm:nb. The disclaimer' presumably was intended to limit the message cor,veyed by the rest of the advenisement, which the Commission h.1s found was a message of " Iow sodium, " Tatham s conference report on one of its discussions of the disclaimer with Mr, Annett records an .1greemcp.t th,!t the disclaimer ('" All Lean Final Order 118 FTC. Stouffer s argument on appeal suggests that it relied on NAD' decision that the "less than 1 gram of sodium" claim was not deceptives Not only is reliance on NAD' s response misplaced, but the fact that Stouffer had considered the competitor s claim sufficiently misleading to challenge it with NAD tends to show that Stouffer was on notice that, regardless of NAD' s decision, a "significant minority of reasonable consumers (Clifdale 103 FTC at 164-66) might well take a misleading "low sodium" claim from the competitor s advertisement and, more importantly, from its own advertisements for Lean Cuisine.

On the basis of the evidence discussed in the Commission opinion and this separate statement, I find that Stouffer s violation was deliberate and, therefore, that the fencing-in relief is appropriate. FINAL ORDER This matter has been heard by the Commission upon the appeals of respondent Stouffer Foods Corporation and complaint counsel and upon briefs and oral argument in support of and in opposition to the appeals. For the reasons stated in the accompanying Opinion, the Commission has determined to affirm the Initial Decision of the Administrative Law Judge, except as otherwise noted, and enter the following order. Accordingly, It is ordered That respondent Stouffer Foods Corporation, a corporation, its successors and assigns, and its officers, representatives, agents, and employees, directly or through any corporation subsidiary, division or other device, in connection with the adver- Cuisine entrees have been reformulated to contain less than 1 gram (1,000 mg.) of sodium ) would be place(dJ in mouse type:' Id. The deliberate decision to put the explanatory material in " mouse type suggests an intention to undemline the disclaimer s effectiveness and to leave virtually intact the overall message conveyed by the advertisement of low sodium content. This intention is further strengthened by the language in the report stating the additional agreement to "(pjbcc in bold type ' Less than J gram of sodium.''' 5 Although the Commission often agrees with the decisions of industry self-regulatory organizations such as AD regarding whether' particular claims are misleading. the decisions of such organizations are not controlling in cases before the Commission. 6 In both Kmji J 14 FTC at 140. and Thollf)lol1 104 FTC at 8J4-35. the Commission relied on the fact that the companies had received wurning from others regarding the potential that the advertisement at issue might not be true Here, the v.arning originated within the respondent company and should be given at leelst as much weight by the Commission. if not more STOUFFER FOODS CORPORATION 819 Final Order tising, labeling, offering for sale, sale, or distribution of any frozen food product in or affecting commerce, as "commerce" is defined in the Federal Trade Commission Act, do forthwith cease and desist from misrepresenting in any manner, directly or by implication through numerical or descriptive terms or units of measurement, or by any other means, the existence or amount of sodium or salt or any other nutrient or ingredient in any such product. Provided, however that if any representation covered by this part either directly or by implication conveys any nutrient content claim defined (for purposes of labeling) by any regulation promulgated by the United States Food and Drug Administration or, if applicable, by the United States Department of Agriculture, compliance with this part shall be governed by the qualifying amount for such term as set forth in that regulation. Provided, further, however, that nothing in this part shall prohibit any representation as to thc amount of sodium or salt or any other nutricnt or ingredient in any frozen food product if such representation is specifically permitted in labeling, for the serving size advertised or promotcd for such product, by any regulation promulgated by the United States Food and Drug Administration or, if applicable, by the United States Department of Agriculture. II.

It isfurther ordered That respondent Stouffer Foods Corporation its successors and assigns shall, for three (3) years after the date of the last dissemination of the representation to which they pertain maintain and upon request make available to the Federal Trade Commission for inspcction and copying all advertisements containing any representation covered by Part I of this order. It is further ordered That respondent Stouffer Foods Corporation shall distribute a copy of this order to its operating divisions, to each of its managerial employees, and to each of its officers, agents, representatives, or employees engaged in the preparation or placemcnt of advertising or other material covered by this order and shall secure from each such person a signed statement acknowledging receipt of this order.

Final Order 118 FTC. IV.

It is further ordered That respondent Stouffer Foods Corporation shall notify the Commission at least thirty (30) days prior to any proposed change in the corporation such as the dissolution, assignment, or sale resulting in the emergence of a successor corporation the creation or dissolution of subsidiaries, or any other change in the corporation which may affect compliance obligations arising out of this order.

It isjilrther ordered That respondent Stouffer Foods Corporation shall, within sixty (60) days after service upon it of this order and at such other times as the Commission may require, file with thc Commission a report, in writing, setting forth in detail the manner and form in which it has complied with the requirements of this order.

By the Commission.

Prior to leaving the Commissior.. former Commissioner Owen and tmmcr Commissioner Yao each regi !ercc! a vote in the affinnativc forthe Final Order and the Opinion of the Commission in this matter TRANS UNION CORPORATION 821 Summar Decision