Tobacco Documents Online

THE ROPER ORGANIZATION INC. BURNS W. ROPER President November 11, 1974 Mr. William Kloepfer Tobacco Institute 1776 K Street, N.W., Washington, D.C. 20006 Dear Bill, Since seeing you in Washington, Shirley and I have discussed further the matter of determining how many teenagers smoke. While we have refined our thinking somewhat, we come up with exactly the same alternatives that you and I discussed--namely: 1. Taking the position with the Cancer Society and others, that you have no knowledge of the incidence of smoking among the young because you have scrupulously avoided researching them in order to avoid any appearances of trying to exploit youth--but that you would be happy to enter into a cooperative study with the Cancer Society, Heart Association or whoever, to conduct a national in- school study along the lines of the San Mateo study. 2. Replicating insofar as possible the telephone study done by Chilton, though possibly improving on it in question wording, etc. (assuming we can learn exactly how their's was done). 3. Conducting a personal interview study, the main thrust of which is an exploration of youth activities and atti- tudes but which in the process, also determines the extent to which kids smoke. The challenge to the Cancer Society, Heart Association, etc. would probably not be picked up by them, but the mere fact of making the challenge might get you off the hook. If the challenge was accepted, of course, you would then be in the position of having to publish the results and that might or might not be good from your point of view. CONMENnAL' MINN"OTA TOBACCO LITIGATION ONE PARK AVENUE • NEW YORK. NEW YORK 10016 - 212 OREGON 9.3523 2 TIMN 262415

Mr. William Kloepfer November 11, 1974 Page 2 Replicating the Chilton study would serve to satisfy your curiosity as to whether the reported results of the Chilton study are accurate, but it probably wouldn't do much else. If it confirmed the Chilton results, it would be nothing new. If it disputed the thilton results, then it would become a battle of which survey was more accurate, and your survey-- our survey--might well be the suspect survey, even though I think our organization has generally better credibility than Chilton does. I say that our survey would likely be the suspect one because you represent the forces of evil and the National Clearinghouse represents the forces of virtue. This, therefore, leads us to favor the third approach, because that would represent quite different methodology and methodology which would widely be recognized as superior to Chilton's telephone approach. As Shirley and I have considered it, what we would suggest would be a personal interview survey using a strict national probability sample, with up to three call backs to obtain designated respondents, to be conducted among a cross section of boys and girls of, let's say, 12-17 years of age (or whatever ages you want). We would design and administer about a fifteen minute questionnaire which would be "kid-oriented". It might deal with such things as clothing styles, pop singers, recreational activities, etc. About the thirteenth minute of the fifteen minute interview would be devoted to a list of activities about which two questions might be asked. The list of activities might include such things as ridden on a motorcycle, participated in a student protest, smoked marijuana, drunk beer, smoked cigarettes, drunk hard liquor, been arrested, driven a car, stayed out all night, run away from home, etc. These activities would be listed on a card, each one preceded by a large capital letter (A, B, C, D, etc.). The card would be given to kids about the thirteenth minute and they would be asked a question along the following lines (wording subject to change and hope- fully improvement): "Here is a list of things that some young people do, even though their parents don't look on them with favor. Would you go down that list and for each item tell me whether you have ever done it. Just call off the letter of the item and then tell me yes or no". After they had answered the question, they would be asked a second question along these lines, "Now for each item you have ever done, would you tell me whether you have happened to do it in the last two weeks. Just call off the letter of each item you have done in the last two weeks." Ourthesis is that almost everyone who has done a given item in the last two weeks can be regarded as a "regular". It is, of course, possible that in a few cases someone will have done an item in the last two weeks and it will be the first and only time either he or she has done it. But that would have to be a very small percentage of the cases. By stressing the letters with which the items were designated and asking people to call off the letter using yes or no, we would avoid the problem of kids having to say out loud that they had or had not done the item if they were in the presence of their parents. CONFIDENTIAL: MINNESOTA TOBACCO LITIGATION 3 TIMN 262416

Mr. William Kloepfer November 11, 1974 Page 3 Moreover, we would further reduce the significance of these questions, both in the kid's mind and in his parents' minds, assuming he was in earshot of his parents, by preceding them with twelve minutes of youth activities questions and following them with two more minutes of comparatively innocuous questions and questions about demographic characteristics. Thus, to the kids and any eavesdropping parents, these two questions would be incidental to the study. From your point of view, of course, they would be the entire study--they and the demographic analysis that would make them meaningful. Conducting the study in this fashion, by personal interview with several call backs, would be superior to the Chilton methodologies and would avoid any questions about non-telephone homes, unlisted phones, not at home biases, etc. While Chilton started out using a 5,000 sample, they have now dropped down to a 2,500 sample. it is our feeling that a sample of 1,500 along the lines that we are suggesting would stack up competitively to the Chilton sample. The personal interviews and call backs would more than offset the larger sample size. Moreover, 1,500 has become the standard by which surveys are judged, since that is what Gallup and Harris traditionally use. Finally, a 1,500 sample would give us at least 100 kids of each sex in each single year age group. A study along these lines would cost $37,500, plus or minus 10%. This is a high cost relative to the regular biennial Tobacco Institute Study. That is more like a forty-five minute questionnaire and employs a 2,500 sample and yet costs less than twice as much. The cost disparity results from the fact that probability samples with repeated call backs are expensive indeed. As a rule we do not recommend a probability sample. But in view of the audience we are trying to reach on the one hand and the fact that the study results may be publicized and hence should be "unassailable" on the other hand, we would recommend this more expensive sampling method in this instance. Questions, reactions, comments will be gratefully received. Cordial , Burns W. Roper CONFIDENTIAL: MINNESOTA TOBACCO LITIGATION TIMN 262417