“ '. g . , ‘ ' J V V. . -Dpi‘EQiflY‘f‘ -. "h :5. 7:‘-°—‘ -' -'- H‘. «who/Ed fig} gym“ I??? ’: .Iiv . 3 ’5‘5’“ I.” . .t'g 7-._,._~. ' ('6‘ u nuk‘fim" Y V 1‘1)?” {:53 ‘M .7. u 337:}? 4.1:, 4-“! W ($13k '7 .;v.~.‘:: '3'? #334,..{14’ Lgi;?‘;EI-1'v 1'33." 9’: .41.. .' 33".“h “' ”if” .h)‘ ‘L’.’ (vai‘ .‘Sg': kg)???“ '"\ v. V 44%”. W R I V. 4‘ 4 .n" -~- ‘I.. . ”“1. {W I" ;?‘5.'4;::‘.: :{17- f t.‘ ‘y{. .1 $3351- \ o ,.,. ix ‘2," air-”rm" «.3... . .\ ('4': 5-1::2‘. . 9-1;:th -~-‘’Qr“lqu2-3;y ' l .‘t .- JUH'. T: 3.44.; ”4"" 92/? {:7 ”£12?" 'l/'. /:”/;:./-':y' a“? "$54); , «1.. 2‘ ‘4 r. .13.. -.‘;'.'.n.* ‘\:q . I...” A'J'f" 4. n 4 45.1,: fir \ ‘ . f.‘ ‘ "fl'l‘ J .- h I ~’~.+. '\_'! 7;. ‘ ';_., "¢ . #42353’. '41.‘ .33.. . - a ' *7 {a “ ' t 9:1; . 4:... 1' "'u'n‘ .~o ‘ ." :Ju'fin M" . u '4'. .. ’1 ‘II‘ "4’ “Hr-“1.51 -' 4,1";- o!- . '54 J . ~ 4'? 4 . 1.414% ’1’"- , g I”! 'IOI-ull‘l‘l‘o .4I' ru') “0'.- .'"‘ " Mimi 5'41”] 7‘1613791939 MICHIGAN ST T UNIVERSITY llBRARIES W l W _ll\l\ll l‘l‘fllll‘l‘l‘lilll 2| n 3 1293 00551 4553 LIBRARY 1 Michigan State University This is to certify that the thesis entitled HEALTH LOCUS OF CONTROL, ATTITUDES AND INTENTION TO REDUCE CANCER RISK presented by Sally Ruth Anger. has been accepted towards fulfillment of the requirements for M. S. degree in Human Nutrition §kgzg :1 2/. gaging Major professor Date June 20, 1988 0-7639 MS U is an Affirmative Action/Equal Opportunin Institution Fr“ - _ - — 4.2—- _m MSU LETURNING MATERIAL; Place in book drop to LJBRARJES remove this checkout from A“... your record. FINES will be charged if book is .«»v«~' returned after the date a Stamped below, JUL 2 2 21104 3EP._2,.0,2001 Eli'é 3 (1] HEALTH LOCUS OF CONTROL, ATTITUDES AND INTENTION TO REDUCE CANCER RISK BY Sally Ruth Anger A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Food Science and Human Nutrition 1988 .3 (1” if) K) ABSTRACT Health Locus of Control, Attitudes and Intention to Reduce Cancer Risk BY Sally R. Anger Health Locus of Control (HLC) and Health value (RV) were compared to change in nutritional knowledge, attitude and behavioral intentions from 94 people enrolled in worksite wellness programs in Hichigan before and after viewing a nutrition/cancer program. A nutrition/cancer knowledge test (NCKT) was developed, validated, and pilot tested for reliability (KR-20) and item analysis. Previously validated instruments were used to assess HLC, RV and control over eating habits. Intention to change behavior and attitude-toward-nutrition were assessed following the elements of the theory of reasoned action developed by Fishbein and Ajzen (1980). Correlational statistics revealed a significant increase in nutritional knowledge from pre- to post-test. No differences were found between internals and non-internals in knowledge gained or attitude change. A cancer-specific attitude scale accounted for more variation in behavioral intention than the HHLC or attitude-toward-nutrition in this study. ACKNOWLEDGEMENTS I would like to express my gratitude to Dr. Maurice Bennink whose guidance and understanding as my major advisor enabled me to pursue my career goals. warm thanks go to my thesis advisor, Dr. Judith Anderson, who provided invaluable encouragement and never-ending belief in my capabilities. Drs. Hoerr and Ignatovich provided guidance and insight as committee members. Thanks are due to Dr. Richard Houang for his statistical guidance. My studies would not have been possible without the financial support of the Department of Food Science and Human Nutrition and the Cooperative Extension Service. I greatly appreciate the assistance of Janet Fabiano of the American Cancer Society (Ingham Division). I would also like to thank Barb Hutch for her sense of humor, non-judgemental listening and friendship. Thanks go to my dad and Kathy who invariably provide support and encouragement. Special thanks are due to my mom who is always my number one fan in whatever guest I undertake. ii TABLE OF CONTENTS LIST OF TABLES..... ......... . ........... ... ...... .. .. vi LIST OF FIGURES..... . ..... .. .. ............ . ........ viii INTRODUCTION......................... ...... .... ....... 1 Research Rationale............................... 1 Objectives.................... ......... .......... 2 PERTINENT LITERATURE AND THEORY.... ...... . ...... . ..... 4 Nutrition and Cancer ...... ........ ...... ......... 4 Locus of Control Theory.......................... 5 Multidimensional Health Locus of Control Scale... 9 Audience Segmentation............................ 14 Fishbein's Theory of Reasoned Action............. 15 Integration of Locus of Control and Reasoned Action18 Conceptual Research Hodel......................... 20 Hypothesis........................................ 23 METHODS Human Use Approval........ ..... . ........ . ..... .... 25 Subjects.......................................... 25 Research Strategy................................. 25 Instrumentation................................... 26 Nutrition/Cancer Knowledge Test .............. 26 Test Development.......................... 26 Pilot Studies............................. 30 Item Analysis........................ ..... 30 Reliability and Validity.................. 32 Multidimensional Health Locus of Control Scale 33 Cancer Locus of Control Test.................. 33 Health value Survey........................... 33 Intention To Change Behavior Questions........ 33 Attitude Toward Nutrition Questions........... 34 Subjective Norm Questions..................... 34 Data Collection................................... 35 Data Analysis..................................... 37 Locus of Control and Health value............. 37 Elements of Theory of Reasoned Action......... 39 iii RESULTSOCODCO......CCICIOIOIOO......... ......... O ..... Population Characteristics......... Group 1 and 2 Characteristics................. Internal versus Non-internal Characteristics.. Worksite Characteristics...................... Reliability and Distribution of Scores............ Nutrition/Cancer Knowledge Test(NCKT)......... Hulitdimensional Health Locus of Control Test Cancer Locus of Control Scale (CALC).......... Attitude Scales..... OOOOOOOOOOOOOOOO Social Norm and Intent........................ Correlations Between Scales....................... HHLC and CALC................................. Between Variables................... NCKT Scores............ Attitudes.............. Social Norm............ Relationships Effect of Effect of Effect of Effect of Effect of Variables variables variables variables variables on on on on on Intent. Cancer Behavior Intent. Summary of Results.............. ....... ........... DISCUSSIONOOOOOOOOOOOO Introduction ........... .. NCKT Scores in Pre- versus Post-test Groups... Effect of Program and Internality on Attitude. Effect of Internality on NCKT Scores.......... Effects of the Elements of the Fishbein Equation on Intention to Change Dietary Behavior... Strengths and Limitations of the Study............ Instrument Reliability and validity........... Internal validity Issues...................... History................................... Haturation................. Testing................................... Selection................................. External validity Issues...................... Interaction of Selection and Treatment.... Interaction of History and Treatment...... SUHMARY AND IMPLICATIONS.... LIST OF REFERENCESOO......OOOOOOOOO.......OOOOOOOOOOOO APPENDICES A. Dietary Recommendations to Americans on Diet/Nutrition/Cancer—-l980-1985............. B. Consent Form................................. C. Final Version NCKT... iv 41 73 74 74 74 75 77 77 78 79 80 81 81 82 82 84 85 88 94 95 96 D. E. F. G. I. J. K. Facilitator's Guide of the American cancer Society Taking Control Program...............100 Content Review - ACS Volunteers..............101 Checklist for Hultiple Choice Items..........104 Content Review - HSU Nutrition Faculty.......105 Scoring Office Results - First Pilot Test....110 Scoring Office Results - Second Pilot Test...116 Questionnaire on Health Beliefs..............122 values, Attitudes and Intention to Change Behavior Survey..............................126 Scoring Office Results - Final version NCKT..l30 LIST OF TABLES Table Page 1 Alpha reliabilities of multidimensional health locus of control (MHLC) scales.......... 10 2 Mean scores for HHLC scales summarized across types Of sobjectSCOOOO0.0.0.0009...00.0.00...O. 11 3 Test specifications....... ..................... 28 4 Summary of demographic variables for total sample..................... .............. 42 5 Summary of background variables for total sample ..... ....... ..... . ................. 43 6 Summary of demographic variables for groups 1 and 2..... ...... .............. ........ 45 7 Background variables for groups 1 and 2. ....... 47 8 Distribution of internality and health value between groups. .... . ..... . .. ....... 49 9 Demographic variables for internals and non-internals.0.0.0....IIOOOOOOOOOOOOOOOOOOOOO. so 10 Background variables for internals and non-Internals.O......OIOIOOIOOOOIOIIOOOOO ...... S3 11 Worksite characteristics......... ...... ........ 55 12 Mean scores and standard deviations for MHLC... 56 13 Mean scores and standard deviations for CALC... 56 14 Correlations between the MHLC and CALC......... 58 15 Two-way ANOVA of NCKT to test for the effects of group and internality....... ..... ... 59 vi 16 17 18 19 20 21 22 23 24 25 26 27 28 Oneway ANOVA of NCKT by internals with high Hv versus others .......................... ' Oneway ANOVA of NCKT by internals with high Hv versus others, pre- and post-test...... Results of regression analysis of the dependent variable: intent to change diet........ ..... .. Results of regression analysis of the dependent variable intent to change diet for internal with low health value and non-internal subjects .................... ...... Results of regression analysis of the dependent variable intent to change diet for internal subjects with low health value and non-internal subjects taking the pre—attitude test..... ..... Results of regression analysis on C81 in internal subjects with high HV ............ ..... Regression analysis on C8! in internal subjects with low Hv and non-internals.................. Regression analysis of C81 in internal subjects with high health value who took the pre-attitude test ....... . .................. .... Regression analysis of C31 in internal subjects with high health value who took the post-attitude test.............. ............... Regression analysis of C81 in internal subjects with low health value and non-internals who took the pre-attitude test............. ........ Regression analysis of C81 in internal subjects with low health value and non-internals who tOOk the pOSt‘attItuae tBSt. o o o o o o o o o oooooooooo Summary of effects on intent to change diet.... Summary of effects on intent to change dietary behavior related to cancer............. vii 61 63 65 66 67 67 68 69 69 69 71 Figure 1 Elements of Ajzen and Fishbein's theory........ 2 Model showing the integration of Fishbein and Rotter theories ............................ 3 Steps for test development ..................... 4 Experimental design ............................ S NCKT by group.. ......... ................. ...... 6 NCKT scores by group for internals with high LIST OF FIGURES health value (squares) versus internals with low health value and non—internals (triangles). viii Page 15 22 27 38 60 62 INTRODUCTION Research Rationale A demonstrated need exists for increased public education on possible methods of cancer risk reduction. The 1984 Cancer Prevention Awareness Survey published by the National Institutes of Health (1984) documented the need for programs to increase public awareness about cancer prevention. About half of the 1,876 respondents to this survey believe that ”there is not much a person can do to prevent cancer” and that "everything causes cancer”. The obvious first step toward dietary change is increasing the awareness that dietary changes may decrease the risk of cancer, as well as increasing the awareness of specific changes needed. The American Cancer Society's (AC8) ”Taking Control” program (1985) is designed to inform people of cancer protective and risk factors. The cancer risks which might be lowered by a change in behavior are stressed in the ”Taking Control" program. One of the goals of the ACS is to: provide an introduction to a healthy, enjoyable lifestyle that may reduce one's risk of developing cancer...The Taking Control program helps people think about health in terms of their overall lifestyles. Taking Control is positive, optimistic, and action oriented. (American Cancer Society, 1985) Research on the factors associated with preventive health behavior has uncovered a variety of determinants of preventive health practices. These include social and demographic variables, previous use of health services, health attitudes and beliefs (McCusker, 1979). Findings from a wide range of studies suggest more adaptability in those persons who believe they have control over their own lives compared to those who believe they have little or no control over their lives. These findings have significant implications with respect to preventive health. Objectives The primary question investigated by this research was whether the health locus of control and health value of viewers of the ”Taking Control" program influenced their acquisition of knowledge and change in attitude and/or intentions to adopt a healthier diet after participating in the program. The answer to this question will provide useful information from which to make recommendations for future program development and accurate targeting of health messages to specific groups within the general public. There were three main objectives to this study: (1) to determine if there was an increase in viewers' nutrition/cancer knowledge following the ”Taking Control” program; (2) to determine if the "Taking Control" program (3) to determine if health locus of control and health value related to change in knowledge, attitudes or intention to change behavior prior to and upon completion of the ”Taking Control” program. PERTINENT LITERATURE AND THEORY Nutrition agg,Cancer The American Cancer Society's ”Taking Control" program (1985) encourages Americans to eat less fat, more fiber, more fresh fruits and vegetables, more whole grains and cereals, less salt-cured, smoked and nitrite-cured foods, and to drink less alcohol than most Americans now do. The ACS also recommends two non-dietary guidelines: to stop cigarette smoking and to avoid over-exposure to the sun. Research in cancer and nutrition suggests that populations which, relative to the 0.8., eat less fatty foods, more fiber-rich foods and more vegetables and fruits tend to have lower rates of some of the more common cancers such as those of the colon, breast, prostrate and uterus (Lanza, 1987). 0011 and Peto (1981) have estimated that diet is a contributing factor for 35% of cancers.‘ Based on the bulk of research in the field, the National Cancer Institute (NCI) (Greenwald, 1986), the American Cancer Society (AC8) (1987) and the National Academy of Sciences (NAS) (1982) have all proposed public dietary recommendations. These have been summarized by Michael v. Pariza (1986) and are included in Appendix 1. These recommendations are based on extensive research conducted in recent decades. While some scientists disagree with one or more of the recommendations, the guidelines reflect a general consensus in the nutrition community. Moreover, when the general healthy public follows the dietary recommendations, they do not experience any additional risk (Brown and Cooke, 1980). Egg g Q£_Contrgl Theory Locus of control is derived from Rotter's Social Learning Theory (1954) which states that the potential of a specific behavior (behavior potential, BP) occurring in a given situation is a function of expectancy (E) that the behavior will lead to a particular reinforcement in that situation, and the value of that reinforcement (RV), e.g. BP = f(E & RV). Of the components of this model, expectancy has been the subject of the most research. The concept of locus of control was developed because of the observation by Rotter et a1 (1975) that increases or decreases in expectancies following reinforcement appeared to vary systematically, depending on the situation and also as a consistent trait of the particular person who was receiving the reinforcement. Locus of control evolved as a variable that might correct or refine the social learning theory prediction of how reinforcements change expectancies. In locus-of—control terms, those individuals who believe that their health is influenced mainly by their own behavior are "internals", with ”externals" being those who believe that their health is determined largely by fate, chance, other people or events beyond their control. These beliefs might influence the outcome of health intervention programs such as smoking cessation and weight control. In general, people with high scores on internality appear likely to adopt behaviors that improve their physical health. For example, studies using Rotter's Internal/External Scale have shown that nonsmokers are more likely to be internals than smokers (Straits and Sechrest, 1963; James, Voodruff and Werner, 1965). Straits and Sechrest (1963) found from internal- external questionnaires given to male college students that smokers were more ”chance-oriented” (p<.10) than non- smokers. James, Ubodruff and Werner (1965) found in a study with undergraduate students that both male and female smokers scored significantly higher in externality than nonsmokers. They found that male smokers who reported that they had stopped smoking following the Surgeon General's report were significantly less external than those who continued smoking. There was no difference between females who continued and those who stopped smoking. Other studies suggest that individuals who were not smokers or individuals who were able to stop smoking were more internal than individuals who smoked (Coan, 1973; Williams, 1973). These results have not always been replicated (Danahar, 1977) and are only correlational; there has been no proven cause and effect relationship. O'Bryan (1972) studied locus of control and weight loss, and found that overweight people were significantly more external than those of normal weight. In a study of negative versus positive covert reinforcement, Hanna and Marston (1972) found that Rotter's Internal-External Scale (I-E scale) scores were correlated with weight loss at follow-up (r = -0.57, p<.05) in their control group. The weight of controls before treatment correlated negatively (r=-0.39, p<.05) with internality on Rotter's I-E scale. Manno and Harston found that externally oriented subjects weighed more at pre- treatment, yet lost less at follow-up. No relation was found between I—E scores and weight loss in the experimental groups (those receiving various forms of reinforcment for weight reduction). Balch and Ross (1975) found significant correlations between internal Rotter I-E scores and both completion and success in a weight loss program. There are several factors to be considered when reviewing the locus of control literature. Rotter (1975) warns against assuming that it is ”good” to be internal and "bad” to be external, a concept which could easily be interpreted from the studies cited above. It may be better for people who are trying to cope with weakening abilities, such as the aged (Rodin, 1986) or cancer patients (Marks, Richardson, Graham and Levine, 1986) to have a feeling that they can control what happens to them, but there must be a limit on personal control. Rotter's (1975) early hypothesis was that locus of control would have a curvilinear relationship to adjustment. Rotter (1975) has emphasized the fact that locus of control is just one component of a complex combination of factors used to predict behavior. Locus of control relates to expectancy of reinforcement, and Rotter's social learning theory equally emphasizes the situation in which the reinforcement occurs as well as the value of the reinforcement. Rotter (1975) urged researchers to include a measure of the value of the reinforcement as well as a situation-specific measure of locus of control. The studies mentioned above (James et a1, 1965; Coan, 1973; Williams, 1973; O'Bryan, 1972; Manno and Marston, 1972; Balch and Ross, 1975) have used Rotter's general I-E Scale without a locus of control scale specific to the situation being researched. With health-related measures of value of reinforcement and expectancy, the theory has been that an individual's health behavior could be explained by the interaction of these two constructs (Wallston and Wallston, 1981). The locus of control theory would predict that those subjects who value their health highly AND expect their behavior to enhance their health would be more likely to adopt healthy behaviors (Hallston and . wallston, 1981). In studies of health value in relation to health behavior, Kristiansen (1985) has shown that health value, as measured by Rokeach's (1973) value survey with the additional value of health (Wallston, 1976), is related to healthy preventive actions of adults. Multidimensional Health Locus of Control Scale Rotter (1975) recommended that specific measures of locus of control be developed for specific applications. One such scale is the Multidimensional Health Locus of Control (MHLC) Scale developed by Wallston, Wallston, Kaplan and Maides (1976). In the MHLC, health locus of control (HLC) is divided into three dimensions: internal health locus of control (IHLC), powerful others health locus of control (PHLC) and chance health locus of control (CHLC). Those individuals with an IHLC tend to feel responsible for their own health. Those with a PHLC tend to feel that their health is in the hands of other people such as their physician, allied health professionals, or family member(s). Those with CHLC tend to feel that no matter what they or others do, their health is influenced largely by chance or accident. Individuals can also have various combinations of high/low scores on the three dimensions, for example one could obtain a high score on internal and powerful others, with a low score on chance. The extensive testing of the MHLC for reliability has been summarized by Wallston and Wallston (1984) from several studies of college students and adults (Table 1). In addition, normative data on the MHLC has been collected and summarized by Wallston and Wallston (1984) (Table 2). 10 no. so. 3. a an a... 5.2:. on. no. 3. < «a 2:39:- 33.... R. 2. n... m .- < as. on. an. a on. a». 3. < 2 u on. :1 no. a a < 2. as. an. m an. on. «a. < .... no. so. so. a a < as. no. on. A a , . 3. vs. on. < co . 3—..an moo—30 6.5.: .310. .82.. 8.8. z ...—.3:— 3 ES... :35 8:25 .3330.— . . .._.__s.=s_ l I 3mm. cot... .63 sort 8.68 6.53: .2360 go 263 .38: 35.265332 so 8.53.53— 23? .a .2 a8. 11 Table 2. Mean scores for MHLC scales summarized across types of subjects.* Mean Scores Powerful Internal Chance Others Sample N (IHLC) (CHLC) (PHLC) Chronic patients 609 25.78 17.64 22.54 College students 749 26.68 16.72 17.87 Healthy adults 1287 25.55 16.21 19.16 Persons engaged in 720 27.38 15.52 18.44 preventive health behaviors * from Wallston and Wallston (1981). 12 The work of several investigators on the MHLC deserves special mention here. Winefield (1982) found the internal and powerful others subscales to be reliable over 7 months. The chance subscale had little stability over time in her study. In addition, Winefield found that the powerful others subscale rose with age, decreasing social status and acute illness. O'Looney and Barrett (1983) found a sex difference in the factor structure of the MHLC in a group of British university students. In their study, the male sample yielded only two factors while the female samples yielded the expected three scale factors. The male factors were defined by a) the internal and chance scale items together and b) by the powerful others scale items. Jordan-Marsh and Neutra (1985) found a change in HLC from admission to discharge from a residential lifestyle change program. Scores tended more towards internality following the program. There were no significant correlations between HLC scores and changes in physical parameters such as weight, blood pressure, low density lipoproteins, or triglycerides. The differences in locus of control due to sex, social status, and program participation are not present in every study. Wallston and Wallston (1984) encourage researchers to view the health locus of control construct not as an unchanging personality variable, but more as a characteristic of an individual at a given point in time. Research with the MHLC has shown that maintenance of smoking cessation was enhanced by an internal locus of 13 control (Kaplan and Cowles, 1978; Shipley, 1981; Horowitz, et a1, 1985). Research by Kaplan and Cowles (1978) indicated that those subjects with internal HLC and a high value of health smoked significantly fewer cigarettes (p<.05) following a smoking cessation program than those subjects with internal HLC-low health value, external HLC- high health value, and external HLC-low health value. Shipley (1981) found subjects who scored highly on the MHLC were more often abstinent following smoking cessation treatment than those subjects with low scores on internality (47‘ vs. 17% at 3 months, p<.05; 40% vs. 13% at 6 months, p<.10). Likewise, subjects with low scores on chance MHLC were more often abstinent than high scorers (47‘ vs. 17\ at 3 months, p<.10; 45$ vw. 9‘ at 6 months, p<.05). In a study of ex—smokers, recidivists and continuing smokers, Horowitz et a1 (1985) found recidivists to be significantly (p<.05) less internal than the other groups and had become less internally controlled between the treatment and a one-year follow-up (t (53) = 3.93, p<.001). Although the powerful others scale did not show significant differences for the group (F(2,152) 2.17, p<.12), a trend was seen for recidivists to have decreased scores on the powerful others MHLC (t (53) = 1.85, p<.10) between treatment and follow-up. Coelho (1985) examined the psychometric properties of the MHLC with cigarette smokers and determined that cigarette smokers were more appropriately divided into a 14 bidimensional scale of IHLC and PHLC measures, since the CHLC dimension was absent in his study. Audience Segmention Research into audience knowledge and attitudes has helped increase the efficiency of public education programs on nutrition (Brown and Cooke, 1980) and cancer risk reduction (Novelli and Novelli, 1984). Audience segmentation is based on the idea that health consumers have important, measurable differences related to health behavior (Novelli and Novelli, 1984). A study by Wallston et a1 (1976) with overweight subjects found that the external HLC group lost more weight in an externally oriented, group program and the internal HLC group lost more weight in an internally oriented, self-directed program. Those matched with the appropriate program reported greater satisfaction than those who were not. In general, internals are more successful in treatments using a self-reward or self- motivation approach. Externals are more successful in programs with reinforcements from an outside source, such as social reinforcement in the form of verbal feedback from program presenters (Wallston et al, 1976). Fishggig'g Theory g;_Re§songg Action The theory of reasoned action developed by Ajzen and Fishbein (1980) is based on the assumption that humans are rational and they behave according to a systematic use of the information available to them (Figure 1). According to the theory, a person's behavioral intention is a .13 uow>qsoa c« on :98 Haauu< uoasason mucosa on couucoucu . .H enemas caonnmam use coufi< mo mucosofim .zuoonu m. muonuc mo oocosmuca vusaoouom uow>a£on a: own no you a « oesuuuu< 16 function of two basic factors, their personal attitude toward the behavior and the subjective norm. An attitude toward a behavior is a function of the individual's perception and evaluation of the consequences of performing a certain behavior. The subjective norm is how a person perceives the social pressures to perform or not perform certain behaviors. Social pressures are the individual's perception of how most people who are important to him or her think that the individual should behave. Fishbein's theory can be written as: BI = (A )w + (SN) w B 1 2 where BI is the intention to perform behavior B; A is the attitude toward performing the behavior; SN is theB subjective norm; and w and w are empirical weights measuring the relativelimportince of the attitude and subjective norm components in affecting behavioral intention. The relative importance of these components will vary according to the individual, the situation, and the behavior of interest. According to Ajzen and Fishbein (1980) factors such as personality traits or demographic variables which are related to behavior are considered "external factors” and do not make up an integral part of their theory. Ajzen and Fishbein recognize the importance of these external variables, but believe that an external variable will influence behavior only to the extent that it influences the beliefs a person holds or the relative 17 importance he or she attaches to attitudes or social norms. Exposure to new information (change in knowledge) is an important base on which to alter attitudes or beliefs and ultimately to influence behavior (Fishbein and Ajzen, 1975). However, knowledge alone is not necessarily sufficient to illicit a change in behavior (Sims, 1981). In a review of attitude assessment in nutrition research, Sims (1981) summarizes this idea: Most of us are equally convinced that knowledge alone ...probably is not enough to change behavior. Several people have reasoned that attitude change must mediate the process, i.e., a person not only must have knowledge of certain information, but also must feel it is important (attitude) and probable (belief). In research using the Fishbein Theory of Reasoned Action, Shepherd and Stockley (1987) found that personal attitudes were good predictors of fat consumption (r = 0.68, p<.001) and more important than perceived social pressure from other persons. Nutrition knowledge was not found to relate to attitudes or to consumption of high fat foods. Dalton, Linke and Simko (1986) used the Fishbein Model to investigate intended versus actual worksite food choices. They found a strong association between consistency of food choice (choosing the food the subject intended to choose) and positive attitudes toward the choice, as measured by sensory appeal, health value, and expediency. Subjects with less positive attitudes were less likely to make choices consistent with their intended food choice. others' influence (perceived social norm) 18 was also associated with consistent food choice (F=61.01 p<.01). Pender and Pender (1986) used the theory of reasoned action to study the relationships among attitudes, subjective norms and intentions to engage in healthy behaviors. Study results showed that attitudes were useful in explaining intentions to exercise regularly, maintain recommended weight and avoid stressful life situations. Social norms influenced only the intention to engage in regular exercise. Godin, Cox and Shephard (1983) found that attitude was correlated with intentions to exercise regularly, but subjective norm was not. In their study, 35% of the variance in intentions to exercise was explained by a combination of two variables: the two components of the Fishbein model (attitude and social norm) and an index of current physical activity. Integration gf_§ggg§_gg,CQntrol gag Reasoned Action Research with the MHLC shows that by itself the MHLC may not be a good predictor of preventive health behaviors (McCusker and Morrow, 1979; Winefield, 1982). In a review of several surveys with different samples of individuals Wallston and Wallston (1982) report that a11 failed to find significant correlations between health locus of control and a wide range of health behaviors. They provided several explanations for the inability of the MHLC Scale to predict preventive health behaviors. First, 19 they suggested that preventive behaviors are multidetermined, and that it is simplistic to expect that a single construct will predict much of the variance in individual health behaviors (Wallston and Wallston, 1981). They suggested that other variables, such as specific beliefs about the the behavior in question, may be equally or more important in predicting behavior. In a study incorporating both the locus of control and Fishbein theories, Saltzer (1978) found that for subjects who valued health and/or physical appearance highly and were described as internal or external using a locus of control scale specific to weight loss, personal attitude was the stronger component of behavior intentions for the internals (p<.005), and the subjective norm was the stronger component for externals (p<.005). However, Saltzer (1981) was unable to replicate her initial findings. Possible explanantions given for the inablility to replicate her initial studies included a) attitudes, normative beliefs and intentions to lose weight were not measured at equal levels of measurement specificity or b) low multiple correlations were obtained between attitudes, normative beliefs and behavioral intentions. Kristiansen and Eiser (1986) attempted to replicate Saltzer's findings while controlling two additional sources of error from Saltzer's (1981) study. The first was that Saltzer (1981) did not use the appropriate sample to identify the beliefs which were used to measure attitudes and normative beliefs. While her major study was with 20 adults attending a weight loss clinc, her pilot subjects were college students. Second, social desirability, a possible confounder, was not evaluated in Saltzer's 1978 study. Kristiansen and Eiser also were unable to replicate Saltzer's 1978 study. Results of Kristiansen and Eiser's study suggest that constructs other than locus of control, values and social desirability may play a more important role in the relationship between attitudes, subjective norm and behavioral intentions. They encourage future research on whether: perceived relevance of values and attitudes to various behaviours affects the influence which locus of control and values have upon the extent to which attitudes, as opposed to normative beliefs, predict behavioural intentions and overt behaviour (1986). Kristiansen and Eiser submitted that locus of control and values might be especially important variables since Saltzer's (1981) data suggest that these variables influence the degree to which intentions are transformed into overt behavior. Conceptual Regearch ugggl, The theoretical construct for this research is based on elements of both Rotter's Social Learning Theory and Ajzen and Fishbein's Theory of Reasoned Action. In a recent review of social psychological models of health behavior, Wallston and Wallston (1984) recommend using an integration of these models for best application of the results and for suggesting future tailoring of interventions to groups of people. The paradigm shown in 21 Figure 2 summarizes the integration of these models. 22 someones cu mucosa Haauo< .motoof. LoIbm new £22m: +0 co_+mcmo+c_ of. @5393 .262 . a 95m: uoasanom omceno t. .. coaucoucn uuueuo mo mucosaucu uo>aoowom uou>asom uncauwo no new occuwuu< IHII . .8923 cuouuoo unease uoc masonuxvasoge seacu nuoeuo one» message >¢ .N can .H mhmuqmn A<¢OH><=MQ 23 Hypothesis This study was designed to assess whether locus of control influences behavioral intentions indirectly by altering the relative importance of the attitude component of Fishbein's model of behavior intentions. More specifically, when compared to non-internals, it was predicted that for individuals who have an internal locus of control orientation and a high value of health, personal attitude toward the behavior will be a stronger predictor of behavioral intentions than social norm. In the ”Taking Control” program people are encouraged to change their behavior to obtain improved health and reduced cancer risk. Because the program encourages people to "take control" of their health, it was predicted that those individuals with an internal locus of control would be more likely to respond to the program than those subjects with a non-internal locus of control. Specifically, it was hypothesized that the internal/high health value subjects would have greater change in knowledge, attitude and intention to change behavior following the "Taking Control” program than non-internal subjects with high or low value of health. The null hypotheses for this research are: 1a. There is no difference in nutrition knowledge scores between subjects who have seen the "Taking Control“ program (group 1) and those who have not (group 2). 1b. There are no differences in attitude between group 1 and group 2. 2a. 2b. 3a. 3b. 24 There is no difference between nutrition/cancer knowledge test (NCKT) scores of the internal and non-internal subjects. There is no difference in attitude scores between the internal and non-internal subjects. In internal/high health value subjects, attitude does not account for more of the variation in intent to change dietary behavior than does social norm. In internal/low health value and non- internal subjects, more of the variation in intent to change dietary behavior is not explained by social norm than by attitude. METHODS Human Use Approval Prior to beginning this study, a copy of the proposal was submitted for review to the University Committee for Research Involving Human Subjects. The nature of the project allowed for exemption from full committee review, and the proposal and consent form were approved (Appendix B). A subject's signature on the consent form was understood as his or her written agreement to participate in the study. Subjects The subjects in this study consisted of 94 men and women between the ages of 20 and 60 from worksite wellness programs. While this was the target age group, no person interested in participating was turned down. Subjects in apparent ill health were not part of the evaluation. The worksites used in this study were recruited either by the local ACS county office or by the program presenter in a way that is similar to the usual operating procedure of ACS. Potential subjects were informed about the project. If interested in participating, they were given a written consent form with a verbal explanation. Subjects were informed that participation was totally 25 26 voluntary, that they could discontinue the experiment at any time without recrimination, and that all results would be anonymous. Instrumeptation NutritionlCancer Knowledge Test Before attempting to devise the Nutrition Cancer Knowledge Test (NCKT), a sampling of available knowledge tests was collected. There are many tests designed to evaluate nutrition knowledge (e.g. Kolasa, et.al., 1979), but relatively few have been validated to test the concepts of nutrition that relate to cancer risk. Since the objective was to evaluate the nutrition concepts specific to the Taking Control program of the American Cancer Society, an original test was developed (Appendix C). Egg; Development. The first phase of test development was the writing, review, and revision of the test objectives (Figure 3). These objectives were adapted from the program outline in the Taking Control Facilitators Guide (see Appendix D). ACS volunteers who were familiar with the Taking Control program were given these adapted objectives to review for accurate representation of the program (Appendix E). The objectives were found to be clearly stated and closely matched to the program's outline, so few revisions were required. To assure a balanced sampling of content, a two-way table of test specifications was developed based on the 27 Project Planning Develop Outlines Review Outlines "‘-—1; Revise Outlines Develop Objectives Review Objectives -~““‘ Revise Objectives Write Jest Items Conduct Content validity rr——v—"""Tests by Program Presenters Approval of Selected Objectives and Review Test Items ---~qb Revise Test Items Conduct Field Test of Test Items / Revise lI tems Content validate Test Items and Recommend Preliminary Passing Scores ~——~.4"Select Final Item Pool Produce Test Forms and Protocol Administer Tests Figure 3. Steps for test development. (Adapted from Mehrens, 1985). Table 3. 28 Test Specifications EBQIEEII!E.EA&IQBE Recognition of Sources (R) Understanding of Benefit/Risk (U) 1.* Cabbage family vegetables """ 5?""§IS§§'"""" ‘ ' 3. 'GTZIJJQ'; """"""""""" """ 3T""GIZZ;I;'E""""""m'm" ""17"”;{32235.1 BISE.EA§IQB§ 6. High-fat diets NA @ 7. Salt-cured, smoked and ‘ NA @ nitrite-cured foods " .. ...;ig: .. (1E3) ..... ;'.'""Ri;;;3{ n 16 T 10. Rxc;;sive Sun Exposure NA 13 Lgy_ Circled numbers correspond to item number on test. NA - Not Applicable '--' I no test item developed * Numbers in the left column are used to code test items to the Table of Specifications. For example, item number 2 is intended to measure understanding of subject area three-and is coded as 03. (See Appendix 3). *9 Items numbered 7 and 8 were omitted from final analysis due to low discrimination. *'* Item six spanned two subject areas. 29 objectives. As shown in Table 3, the general subject matter of the test is described down one side, with the level of learning outcomes across the top. More questions were developed for those sections of Taking Control that were emphasized. For example, more time is spent on the smoking section of Taking Control than the other risk factors, and two questions on the NCKT relate to smoking. Other risk factors are addressed in one question each. The shortness of the Taking Control program and the logistics of presenting the program to a worksite wellness population restricted both the level of learning which could be achieved and the length of time available to administer the test. For these reasons, only basic knowledge was tested, and the levels of learning outcomes were limited to recognition and understanding, with no attempt to ascertain application of the concepts. A five- option multiple choice test was used to minimize the effect of guessing. The table of specifications shows that there were several areas where recognition was ”not applicable”. The ”tobacco”, ”alcohol” and ”excessive sun exposure" items were considered to have obvious sources. For the ”nitrite-cured foods", sources were given in the stem of item 15. The sources of foods high in fat were not stressed in the Taking Control program and were not represented in the test. Test items were writen, reviewed and revised based on the specifications outlined in Table 3. Test item construction followed accepted guidelines on test 30 construction (Gronlund, 1977; Horris & Fitz-Gibbon, 1978; Ebel & Frisbie, 1986) (Appendix F). The final draft of the NCKT is included in Appendix C. Test item review and revision is discussed below. The NCKT was reviewed by the nutrition faculty in the Department of Food Science and Human Nutrition at Michigan State University (Appendix G). Suggestions from the faculty were incorporated into the revised version of the NCKT. 2119; Studies. The NCKT went through two pilot tests. The first pilot consisted of 33 subjects, the second had 40 subjects. The pilot sample populations were obtained from worksite presentations of the Taking Control program. The pilot subjects had the following similarities to the study population: same age range (30-60 years), consent to participate in a worksite program on nutrition and cancer risk reduction, and inclusion of both men and women in apparent good health. Worksite populations were chosen for this study because ACS programs are usually presented to this type of group (as opposed to a group of college students, for example). By choosing a similar population, the information obtained from administration of the NCKT is likely to show whether the Taking Control program is a useful educational tool for the ACS. Iggp_Apglysis. All test items underwent item analysis including measures of difficulty and discrimination using the data obtained from the pilot studies. Item analysis 31 was performed by the MSU Scoring Office. In the first pilot test, the term "difficulty" was used (and reported in Appendix H) as the percentage of a group of examinees who answered a test item correctly. As can be seen in Appendix H, the first pilot test revealed that the test had high difficulty scores (meaning that many subjects answered the question correctly). A medium range of difficulty was sought to contribute to maximum discrimination (Allen and Yen, 1979, p.121). Because the test reliability was relatively low, further test revision was undertaken. Distractors on this version of the NCKT that were not attractive enough for individuals to choose as possible answers were discarded and new distractors were written. The new distractors were then pilot tested. In the second pilot test (Appendix I), difficulty was defined as those subjects who did not answer a question correctly. The medium range of difficulties (ranging from .20 to .60) (Lord, 1953) was achieved. Likewise the discrimination was higher and reliability increased. Prior to pilot testing, it was decided that items with discrimination indices less than .30 were to be eliminated or modified to increase discrimination because items with high discrimination contribute most to test reliability (Ebel, 1986). However, after the second pilot test, two items with a discrimination index of .20 were retained in the final version of the NCKT due to the satisfactory overall test reliability obtained. Items number 7 and 8 were eliminated from the final analysis due to low 32 discrimination. Despite the omissions and alterations due to pilot testing, the table of specifications (Table 3) shows that there is a representative sampling of the different subject areas of the Taking Control program tested by the NCKT. Reliability gpg_Valigity. Reliability was determined from information obtained from pilot studies. Internal— consistency reliability was measured by the Kuder- Richardson formula 20 (KR20) (Allen and Yen, 1979). Before pilot testing, a reliability coefficient of .70 was considered acceptable (Ebel, 1979). The second NCKT pilot test achieved a reliability coefficient of .67. Two types of content validity were determined for the NCKT. The review of the NCKT items by volunteer program presenters of the ACS constitutes one method of determining its content validity. These volunteers were familiar with the ”Taking Control” program and evaluated the test in terms of appropriateness of questions, relevancy of the test to objectives and relevancy of the objectives to the ”Taking Control" program (Appendix E). In cases where the reviewers were not in agreement with the questions or objectives, the item was either modified or deleted. Another type of content validity was determined by requesting input from MSU faculty members who are knowledgeable in the cancer/nutrition field (Appendix G). They were asked to judge the information in the "Taking 33 Control" program in regard to the accuracy of the message based on current research. Comments and suggestions received by the faculty were taken into account in the revision of the NCKT. Multidimensional Health Locus pg_Control ggglg, The Multidimensional Health Locus of Control Scale developed by Wallston, Wallston and Devellis (1978) was used to assess the HLC of the subjects (Appendix J). Cancer Lpgp§,p;_Control Ig§§_ A 6-item Cancer Locus of Control scale (CALC) adapted from the Weight Locus of Control scale by Saltzer (1978) by changing the word "weight” to ”cancer" or ”getting cancer”. This scale was developed as a more specific scale than the MHLC for this application. The CALC has three internally worded items and three externally worded items. It has a 6-point Likert-type format and is scored in the internal direction (Appendix J). Health value Survey Health value was measured with a version of Rokeach's (1973) value survey as modified by Wallston (1976). Wallston used nine of Rokeach's values with an added tenth value of health. Subjects were asked to rank order the values in order of importance to them. ”High” or ”Low" health value was determined by a median split of the rank order scores on the survey (Appendix K). Intentiop §p_ggppgg Behavior Questippg Intention to change behavior was assessed using 34 fixed alternative measures designed to use the elements of the Theory of Reasoned Action developed by Fishbein and Ajzen (1980). Questions were developed following the procedures suggested by Saltzer (1978) (Appendix K). The dependent variable was the response to the question "I intend to start eating a more healthy diet in the near future”. The cancer-specific intent question was “It is very likely that I will start eating a diet aimed at lowering my risk of getting cancer". These questions were rated on a 6-point Likert-type scale ranging from Strongly Agree (scored as a 6) to Strongly Disagree (scored as a 1). This method was chosen to differentiate among several levels of feelings about the issue. Attitude Toward Nutrition ngptiops The personal attitude component of the Fishbein model was adapted from an attitude toward personal control over eating habits scale developed by Rosander and Sims (1981) (Appendix K). In addition, subjects' attitudes toward changing nutrition behaviors was assessed by scores on questions taken from an attitude scale developed by Boren et a1 (1983). Those questions from Boren's scale which loaded highly in factor analysis were chosen for use in this study. Subjective Np;p_Qpestions The subjective norm component of the model was assessed with regard to four referents: a) doctor b) parents c) spouse, fiance, boyfriend or girlfriend; and 35 d) close friends. The questions were adapted from the format used by Saltzer (1978) changing the word ”weight" for the word ”cancer”, and were rated on a 6-point Likert scale ranging from strongly agree to strongly disagree. All of the social norm, attitude and intention questions were combined into one questionnaire (Appendix K). Q§§g_Collgc§ipn At an informational meeting, subjects were given the HLC test, accompanied by the CALC and demographic questionnaires. The questionnaires were evaluated during the following week to determine the level of internality of each subject by adding up the values for the internally worded questions. Scores on the internal health locus of control questions were then randomly distributed into two groups (group 1 and group 2) such that each group had an equal number of internal and non-internal health locus of control subjects. This was to assure proper respresentation of locus of control in each group. Randomization was accomplished in the following manner: A cut-off score of 26 on internality was used for separating the internals from the non-internals. This value was determined by weighting and pooling the normative data from previous studies (Table 2) (Wallston and Wallston, 1981). This method provided a standard cut-off score to be used with groups from several different worksites. On the internal questions of the MHLC scale, subjects scoring less than or equal to 26 were classified as internals (group A) and subjects with scores above 26 36 were classified as non-internals (group B). Prior to the day of intervention, all A subjects were assigned to group 1 or 2 according to a number drawn from a box containing equal sets of paper labeled 1 or 2 . The same was done for B's. Those assigned a 1 were in group 1 and received a packet upon entering the room the day of intervention which contained the test of health behavior intentions, attitudes and health value to be completed first. The test of nutrition/cancer risk knowledge was included in the second portion of the packet, to be completed after the "Taking Control" program. Those assigned a 2 were in group 2 and received a packet with the nutrition/cancer risk knowledge test first; with the intentions, attitudes and health value in the latter portion. All tests were color coded so that the first test for both groups was blue, and the last test was a buff color. Both groups 1 and 2 viewed the ”Taking Control” program and participated in a question/answer session with a set list of answers to commonly asked questions. All subjects in the study had the same program presenter, and the same protocol was followed for each presentation of the program. The design of this study controlled for possible pre- test learning effects and was a modified version of the ”separate-sample pretest-posttest design” (quasi- experimental design .12) described by Campbell and Stanley (1963). Because both groups received the exact same 37 program at the same time, differences in presentation of the program and outside events which might affect responses were controlled. Randomizing between groups made possible comparing the pre—test of group 1 to the post-test of group 2 (and vice versa). See the experimental design outline in Figure 4. Q;§g,§pply§is All data were analyzed using the Statistical Package for the Social Sciences (SPSS/PC) (Norusis, 1984) on an IBM Personal Computer (IBM PC/XT). Subject response to the questionnaires and Knowledge Test was scored by computer scanning at the MSU Scoring Office. The reliability of the attitude and social norm scales were determined by Cronbach's alpha coefficient (Ebel and Frisbee, 1986). Coefficient alpha is the appropriate reliability estimate for scales composed of items scored with values other than 0 and 1 (i.e. non- dichotomously scored data such as the Likert scale). Lpgp§,p;_Control app Health Vplue. Based on the scores of internality and non-internality, the data were analyzed factorially by crossing the HLC category with health value, using a median split of health value based on sample rank frequencies. An analysis of variance (ANOVA) was computed to determine if these groups differed significantly in their attitudes toward the belief that one can change behavior to reduce cancer risk, and in their intention to change cancer-risk behavior. The F- ratio obtained from an ANOVA is a fraction in which the 38 HLC,CALC and Demographic Questions to All Subjects RANDOMIZE [Group J) Test of Health Behavior Intentions Attitudes and Hv ”Taking Control” Test of Nutritihn/Cancer Risk Knowledge Figure 4. Experimental design. [Group g} Test of Nutrition/Cancer Risk Knowledge 1 ”Taking Control" l . Test of Health Behavior Intentions Attitudes and Hv 39 numerator is the group (between group) mean square, and the denominator is the residual (within group) mean square (Iverson and Norpoth, 1987, p.37). The more significantly different two groups are, the larger the F—ratio. When the dependent variable is metric (interval) and the independent variables are categorical, ANOVA is the appropriate method of analysis (Iverson and Norpoth, 1987). A 0.05 level was the pre-determined criterion for significance. Elgpgpg§_pg Theory p£_Reasoned Aptiop. The personal attitude component of the Fishbein model was assessed by scores obtained on the attitude toward changing nutrition behaviors scale adapted from Boren, Dixon and Reed (1983) and from the personal-control-over-eating-habits scale developed by Rosander and Sims (1981). Social norm was assessed by questions developed according to the method described by Saltzer (1978). In the Fishbein equation, attitudinal and normative components are given empirical weights according to their relative importance in the prediction of behavior intentions (Ajzen and Fishbein, 1980). The weights are determined by a multiple regression equation where attitude toward the behavior and subjective norms are the two independent variables and the behavioral intention is the dependent variable. The empirical weights can be expressed as standardized regression coefficients, or beta weights (Saltzer, 1978). In the stepwise multiple regression analysis used here, the multiple correlation coefficient R is computed. 40 R-Squared is an estimate of the "proportion of the variation in the dependent variable 'explained' by the model” (Norusis, 1984, p. B-18l). The ”adjusted" R- squared attempts to correct R-squared to more closely reflect the goodness of fit of the model in the population (Norusis, 1984, pB—180). At each phase of the stepwise procedure, the variables are entered into the equation according to how well they fit the linear model. The standardized beta weights used to express the empirical weights of the Fishbein equation are the coefficients of the independent variables expressed in standardized (Z-score) form (Norusis, 1984, pB-197). The values of the standardized beta coefficients are contingent on the other independent variables in the regression equation. When there are correlations among the independent variables, the unique contribution of each beta coefficient is difficult to assess. RESULTS In this section, distribution and significance of differences in demographic and background variables between group 1 (pre-attitude, post-NCKT) and group 2 (pre-NCKT, post-attitude), between internals and non- internals and between worksites are given first. The significance of differences in the distribution of these variables was determined by the SPSS/PC crosstabs procedure (Norusis, 1984) using the Chi-Square test of independence. A significance level of p<.05 was predetermined as the criteria for true differences. variable interactions, as assessed by analysis of variance (ANOVA), and multiple regression follow. Population Cparactaristias A total of 94 subjects, 70 women and 24 men between the ages of 20 and 70, were included in this study. Subjects participated in the study through seven Michigan worksite wellness programs: Worksite One 19 subjects Worksite Two 7 subjects Worksite Three 10 subjects Wbrksite Four 12 subjects Worksite Five 14 subjects Worksite Six 12 subjects Worksite Seven 20 subjects Tables 4 and 5 give a summary of the descriptive 41 42 Table 4. Summary of demographic variables for total sample. Sex Female 70 Male 24 Age (years) 20-29 17 30-39 34 40-49 26 50-59 12 60-69 5 Education High School 33 HS + some college 11 Associate's Degree 6 Bachelor's Degree 31 Graduate Degree 12 Not Given 2 Occupation Clerical 29 Technical 8 Sales/Accounting 20 Managerial or Supervisory 16 Health Aide 8 Professional: Non-Health 4 Health 5 Not Given 4 43 Table 5. Summary of background variables for total sample. Subject has Cancer: Yes 0 No 86 Don't know 8 Family Member has Cancer Yes - 29 No 59 Doesn't know 3 AC8 Program in Past Yes/Nutrition Related 3 No 6 Yes/Not Nutrition 85 Dietary Change Last Two Years Weight/Cholesterol 74 Cancer-Related 4 Other 6 None 10 44 statistics for the entire sample. Group ;_apa ; Characterisgigs Before the null hypotheses were tested, descriptive data was analyzed for the two experimental groups (Group 1 = attitude pre-test, NCKT post-test; Group 2 = NCKT pre- test, attitude post-test) to verify that the two groups were alike in demographic and background variables. No significant differences were found between Group 1 and Group 2 for any demographic or nutrition/cancer background variable measured (p >.05) . Group 1 had 32 (70%) females and 14 (30%) males (Table 6). Group 2 had 38 (79%) females and 10 (21%) males. All subjects were between the ages of 20 and 60 years. Five subjects (11%) in group 1 were between 20 and 29 years old, 18 (39%) were between 30 and 39 years, 13 (28%) were between 40 and 49 years, 9 (20%) were between 50 and 59 years, and 1 (3%)-was between 60 and 69 years old. Twelve subjects (25%) in group 2 were between 20 and 29 years old, 16 (33%) were between 30 and 39 years, 13 (27%) were between 40 and 49 years, 3 (6%) were between 50 and 59 years, and 4 (8t) were between 60 and 69 years old (Table 6). All subjects had received at least a high school education. In group 1, 18 (40%) were high school graduates, 2 (4%) had gone through high school plus some college, 3 (7%) had an associate's degree, 17 (38%) had a bachelor's degree and 5 (11%) had earned a graduate 45 Table 6. Summary of Demographic variables for Groups 1 and 2. * Grpup ;, Gropp2 (post-NCKT) (pre-NCKT) Sex Female 33 (70%) 38 (79%) Male 14 (31%) 10 (21%) Age (years) 20-29 5 (10%) 12 (25%) 30-39 18 (38%) 16 (33%) 40-49 14 (30%) 13 (27%) 50-59 9 (19%) 3 ( 6%) 60-69 1 ( 2%) 4 ( 8%) Education High School 19 (40%) 15 (31%) HS + some college 2 ( 4%) 9 (19%) Associate's Degree 3 ( 6%) 2 ( 4%) Bachelor's Degree 17 (36%) 14 (29%) Graduate Degree 5 (11%) 7 (15%) Other 1 ( 2%) 1 ( 2%) Occupation Clerical 13 (28%) 16 (33%) Technical 1 ( 2%) 6 (13%) Sales/Accounting 13 (28%) 7 (15%) Managerial or Supervisory 6 (13%) 10 (21%) Health Aide 5 (11%) 3 ( 6%) Professional: Non-Health 2 ( 4%) 2 ( 4%) Health 2 ( 4%) 3 ( 6%) Other 5(11%) 1 ( 2%) *No significant differences on any characteristic (Chi Square, p>0.05). 46 degree. In group 2, 15 (31%) were high school graduates, 9 (19%) had gone through high school plus some college, 3 (6%) had an associate's degree, 14 (29%) had a bachelor's degree and 7 (15%) had earned a graduate degree (Table 6). Thirteen subjects (31%) in group 1 were employed in clerical or word processor positions, 1 (2%) was in a technical or computer-related field, 13 (31%) were accountants or salespeople, 6 (14%) held managerial or supervisory positions, 5 (12%) were health aides, 2 (5%) were non-health professionals, 2 (5%) were health professionals and 5 (11%) subjects did not state their occupation (Table 6). Sixteen subjects (33%) in group 2 were employed in clerical or word processor positions, 6 (13%) were in technical or computer-related fields, 7 (15%) were accountants or salespeople, 10 (21%) held managerial or supervisory positions, 3 (6%) were health aides, 2 (4%) were non-health professionals, 3 (6%) were health professionals, and 1 (2%) was a non-health professional. As outlined in the methods section, subjects with cancer were excluded from data analysis. Thus, no subjects had cancer (Table 7). In group 1, 41 (89%) did not have cancer, and 5 (11%) did not know if they had cancer. Eleven (24%) had a family member with cancer. Thirty-one (67%) did not have a family member with cancer and 4 (9%) did not know. In group 2, forty-five (94%) did not have cancer and 3 (6%) did not know if they had 47 Table 7. Background variables for Groups 1 and 2. Group A Group a Family Cancer yes 11 (24%) 18 (38%) no 31 (67%) 28 (58%) don't know 4 ( 9%) 2 ( 4%) ACS Program in Past Yes/Nutrition 0 3 ( 6%) No 5 (13%) 1 ( 2%) Yes/Not Nutrition 41 (87%) 44 (92%) Dietary Change Weight/Cholesterol 34 (74%) 40 (83%) Cancer-Related 2 ( 4%) 2 ( 4%) Other 2 ( 4%) 4 ( 8%) None 8 (17%) 2 ( 4%) 48 cancer. Eighteen (38%) had a family member with cancer, 28 (58%) did not have a family member with cancer and 2 (4%) did not know if they had a family member with cancer (Table 7). Most subjects had previously seen a program by the American Cancer Society (AC8). In group 1, 41 (87%) had seen an AC8 program before, but the program was not related to nutrition and cancer. Forty-four subjects in group 2 (92%) had seen an AC8 program, but only 3 (6%) had seen a program related to nutrition and cancer (Table 7). Most of the subjects had tried some type of dietary change within the past two years (Table 7). Thirty-four (74%) of the subjects in group 1 had attempted a weight reduction diet and/or a diet low in cholesterol, 2 subjects (4%) had tried a diet change that was cancer related (for example, increase fiber, eat more fruits and vegetables), 2 (4%) had tried an other diet, such as a low salt diet and 8 (17%) of the subjects had not tried to change their diets in the past two years. For group 2, 40 (83%) had attempted a weight reduction or low cholesterol diet, 2 (4%) had tried a diet change that was cancer related, 4 subjects (8%) had tried an ”other diet" and 2 subjects (4%) had not tried to change their diets. Due to the block randomization design of the study, the two groups had a nearly equal distribution of internal and non-internal subjects. Group 1 had 31 internal subjects (67%) and 15 non-internal subjects 49 (33%). Group 2 had 33 internal subjects (68%) and 15 non- internals (31%) (Table 8). The distribution of high and low value of health was not significantly different between the two groups (p>.05). Group 1 had 33 subjects (72%) who ranked health highly and 13 (30%) who ranked it relatively low on the value survey. Thirty-eight subjects (79%) in group 2 gave health a high rank and the remaining 10 subjects (21%) rated health as a relatively lower value. Table 8. Distribution of internality and health value between groups. Group 1 (post-NCKT)* Group 2 (Pre-NCKT) Internals 31 (67%) 33 (68%) Non-Internals 15 (33%) 15 (31%) High HV** 33 (72%) 38 (79%) Low HV 13 (30%) 10 (21%) *NCKT = Nutrition Cancer Knowledge Test **Hv= Health value Internal versus yon-Internal Saaple Characaaaistics There were more subjects with an internal score on the MHLC (67%) than those with a non-internal score on the MHLC (33%). The distribution of health value was significantly different between the internals and non- internals (p=.02). Fifty-three (83%) of the internals gave health a relatively high rank compared to 18 (58%) of the non-internals. No differences existed between internals and non- internals on demographic characteristics (Table 9). Fifty 50 Table 9. Demographic variables for internals and non-internals.* Internal Non-Ipterpala *No significant differences on any characteristic (Chi Square, p>0.05). Sex Female 50(78%) 20(67%) Male 14(22%) 10(33%) Age (years) 20-29 11(17%) 6(20%) 30-39 25(39%) 9(30%) 40-49 20(31%) 6(20%) 50-59 5( 8%) 7(23%) 60-69 3( 5%) 2( 7%) Education High School 25(39%) 8(27%) H8 + some col 10(16%) 1( 3%) Associate's 3( 5%) 2( 7%) Bachelor's 17(27%) 14(47%) Graduate 7(11%) 5(17%) Other 2( 3%) 0 Occupation Clerical 21(33%) 8(27%) Technical 5( 8%) 2( 6%) Sales/Account 11(17%) 9(30%) Managerial or Supervisory 11(17%) 5(17%) Health Aide 7(11%) 1( 3%) Professional: Non-Health 3( 5%) 1( 3%) Health 5( 8%) 0 Other 1( 2%) 4(13%) 51 (78%) of the internals were female, 14 (22%) were male. The non-internal subjects were comprised of 20 (67%) females and 10 (30%) males. Of the internal subjects, 11 (17%) were between 20 and 29 years old, 25 (39%) were between 30 and 39 years, 20 (31%) were between 40 and 49 years, 5 (8%) were between 50 and 59 years, and 3 (5%) were between 60 and 69 years old. Six (20%) of the non- internals were between 20 and 29 years old, 9 (30%) were between 30 and 39 years, 6 (20%) were between 40 and 49 years, 7 (23%) were between 50 and 59 years, and 2 (7%) were between 60 and 69 years old. Twenty-five (39%) of the internal subjects were high school graduates, 10 (16%) had gone through high school plus some college, 3 (5%) had an associate's degree, 17 (27%) had a bachelor's degree and 7 (11%) had earned a graduate degree (Table 9). For non-internals, 8 (27%) were high school graduates, 1 (3%) had gone through high school plus some college, 2 (7%) had an associate's degree, 14 (47%) had a bachelor's degree and 5 (17%) had earned a graduate degree. Twenty-one (33%) of the internal subjects were employed in clerical or word processor positions, 5 (8%) were in technical or computer-related fields, 11 (17%) were accountants or salespeople, 11 (17%) held managerial or supervisory positions, 7 (11%) were health aides, 3 (5%) were non-health professionals, and 5 (8%) were health professionals and 1 (2%) did not provide occpational information. Eight (27%) of non-internal subjects were 52 employed in clerical or word processor positions, 2 (7%) were in technical or computer-related fields, 9 (30%) were accountants or salespeople, 5 (17%) held managerial or supervisory positions, 1 (3%) was a health aide, 1 (3%) was a non-health professional, none were health professionals and 4 (13%) did not give their occupation (Table 9). Of the total sample, some people had cancer. They were omitted from the data analysis. Of the remaining 94 subjects, 5 (8%) of the internals and 3 (10%) of the non- internals did not know if they had cancer. Twenty-one internal subjects (33%) had a family member with cancer. Thirty-eight (59%) did not have a family member with cancer and 5 (8%) did not know. Eight (27%) of the non- internals had a family member with cancer. Twenty-one (70%) did not have a family member with cancer and 1 (3%) did not know. Of the internal subjects, 57 subjects (89%) had seen an ACS program before. However, only 1 (2%) had seen an AC8 program related to nutrition and cancer. Twenty-eight (90%) of the non-internal subjects had seen an ACS program, but only 2 (10%) had seen a program related to nutrition and cancer (Table 10). Fifty (93%) of the internal subjects had attempted a weight reduction diet and/or a diet low in cholesterol, 4 subjects (6%) had tried a diet change that was cancer related 7 (11%) had tried an "other diet” and 8 (13%) had 53 Table 10. Background variables for internals and non-internal subjects.* Ingernals Non-iptegpals Family Cancer yes 21(33%) 8(27%) no 38(59%) 21(70%) don't know 5( 8%) 1( 3%) ACS Program in Past Yes/Nutrition 1( 2%) 2( 7%) No 6( 9%) 0 Yes/ Not Nutrition 57(89%) 28(93%) Dietary Change Weight/Cholesterol 50(78%) 24(80%) Cancer-Related 4( 6%) 0 Other 2( 3%) 2( 6%) None 8(13%) 4(13%) *No significant differences on any characteristic (Chi Square, p>0.05). ' 54 not tried to change their diets in the past two years. For non-internals, 24 (80%) had attempted a weight reduction or low cholesterol diet, none had tried a diet change that was cancer-related, 3 (10%) had tried an other diet such as low sodium, and 3 (10%) had not tried to change their diets. Worksite Characteristics Several worksites participated in this study, as indicated in the beginning of this section. Statistical analysis revealed differences between worksites and these were significant at the .05 level. Differences were found in the distribution of sex, educational level, age and occupation between different worksites. The sampling technique used in this study controlled for worksite differences by dividing each worksite sample into groups 1 and 2 as described in the methods section. Thus the differences between these sites were equally distributed between the two groups: pre- and post-test. That the' pre- and post-test groups were the same was verified by the descriptive statistics above. The remaining demographic and background variables of cancer incidence, cancer of a family member, previous AC8 program participation, dietary changes, distribution of internal and non-internal subjects, or distribution of health value were not significantly different between worksites (p>.05) (Table 11). 55 Table 11. Worksite characteristics.* Worksitel: One Two Three Four Family Cancer yes 3(16%) 3(43%) 4(40%) 3(25%) no l4(74%) 4(57%) 6(60%) 9(75%) don't know 2(11%) 0 0 0 Dietary Change Weight/Cholesterol 17(89%) 6(86%) 8(80%) 8(67%) Cancer-Related 1( 5%) 0 0 1( 8%) Other 1( 5%) 0 0 1( 8%) None 0 1(l4%) 2(20%) 2(16%) Table 11 (cont). Worksitel: Five Six Seven Family Cancer yes 3(21%) 4(33%) 9(45%) no 11(79%) 5(42%) 10(50%) don't know 3(25%) 1( 5%) Dietary Change Weight/Cholesterol 9(64%) 10(83%) 16(80%) Cancer-Related 1( 7%) 1( 8%) 0 Other (eg low Na) 2(14%) 1( 8%) 1( 5%) None 2(14%) 0 3(15%) *No significant differences on any characteristic (Chi-Square, p>0.05 with similar worksites collapsed to keep cells with expected frequencies of (5 to a minimum). 56 Reliability and Distributiop 2; Scores NutritionZCancer Knowledge Test (NCKT) The mean score on the NCKT was 8.1 (SD = 3.3, variance = 10.9) (Appendix M). The final reliability (KR- 20) was 0.73. The mean item difficulty (based on incorrect responses) was 46 and mean item discrimination was 54. These values met the criteria set forth in the methods section. Multigipansional Health Locus g£_Coptrol Test (MHLC) The mean scores and standard deviations for the MHLC are given in Table 12. Table 12. Mean scores and standard deviations for MHLC. Mean Std. Dev. Internal HLC 28.15 3.67 Chance HLC 15.39 5.67 Powerful HLC 16.66 5.56 Cangar Locus g Control Scale (CALC) The mean and standard deviations for the CALC are given in Table 13. Table 13. Mean scores and standard deviations for CALC. Mean Std. Dev. Internal CALC 10.31 3.29 External CALC 8.83 3.73 57 aggagpga_89alas A mean score of 19.5 (SD = 3.1) was obtained on the Sims personal-control-over-eating-habits scale. The Cronbach's alpha for this scale was 0.45. On the attitude questions from Boren, the mean score was 15.2 (8D x 2.1). Cronbach's alpha for this scale was -0.74. The cancer attitude questions had a coefficient alpha of 0.34. Social up Lug Intent The social norm score was 17.53. The coefficient alpha for this scale was 0.39. The average score on the intent question was 1.81 (SD = 1.1). The mean for the cancer behavior intent (cbi) question was 2.0 (SD = 1.1). Correlatigns getweea Saales MHLC apa,CALC On the MHLC scale, PHLC correlated positively and significantly with the CHLC (Table 14). The internal cancer locus of control (ICALC) scale correlated positively and significantly with the internal health locus of control (IHLC) scale. The external cancer locus of control (ECALC) correlated positively and significantly (p<.05) with both the chance (CHLC) and powerful others (PHLC) health locus of control scales. 58 Table 14. Correlations between the MHLC and CALC. Correlations: IHLC CHLC PHLC ICALC ECALC IHLC 1.0000 -.1325 -.0702 .3880** -.1115 CHLC 1.0000 .5032** -.0919 .5050** PHLC 1.0000 .1269 .2678* ICALC 1.0000 -.2038 N of cases = 94 * = p<.01 ** = p<.001 Relationshipa Betwaen Vagiaplas Eggact g;_Vaaiablas pp_§CKT Scares A two-way ANOVA with NCKT as the dependent variable and group and internality as the independent variables, revealed that both internals and non-internals taking the NCKT after the "Taking Control" program had significantly (p<.0001) higher scores than those taking the NCKT prior to the program (Table 15). There were no significant differences in NCKT scores between internal and non-internal subjects (Figure 5). 59 Table 15. Two-way ANOVA of NCKT to test for the effects of group and internality. Sources Degrees Sum of Mean F p of of Squares Square value value variation Freedom Main Effects 2 292.830 146.415 17.953 <.0001** Group* 1 286.629 286.629 35.145 <.0001** Internality 1 4.463 4.463 .547 .461 Interaction Effect Group by Internality 1 4.252 4.252 .521 .472 Explained 3 297.082 99.027 12.142 .0001** Residual 89 725.842 8.156 Total 92 1022.925 11.119 * Group = Pre- or post-test **p (.001 6O .mcouw an .562 .m use»: c.5275: ...... :83... 53 ..e 83.22. 3 ...... .3... 892.35. ... 856 as.» 3 sites: . niecesszfiu—afissn an: r8... zfi’XI-e 61 A one-way analysis of variance with NCKT as the dependent variable and groups of subjects classified as either high health value internals (N853), or low health value internals and non-internals (N=41) was done to determine whether internal subjects with high health value performed differently on the NCKT than other subjects. The results of this analysis revealed no significant differences as shown in Table 16. Table 16. Oneway ANOVA of NCKT by Internals with High Hv versus Others. Sources Degrees Sum of Mean F of of Squares Square Ratio value variation Freedom Between Groups 1 21.528 21.52 1.88 .1732 Within Groups 92 1051.376 11.43 Total 93 1072.904 *p (.05 *‘p (.001 When these groups of subjects were divided into groups according to whether they were pre- or post-NCKT tests, there were significant differences between the four groups formed (i.e. high HV/internal post-NCKT, n-24; high HV/internal pre-NCKT, ns28; low HV/interna1+non-interna1s post-NCKT, n-21; and low HV/internal+non-internals pre- NCKT;n-20). The results of a one-way ANOVA with Fisher's Least Significant Difference Test for multiple comparisons between means (SPSS Ranges subcommand, Norusis, 1984) are summarized in Table 17 and Figure 6. The pre- to post- test scores on the NCKT were the only significant 62 .358» cooauoa neocououuao oz .253» 52— oooo 3 one .8539. A3 ooocouomuuo vacuums»: a 3393.“an caucuses—«loos one on?» canoe..— aon cums oHoEoucn compo» 33¢:ro oases guano..— :u«: so? mus—£35 you moon» ma monsoon E: ..w vacuum Ln: -..: 229-. 5.. 20:“ 63 differences found. No differences in the four means were attributable to internality/Hv. Table 17. Oneway ANOVA of NCKT by Internals with High Hv versus Others, Pre- and Post-Test. Sources Degrees Sum of Mean F p of of Squares Squares Ratio value variation Freedom Between Groups 3 327.385 109.1282 13.174 .0000** Within Groups 90 745.520 8.2836 Total 93 1072.904 *p <.os tap <.001 Effegt p;_variaples pa_§ttitudea A two-way ANOVA revealed that there were no significant differences in attitude between the two experimental groups or between internals and non-internals as measured by either the personal- control-over-eating-habits scale of Rosander and Sims, by the attitude scale adapted from Boren et a1. or by the two scales combined. One-way ANOVAs with NCKT as the dependent variable and groups of subjects classified as either high health value internals or low health value internals and non- internals were done to determine whether internal subjects with high health value performed differently on the attitude scales. There were no significant differences between internals and non-internals even when health value was taken into account, on either the Sims scale or the Boren scale. When these groups of subjects were divided into 64 groups according to whether they were pre- or post- attitude tests, there were no significant differences between the four groups. Effect gj_Variaples pp_Social Npaa_ ANOVAs revealed no significant differences in social norm scores between groups 1 and 2 (pre- and post-), between internal and non- internal subjects, or between internal subjects with a high health value and the remaining subjects. When these groups of subjects were divided into groups according to whether they were pre- or post- attitude tests, there were no significant differences between the four groups. Eggaga p; Variaples ga_1p§ent ANOVA showed that there were no significant differences in the variation in intent scores of internal subjects with high health value versus internal subjects with low health value and non-internal subjects in either group (pre- or post-test). Regression analysis revealed that the Roasander and Sims personal-control-over-eating-habits attitude scale and social norm both accounted for significant amounts of the variation in intent to change dietary behavior. These results are shown in Table 18. The scores on the Boren attitude-toward-nutrition scale did not reach the significance level required for entry into the equation. 65 Table 18. Results of regression analysis of the dependent variable: intent to change diet. Step/ Multiple Adjusted F Sig. variable R R-squared value of F Beta 2.Social Norm .3965 .13665 7.648 .0009 .3038 For the internals with high health value, stepwise regression analysis of the independent variables Boren- attitude-scale scores, Sims-attitude-scale scores, and social norm scores revealed no significant differences in the amount of variablility in the intention to change dietary behavior explained by these independent variables. Stepwise regression analysis of the independent variables Boren attitude scale scores, Sims attitude scale scores, and social norm scores on the dependent variable intent to change dietary behavior was also done for the internals with low health value and non-internal subjects. In these subjects, scores on the Sims scale accounted for a significant amount of the variation in intent (Table 19). The other independent variables in the regression, scores on the social norm and Boren scales, did not reach the required significance level (p<.05) to enter into the regression equation. Thus these variables did not account for significant variation in dietary behavior change intention in low health value internals and non-internals. 66 Table 19. Results of regression analysis of the dependent variable intent to change diet for internal with low health value and non-internal subjects. Step/ Multiple Adjusted F Sig. variable R R-squared value of F Beta Score on Sims .3979 .1343 6.58 .015 -.3979 The low health value internals and non-internals were divided into groups according to whether they were pre- or post-attitude tests (i.e. high HV/internal pre-attitude test, high HV/internal post-attitude, low HV/interna1+non- internals pre-attitude, and low HV/internal+non-internals post-attitude). Regression analysis of intent by social norm, the Sims scale and the Boren scale for each of these four groups showed that only the low HV/internal+non- internals pre-attitude group had an independent variable (score on the Sims scale) which achieved the required significance level to enter the stepwise regression equation. These results from this group are shown in Table 20. The independent variables social norm and score on the Boren scale did not enter into the stepwise equation. 67 Table 20. Results of regression analysis of the dependent variable intent to change diet for internal subjects with low health value and non-internal subjects taking the pre- attitude test (N=22). Step/ Multiple Adjusted F Sig. variable R R-squared value of F Beta Score on Sims .7178 .4898 20.20 .0002 -.7178 Effect pg_va;iables pa Cancer Bepavio; latent A cancer-specific attitude scale was used in conjunction with social norm and Sims as the independent variables, and intention to change dietary behavior related to cancer as the dependent variable, to determine if a scale more specific to cancer issues than Boren would account for variation in cancer behavior intention (C81). The results of stepwise regression analysis of these variables in internal subjects with high health value are summarized in Table 21. Social norm and the score on Sims did not reach the required significance level for entry into the equation, although social norm was close (p=.062, Beta=.244$) Table 21. Results of regression analysis on C81 in internal subjects with high Hv (N=49). Step/ Multiple Adjusted F Sig. variable R R-squared value of F Beta Cancer Attitude .4427 .1789 11.46 .0014 .4427 68 The same regression as described above was run on the scores from internal subjects with low health value and non-internal subjects. In these subjects, social norm and the Sims scale were significant components of the variation in intent to change dietary behavior related to cancer. These results are summarized in Table 22. The cancer-specific attitude scale (p=.1311) did not reach the .05 significance level required for entry into the regression equation. Table 22. Regression analysis on C81 in internal subjects with low Hv and non-internals. Step/ Multiple Adjusted F Sig. variable R R-squared value of F Beta 1.Social Norm .4925 .2209 11.21 .002 .4925 When the internals with high Hv versus internals with low Hv and non-internals were divided into groups according to whether they were pre- or post-attitude tested, all four groups had components which reached the required significance level for entry into the regression equation. These results are summarized in Tables 23 to 26. Neither the Sims nor the social norm scale reached the required significance level for entry into the equation to account for variation in cancer behavior intent. 69 Table 23. Regression analysis of CBI in internal subjects with high health value who took the pre-attitude test. Step/ Multiple Adjusted F Sig. variable R R-squared value of F Beta Cancer Attitude .4776 .1930 6.50 .018 .4776 Table 24. Regression analysis of CBI in internal subjects with high health value who took the post-attitude test. Step/ Multiple Adjusted F Sig. variable R R-squared value of F Beta 1. Social Norm .4658 .1829 6.3730 .0189 .4658 2. Sims .6450 .3629 7.8345 .0027 -.4966 Cancer attitude scores did not reach the required significance level for entry into the stepwise equation to account for variation in CBI. Table 25. Regression analysis of CBI in internal subjects with low health value and non-internals who took the pre- attitude test. Step/ Multiple Adjusted F Sig. variable R R-squared value of F Beta 1. Sims .5492 .2628 7.7735 .0121 -.5492 2. Social Norm .7413 .4966 10.3696 ,.0011 .4982 Cancer attitude scores did not reach the required significance level for entry into the equation to account for variation in CBI. 70 Table 26. Regression analysis of CBI in internal subjects with low health value and non-internals who took the post- attitude test. Step/ Multiple Adjusted F Sig. variable R R-squared value of F Beta Cancer Attitude .6163 .3355 8.5725 .0110 .6163 Neither scores on the Sims scale (p=.7558) nor the social norm (p=.2447) met the required significance level for entry into the equation to account for variation in CBI. Sumaary p;_§asplts The results of regression analyses with intent to change diet as the dependent variable and scores on the Sims and social norm scales as the independent variables are summarized in Table 27. A summary of the regression analyses with intent to change diet specific to cancer as the dependent variable and scores on the Sims, cancer attitude and social norm scales as independent variables is given in Table 28. With the Sims scale, all significant results (denoted by a ”+”) are based on negative correlations (i.e. negative standardized betas were obtained). These results are discussed in the next section. 71 Table 27. Summary of effects on intent to change diet. Sims' Attitude Social Norm Inteanaleiga _! - overall ' - - - pre-test - - - post-test - - Internal/Low Hv a,Nop-inta;pal - overall + - - pre-test + - - post-test - - - = no significant effect + = significant effect (p<.05). 72 Table 28. Summary of effects on intent to change dietary behavior related to cancer. Sims' CA* Social Norm InternallHigh fl! - overall - + - - pre-test - + - - post-test + - + Internal/Low Hv a,Non-ipte£nal - overall + - + - pre-test + - + - post-test - + - * CA = Cancer Attitude - = no significant effect + a significant effect (p<.05). DISCUSSION Introdpction In this chapter, the effect of demographic variables is considered first, followed by a look at the interactions of variables and support or rejection of the null hypotheses set forth previously. Strengths and limitations of the study are discussed at the end. The primary question investigated by this research was whether the health locus of control and health value of viewers of the ”Taking Control” program influenced their acquisition of knowledge, change in attitude and/or intentions to adopt a healthier diet after participating in the program. The answer to this question will provide useful information from which to make recommendations for future program development and to provide more careful targeting of health messages to specific groups within the general public. Because the demographic variables were not significantly different between internals/non-internals and between groups 1 and 2, it is assumed that the worksite differences were randomly distributed between experimental groups and therefore considered to have a ‘ negligible effect on outcomes. In addition, the study 73 74 design controlled for worksite differences by dividing subjects at each worksite into groups 1 and 2. NutritionlCancer Knowledge Test (QCKT) Scores Lp_ Pre- versus Post-test Groupa HYPOTHESIS 1a: There is no difference in nutrition knowledge scores between subjects who have seen the ”Taking Control" program (group 1) and those who have not (group 2). The consistently significant results of higher post- than pre-test scores on the NCKT indicates that a significant amount of the variation between scores on the NCKT can be attributed to program participation. Based on the results of the ANOVA, hypothesis 1a can be rejected. Effect g§_Programland Internality pp Attitude HYPOTHESIS 1b. There are no differences in attitude between group 1 and group 2. HYPOTHESIS 2b. There is no difference in attitude scores between the internal and non-internal subjects. Due to the lack of significant differences in attitude for groups 1 and 2, hypothesis 1b cannot be rejected. Likewise the internal and non-internal subjects had no differences in attitude scores, and hypothesis 2b was not rejected. Effect a; Interpality pa_NCKT Scores HYPOTHESIS 2a: There is no difference between NCKT scores of the internal and non-internal subjects. ANOVA was employed to test the second null hypothesis. The lack of significant differences in NCKT scores between internal and non-internal subjects suggests that internality was not a factor in knowledge acquisition 75 in this study. Therefore hypothesis 2a cannot be rejected. A possible explanation for the lack of an effect of internality on NCKT outcomes may be because the program influenced the internality of the subject. This is discussed in more detail in the Strengths and Limitations section below. Effects g;_the Elamants of age Fis ape ia Eguatiop p_, Iptentigp ;_ Change pietary Be pav igr HYPOTHESIS 3a. In internal/high health value subjects, attitude does NOT account for more of the variation in intent to change dietary behavior than does social norm. 3b. In internal/low health value and non- internal subjects, more of the variation in intent to change dietary behavior is NOT explained by social norm than by attitude. To test the third hypotheses, the analysis technique used by Saltzer (1978) was employed. The standardized beta weights obtained from multiple regression analysis were used as weights in the Fishbein equation. Using this method, data analysis suggests that when a cancer-specific attitude scale was used to assess intention to change dietary behavior to lower cancer risk, in internal/high health value subjects, attitude toward cancer was the dominant component of the variation in intention, when compared to social norm. In internal/low health value and non-internal subjects, social norm was the dominant component of the variation in intention to change dietary behavior to lower cancer risk. The finding that attitude was more important than social norm for internal subjects, 76 and that social norm was more important for non-internal supports the rejection of null hypotheses 3a and 3b. The results from the general attitude and intention scales are more difficult to interpret. The analysis of the internal/high health value data are not significant and do not support rejection of hypothesis 3a. However, the results of the analysis of the internal/low health value and non-internals suggest that attitude accounts for more variation in intent to change dietary behavior than does social norm. This finding does not support rejection of hypotheses 3b, and even implies the opposite relationship. This may have been due to the high health value-internals' perception that their diets were already nutritionally sound, and therefore did not intend to change them, whereas internal subjects with low health value and non-internals may have felt their diets needed to be changed. From the results of regression analyses, it appears that the cancer-specific attitude and intention questions formed a better "fit” to the linear model representing the Fishbein equation. This is similar to Saltzer's findings (1978). The elements of the Fishbein equation were not significant predictors of intention to lose weight when subjects were classified by the MHLC in her study. When she classified subjects according to a specific weight locus of control scale, she found the attitude and social norm elements to predict intention to lose weight. 77 A possible explanation of the results summarized in Tables 27 and 28 is that the internal subjects with high health value may have been concerned about cancer risk prior to participation in the ”Taking Control” program. After viewing the program, they may have been less concerned about cancer because they felt that their diets already posed little risk. Conversely, the internal subjects with low health value and non-internals may not have been too concerned with cancer risk prior to the program, but after learning of the cancer risks, they may have become more concerned about their cancer risk due to diet. Strepgths app Lipitagipns p;_§ae ggpay Before accurate conclusions can be drawn from any survey method of research the question of instrument reliability and validity must be addressed. Instrupapt Reliability apg_valigi§y A strong point of this study was that appropriate measures were taken in the preparatory stages to insure that the NCKT and MHLC instruments were valid and reliable. A limitation of this study is that the scales used for assessing health value and attitudes were less clearly validated. While factor analysis had been done on the Boren scale, not all of the questions given on Boren's scale were included. This compromised the validity of the scale. However, those items from Boren's scale with the highest factor loading were the ones chosen for use in 78 this study. Limited validation and factor analysis had been previously performed with the personal-control-over- eating-habits scale by Sims. The health value survey adapted from Rokeach by Wallston was the least validated of the scales used in this study. Wallston (1978) has used it in conjunction with the MHLC, but little reliability and validity of the modified version has been determined. Because the value survey asks subjects to give the relative value of health in comparison to other values, a "low” score of health should be viewed as only relative since everyone in our society values health rather highly (Wallston, 1978). Although the general dietary intention scale revealed few significant results, the cancer specific scales did provide a possible explanation of variance in intent to change dietary behavior related to cancer. The ability of these scales to provide significant results despite their low reliability suggests that the magnitude of the trends observed must be so great as to be detected despite the unrefined tools employed. The lack of detection of significant results with the general dietary scale may have been due to the inability of the low reliability scales to uncover the trends. Interpal validity Isspes Cook and Campbell (1979) define internal validity as: the validity with which statements can be made about whether there is a causal relationship from one variable to another in the form in which the variables were manipulated or 79 measured. (Campbell and Cook, 1979, p.38) Possible threats to internal validity described by Cook and Campbell which are relevant to the study include history (or previous experiences of the subject), maturation (how the subject might change during the time of the study), testing (effects of test taking on the subject), selection (how the subjects are chosen for participation in the study), and diffusion of treatments (when subjects in one group learn about the program from subjects in another group). Each of these will be discussed below. Histpgy. Cook and Campbell (1979) describe history as a threat when an observed effect might be due to an event other than the treatment which takes place between the pre- and post-test. For example, when programs take place over extended periods of time, subjects might learn about topics covered by the program when they go home and read the evening paper. In this study, the effect of history was minimized by (1) scheduling the post-test immediately after the program, (2) having both groups in the same room at the same time for the same program (3) randomly allocating subjects to group 1 or 2 so that previous knowledge should be randomly distributed between the pre- and post-test groups. A possible effect of history on the internal validity of this study concerns the attribution of performance on the NCKT to the "Taking Control” program. In comparing group 2's pre-test of NCKT to group 1's post-test of NCKT, 80 the difference between these two groups is used to explain the effects of the treatment (viewing the "Taking Control” program) on NCKT scores. However, group 1 had not only received the treatment, but had also taken the attitude questionnaire and health value survey. In a sense, the treatment and attitude are confounded in this study. The design of this study did not allow for the determination of the unique contributions of the attitude scale or the treatment. However, the fact that there were no differences in attitude between group 1 and 2 (unable to reject hypothesis 1b) suggests that taking the attitude test had negligible effects on the NCKT scores. Maturatiop. The threat of maturation is minimized by the shortness of the intervention in this study. The entire program including the pre- and post-tests took approximately one hour. However, one concern with the study was the possibility that the HLC of the subjects may have been changed by the ”Taking Control" program. The program strongly encourages people to understand that they do have some control over their health. To test whether the HLC of subjects was changed by participation in "Taking Control”, the MHLC was given to a separate group of subjects who were members of a civic club. Half of these subjects took the MHLC prior to seeing ”Taking Control" and the other half took the MHLC after the program. This 81 population was significantly different from the worksites in sex (mostly male) and age (older population), but the information on the effect of ”Taking Control" on HLC was still of interest. Analysis of this data revealed no differences in internality on the MHLC or Cancer locus-of- control scales from pre- to post-test. Wallston and Wallston (1984) encourage researchers to view the health locus of control construct not as an unchanging personality variable, but more as a charactersistic of an individual at a given point in time. Tesaipg. The design of this study controlled for the effects of test-taking. Because each test was given only once, there was no chance of enhancement of performance due to increased familiarity with the test after multiple testing situations. Selecgion. One of the limitations of this study is that subjects were self-selected into the program. ”Taking Control" was offered to all employees at each site, however many factors could be involved in the subjects' self-selection into the program. For example, internal subjects may be more likely than non-internals to attend a program that emphasizes personal control over cancer risk. On the other hand, internals may feel like they already have control over their cancer risk and therefore are not in need of a such a program. To examine the differences between self-selected participants in this study and non-participating employees, the MHLC and demographic information was 82 collected from employees who did not attend at three worksites (N840). Chi-square analysis of these data showed that the distribution of internal scores was not significantly different between those subjects who participated and this sample of non-participants. There were no significant differences (p>.05) for any other demographic or background variables except occupation. This difference in occupation may have been due to the three worksites from which this comparative data was collected, because these subjects were all government employees. Diffusipn p; Traagaants. Because all subjects in each worksite viewed the program at the same time, diffusion of the information in the program was controlled. Exterpal validity Laapaa Campbell and Cook (1979) describe external validity as: the approximate validity with which conclusions are drawn about the generalizability of a causal relationship to and across populations of persons, settings, and times. (Cook and Campbell, 1979, p.39). Potential sources of threats to external validity relevant to this study include: a) the interaction of subject selection and treatment; b) interaction of treatment setting and treatment; and c) the interaction of subjects' history and treatment. Lpgaaaggagp_p;,galacgipn gag Tyeataapt. In many of the 83 studies in the health locus of control area college students have been selected as subjects. Because college students score slightly higher than healthy adults (Wallston, 1982) on the MHLC it is difficult to draw relevant conclusions from previous studies to the general public. In addition, college students are likely to have good health and to be well-educated, they may not value health as highly as older people. It is therefore difficult to draw relevant conclusions from these studies in comparison to the general public. This study differed from most in that the observed population was comprised of worksite wellness program participants. The results can be more easily generalized to adult populations pertinent to the ACS than can results from studies on college students. Because the audience in this study is similar to ACS's usual target audience, the information on the ”Taking Control” program should be useful to the ACS for evaluative purposes. Another strength of this study is that the MHLC was combined with a measure of health value as Rotter (1975) recommends. Locus of control relates to expectancy of reinforcement, and Rotter's social learning theory equally emphasizes the situation in which the reinforcement occurs as well as the value of the reinforcement. It is for this reason that Rotter (1975) urges researchers to include a measure of the value of the reinforcement as well as a 84 situation-specific measure of locus of control. In this study, the MHLC scale specific to health issues was used rather than the generalized I-E locus of control scales. In addition, a scale was developed to measure the cancer-specific construct of locus of control. Rotter has recommended that specific measures of locus of control be developed for specific applications. Interaction g;_Histo;y aag,Traatment. This threat to external validity is difficult to distinquish in this study. Because all of the different worksites participated in the program on different days, spanning a time period of several months, it was assumed that any causal effect on outcomes due to special days of the year or special conditions of the day of presentation were cancelled. In addition, the subjects were randomly distributed into the two groups at each worksite, which should also have served to control for any special conditions of a particular presentation day. SUMMARY AND IMPLICATIONS The results of this study suggest that the "Taking Control" program effectively conveys nutritional inform- ation to worksite wellness program participants. The locus of control of subjects did not influence the amount of knowledge gained from participation in "Taking Control". Intent to change dietary behavior was significantly greater in internals with low health value and non- internals than in internal subjects with high health value in this study. As discussed previously, this may have been due to the high health value-internals' perception that their diets were already nutritionally sound, and therefore did not intend to change them, whereas internal subjects with low health value and non-internals may have felt their diets needed to be changed. In the internal subjects with high health value, the scores on scales used to measure attitudes and social norms did not account for variation in intent to change behavior. With intent to change dietary behavior related to cancer, attitude toward cancer was important in the pre-attitude test sample; social norm and attitude toward personal-control-over-eating-habits (Sims' scale) were 85 86 significant components of variation in the post-test sample. In the internal subjects with low health value and non-internal subjects, scores on the Sims' scale was a significant component of the variance in intent to change dietary behavior. In the pre-attitude test group, the same results were found. In the post-test sample, none of the scores on these scales were significant components of variance in intent to change dietary behavior. Social norm and scores on the Sims' scale were significant components in intention to change dietary behavior relative to cancer in these subjects. In the pre-attitude test sample, social norm and the Sims' scale were the only significant components of the variation in intention to change behavior in internals with low health value and non-internals. In the post-test group of these subjects, cancer attitude was the significant component of variation in intent to change cancer behaviors. A possible explanation of these results is that the internal subjects with high health value may have been concerned about cancer risk prior to participation in the ”Taking Control” program. After viewing the program, they may have been less concerned about cancer because they felt that their diets already posed little risk. Conversely, the internal subjects with low health value and non-internals may not have been too concerned with cancer risk prior to the program, but after learning of 87 the cancer risks, they may have become more concerned about their cancer risk due to diet. If the relationship between attitudes and social norm to health behavior intention is to be fully recognized, reliable scales need to be developed and validated for the measurement of these constructs. Future research also needs to study the relationship between behavior intention and overt behavior with respect to these variables. The results of the NCKT suggest that the ”Taking Control" video of the American Cancer Society provides an effective method of public education in a worksite setting. ”Taking Control" is recommended as a prototype for future AC8 programming. LI ST OF REFERENCES LIST OF REFERENCES Ajzen I. and Fishbein M. Understanding Attitudes and Predicting Social Behavior. Englewood Cliffs, NJ: Prentice-Hall, 1980. Allen M.J. and Yen W.M. Introduction to Measurement Theory. Belmont, CA: Wadsworth, 1979. American Cancer Society. Cancer Facts and Figures - 1988. American Cancer Society: New York, 1988. American Cancer Society. Taking Control - Public Education Program on Cancer Prevention - Risk Reduction. American Cancer Society: New York, 1985. Balch P. and Ross A.W. Predicting success in weight reduction as a function of locus of control: A unidimensional and multidimensional approach. Journal of Consulting and Clinical Psychology, 43 (l):119, 1975. Boren, A.R., Dixon, P.N. and Reed, D.B. Measuring nutrition attitude among university students. Journal of the American Dietetic Association, 82:251-253, 1983. Brown, K.H. and Cooke, T.M. The general public. Journal of Nutrition Education, 12 (2) (supp):117-120, 1980. Campbell D.T. and Stanley J.C. Experimental and Quasi- Experimental Designs for Research. Chicago, IL: Rand McNally College Publishing Co., 1963. Cancer Prevention Awareness Survey. U.8. Dept. of Health and Human Services, Public Health Service, National Institutes of Health, National Cancer Institute, NIH Pub. No. 84-2676,1984. Coan, R.W. Personality variables associated with cigarette smoking. Journal of Personality and Social Psychology, 26:86-104, 1973. Coelho R.J. A psychometric investigation of the multidimensional health locus of control scales with cigarette smokers. Journal of Clinical Psychology, 41 (3):372-376, 1985. 89 Cook, T.D. and Campbell, D.T. Quasi-experimentation: Design and Analysis Issues for Field Settings. Houghton Mifflin Co., Boston, 1979. Dalton, 8.8., Linke, R.A. and Simko, M.D. Worksite food choices: An investigation of intended and actual selections. Journal of Nutrition Education, 18 (4):182-186, 1986. Danahar, B.G. Rapid smoking and self-control in the modification of smoking behavior. Journal of Clinical and Consulting Psychology, 45:1068-1075, 1977. Doll R. and Peto R. The Causes of Cancer. New York: Oxford University Press, 1981. Ebel, R. L. Essentials of Educational Measurement, 3rd ed. Englewood Cliffs, NJ: Prentice-Hall, 1979. Ebel, R. L. and Frisbie, D.A. Essentials of Educational Measurement, 4th ed., Englewood Cliffs, NJ: Prentice Hall, 1986. Fishbein, M. and Ajzen, I. Belief, attitude, intention and behavior: An introduction to theory and research. Reading MA: Addison-Wesley Publishing Co., 1975. Godin, G., Cox, M.H., and Shephard, R.J. The impact of physical fitness evaluation on behavioral intentions towards regular exercise. Canadian Journal of Applied Sports Sciences, 8:240-245, 1983. Greenwald P. and Sondik R.J., eds. Cancer Control Objectives for the Nation 1985 - 2000. U. 8. Dept. of Health and Human Services, Public Health Service, National Institutes of Health, National Cancer Institute, Bethesda, 1986. Gronlund, N.E. Constructing Achievement Tests, 2nd ed., Englewood Cliffs, NJ: Prentice Hall, 1977. Horowitz M.B. Psychosocial mediators of abstinence, relapse, and continued smoking: a one-year follow-up of a minimal intervention. Addictive Behaviors 10:29-39, 1985. Iverson, G.R. and Norpoth, H. Analysis of variance, 2nd ed., Sage University Paper series on Quantitiative Applications in the Social Sciences, 07-001. Beverly Hills: Sage Pubns, 1987. James W.H., Woodruff A.B. and Werner W. Effect of internal and external control upon changes in smoking behavior. Journal of Consulting Psychology, 29:184-186, 1965. 90 Jordan-Marsh, M. and Neutra, R. Relationship of health locus of control to lifestyle change programs. Research in Nursing and Health, 8:3-11, 1985. Kaplan, G.D. and Cowles, A.C. Health locus of control and health value in the prediction of smoking reduction. In Wallston, K.A. and Wallston, B.S. (eds) Health Education Monographs: Health locus of control, 6(2):129-137, 1978. Kolasa, K., Lackey, C., Penner, K. and Hutch, B. Nutrition Education and Training Activities - Michigan State University 1978 - 1979, Volume 3: Nutrition Knowledge and Attitude/Behavior Test. Michigan State University, 1979. Kristiansen, C.M. value correlates of preventive health behavior. Journal of Personality and Social Psychology, 49:260-271, 1985. Kristiansen, C.M. and Eiser, J.R. Predicting health- related intentions from attitudes and normative beliefs: The role of health locus of control. British Journal of Social Psychology 25:67-70, 1986. Lanza, E. Diet and Cancer. In: The Surgeon General's Report on Diet and Health, Office of Health Promotion, U.S. Dept. of Health and Human Services, 1987. Lefcourt, H.M. Research with the Locus of Control Construct (Vol 3), New York: Academic Press, 1982. Lord, F.M. and Novick, M.R. Statistical Theories of Mental Test Scores. Reading, MA: Addison-Wesley, 1953. Manno, B. and Marston, A.S.: Weight reduction as a function of negative covert reinforcement (sensitization) versus positive covert reinforcement. Behavior Research and Therapy, 10:201-207, 1972. Marks, G., Richardson, J.L., Graham, J.W. and Levine, A. Role of health locus of control beliefs and expectations of treatment efficacy in adjustment to cancer. Journal of Personality and Social Psychology 51(2):443-450, 1986. McCusker, J. and Morrow, G. The relationship of health locus of control to preventive health behaviors and health beliefs. Patient Counseling and Health Education, 1:146- 150, 1979. Mehrens, W.A. A Standard of Quality: The Georgia teachers' certification testing program. Atlanta, Georgia Department of Education, 1985. Morris, L.L and Fitz-Gibbon, C.T. How to Measure Achievement, Beverly Hills, CA: Sage Publications, Inc., 1978. 91 National Academy of Sciences: Diet, Nutrition and Cancer Prevention. National Academy Press, 1982. Norusis, M.J. SPSS/PC for the IBM PC/XT. SPSS inc., Chicago, IL, 1984. Novelli, W.D. and Novelli, N.P.: From general audience to quiet pessimists: sharpening the focus in cancer audience segmentation. In Communicating Cancer Information: New Messages, New Approaches, and New Potential. Proceedings of the Fifth National Cancer Communications Conference. National Cancer Institute and American Cancer Society, 1984. O'Bryan, 6.6.: The relationship between an individual's I - E orientation and information-seeking, learning, and use of weight control relevant information. Dissertaion Abstracts International 33B:447B, 1972. O'Looney, B.A. and Barrett, P.T. A psychometric investigation of the multidimensional health locus of control questionnaire. British Journal of Clinical Psychology, 22:217-218, 1983. Pariza, M.W. Analyzing current recommendations on diet, nutrition and cancer. In Food and Nutrition News, National Livestock and Meat Board, Research and Nutrition Dept. 58 (1), Jan/Feb 1986. Pender, N.J. and Pender, A.R. Attitudes, subjective norms and intentions to engage in health behaviors. Nursing Research, 35(1):15-18, 1986. Rodin, J. Aging and health: effects of the sense of control. Science 233:1271-1276, 1986. Rokeach, M. The Nature of Human Values. New York. Free Press, 1973. Rosander, K. and Sims, L.8. Measuring effects of an affective-based nutrition education intervention. Journal of Nutrition Education 13(3):102-105, 1981. Rotter, J.R. Social Learning and Clinical Psychology. Englewood Cliffs, N.J.: Prentice-Hall, 1954. Rotter, J.B. Some problems and misconceptions related to the construct of internal versus external control of reinforcement. Journal of Consulting and Clinical Psychology 43:56-67, 1975. Saltzer, E. B. Locus of control and the intention to lose weight. In Wallston, R.A. and Wallston, B.S. (eds) Health Education Monographs: Health Locus of Control, 6(2):118- 92 128, 1978. Saltzer, E.B. Cognitive moderators of the relationship between behavioral intentions and behavior. Journal of Personality and Social Psychology, 41:260-271, 1981. Shepherd, R. and Stockley, L. Nutrition knowledge, attitudes and fat consumption. Journal of the American Dietetic Association, 87(5):615-619, 1987. Shipley, R.H. Maintenance of smoking cessation: effect of follow-up letters, smoking motivation, muscle tension, and health locus of control. Journal of Consulting and Clinical Psychology, 49(6):982-984, 1981. Straits, B. and Sechrest, L.: Further support of some findings about the characteristics of smokers and non - smokers. Journal of Consulting Psychology 27:282, 1963. Wallston, B.S., Wallston, K.A., Kaplan, G.D. and Maides, B.A.: Development and validation of the health locus of control (HLC) scale. Journal of Consulting and Clinical Psychology 44:580-585, 1976. Wallston, K.A., Wallston, 8.8. and Devellis, R. Development of the multidimensional health locus of control (MHLC) scales. In Wallston, K.A. and Wallston, B.S., eds. Health Education Monographs: Health Locus of Control, 6(2):161-170, 1978. Wallston, K.A. and Wallston, 3.8. Health locus of control scales. In Research with the Locus of Control Construct (Vol 11): Assessment Methods, Academic Press, pp. 189-243, 1981. Wallston, K.S. and Wallston, 8.8. Who is responsible for your health? The construct of health locus of control. In Sanders, G. and Suls, J. eds. Social Psychology of Health and Illness, Lawrence Erlbaum & Associates, Hillsdale NJ, pp.65-95, 1982. Wallston, K.A., What to do with the MHLC scores once you get them. March, 1983, unpublished paper. Wallston, B.S. and Wallston, K.A. Social psychological models of health behavior: an examination and integration. In Baum, A., Taylor, S. and Singer, J.E., eds. Handbook of Psychology and Health, Volume 4: Social Aspects of Health, Lawrence Erlbaum & Associates, Hillsdale, NJ, pp. 23-53, 1984. Williams, A.F. Personality and other characteristics associated with cigarette smoking among young teenagers. Journal of Health and Social Behavior, 14:374-380, 1973. 93 Winefield, H.R. Reliability and validity of the health locus of control scale. Journal of Personality Assessment, 46(6):614-619, 1982. APPENDICES APPENDIX A DIETARY RECOMMENDATIONS TO AMERICANS 0N DIET/NUTRITION/CANCER 1980 - 1985 Dietary Recommendations to Americans on Diet/Nutrition/Cancer-- 94 APPENDIX A 1980-1985 (Pariza, 1986). Senate 7m .4 dietary cl .mhmm pkhmm Aunts: cums Cums are" I984 thmu cums « hniuue 1984 «me mmbmms Rx Cancer Research 1982 NMAMK? came I982 MAS-INS! heathiul i980 dies Ammtm: any (human: cums fihmmb Health 1985 ‘ American diet! Imufifig. cums g0 - o . OM. 7119/8! USDA-m ”I. 3:: Rev. (highs 30% 30$ NR NR Mkdkm.mnw. d- wegs. rich in vitamins A, C; «mumuuwm Mm“€am.~uuflydihm 4. 0.3. On! ”momma. lat nave limit fiber “has 73 N" we: we yes NR NR NR NR Ye, 30: ad» NR NR NR . NR 7" 7'5 H mm H Ssh-see Nhumlh moderation air Maintain Right; exercise mom was: NR Vedeai less: his. oils, “0 35...: APPENDIX B CONSENT FORM 95 APPENDIX B Department of Food Science and Human Nutrition MICHIGAN STATE UNIVERSITY East Lansing, Michigan 48824 Informed Consent The purpose of this research project is to find out: (1) Can the "Taking Control” program developed by the American Cancer Society effectively teach people about reducing risk of cancer? (2) Does a person's ideas about whether they can control their health influence what they learn about cancer risk? People who agree to take part in this project will learn about nutrition and cancer risk. They will: (1) Answer a set of questions about their feelings about health just after this meeting today. (2) Attend a one- to two- hour program next week which will include slides, a talk and time for questions. (3) Fill out two (2) short sets of questions - one just before the slide show and one after the question period. Your participation in this project is completely voluntary and will not affect the services you obtain from the American Cancer Society or your position at work. This research program cannot give you the promise of avoiding cancer. This program involves no risk. I. . agree to take part in this project to test the ”Taking Control” program. I understand that participating does not promise me the possibility of avoiding cancer. I was allowed to ask questions about the program and the two sets of questions I will answer. I have been told that I may stop participating at any time without penalty. I understand that the study results will not in any way affect me as an individual. The data from individuals will be treated with strict confidence and my anonymity will be preserved. At the end of the study, I may ask for a copy of the summarized results if I want to. Investigator Signature Participant Signature Date Date 8/87/SRA APPENDIX C FINAL VERSION NCKT 96 APPENDIX C FINAL VERSION NCKT QUESTIONNAIRE OP NUTRITION KNOWLEDGE DIRECTIONS: Choose the one answer which best completes the question. Then, with a No. 2 pencil, blacken the circle on your answer- sheet which corresponds to your answer. PLEASE DO NOT USE ink, ballpoint, or felt-tip pens. IT IS IMPORTANT TO GIVE ONE ANSWER TO EACH QUESTION. THANK YOU FOR YOUR HELP IN OUR STUDY! 1. Which one of the following statements is true ? a. It has NOT been proven that cigarette smoking is the main cause of lung cancer. b. For American women, lung cancer has overtaken breast cancer as the number one cause of cancer death. , c. Ninety percent of all cancer is smoking related. d. In the United States, the rate of cigarette smoking in adults has been steadily rising. e. Compared to pregnant women who don't smoke, pregnant women who smoke are more likely to have babies who will grow up to get cancer. 2. One advantage to eating foods like broccoli and sweet potatoes is that they are: a. a concentrated source of calories for energy. b. high in beta carotene, a form of vitamin A. c. ab e to protect you from skin cancer. d. able to lower your risk of leukemia. e. cruciferous (cabbage-family) vegetables. 3. Which one of the following foods is a poor source of fiber? a. fish b. tomatoes c. potatoes d. beans e. oatmeal 4. Which one of the following is a cabbage-family (cruciferous) vegetable? a. spinach b. parsley c. cauliflower d. lettuce e. green pepper 9. Which one a. b. c. d. e. of the following foods is now in beta carotene? apricots broccoli peaches squash oranges Of the following foods, which one is the best source of both vitamins A and C? a. b. c. d. e. orange apricot broccoli carrot cauliflower Eating vegetables from the cabbage (cruciferous) family is important a. b. c. d. because eating these vegetables may: reduce the risk of cancer of the colon and stomach. reduce the risk of breast cancer. increase the amount of iron absorbed from the diet. be the best source of herbicide and pesticide free vegetables. control the amount of toxins one's body produces. Eating foods which are high in fiber is important because dietary fiber: Which one vitamin C? a. b. c. d. e. may decrease your risk of getting cancer of the colon. may decrease your risk of getting cancer of the lungs and larynx. is one of the highest quality, naturally potent and balanced food components available today. is an essential nutrient. contains live enzymes necessary to trigger bodily functions which protect you from cancer. of the following foods is a poor source of oranges tomatoes broccoli green peppers apples 98 10. An increased risk of cancer of the breast and colon is linked to: 0000' being overweight. high dietary cholesterol levels. being ”out of control” of yourself. low triglyceride levels in the blood. over-consumption of sugar. ll. Of the possible cancer risk factors that can be controlled by an individual, the most harmful one is: 0000'. excessive sun exposure. eating too much fat. smoking tobacco. eating nitrite-cured foods. all risk factors are equal. 12. Which one of the following statements is true? A high intake of fat has been linked with increased risk of lung cancer. A high intake of fat has been linked with increased risk of colon cancer. Americans get about 20 percent of their calories from fat, whereas a healthy diet would include a maximum of no more than 10 percent fat. Being underweight increases your risk of getting cancer. Eating more saturated fats than unsaturated fats increases your risk of getting cancer. 13. Which one of the following statements is true? The sun's rays are most damaging in the late afternoon. The only way to lower your risk of getting skin cancer is to stay out of the sun. People who tan easily do not get skin cancer. A sunscreen with an SPF of 8 is best for protection from the sun's harmful rays. Some exposure to sunlight is necessary for good health. ' 14. Which one of the following would most likely decrease your risk of getting cancers of the esophagus, larynx and lung ? a. b. c. d. e. Jog three miles every day. Eat plenty of whole-grain, high-fiber foods. Eat plenty of foods containing beta carotene. Avoid over-exposure to the sun. Abstain from alcohol. 99 15. The American Cancer Society recommends that one restrict consumption of smoked, salt-cured, and nitrite-cured foods (such as bacon, ham and hotdogs) because they are a. gesponsible for most of the fat in the American et. b. used as preservatives at an alarmingly high rate in the United States. c. associated with an increased risk of lung cancer. d. associated with an increased risk of stomach and esophageal cancer. e. high in salt. 16. Which one of the following statements is true for alcohol? a. To avoid getting cancer of the liver, alcohol must be removed from the diet. b. Drinking alcohol and smoking cigarettes is worse when one does both than either one alone. c. Moderate drinking of alcohol lowers your risk of getting cancer of the liver and pancreas. d. Drinking alcohol is the single most preventable cause of cancer in the United States today. e. Alcohol is linked to more than 30‘ of cancer in Americans. 1?. Which one of the following risk factors has been associated with an bladder a. b. c. d. .0 increased risk of cancers of the uterus, gall and colon? I Not eating enough foods high in beta carotene. Consumption of alcohol. Eating to many foods high in nitrates. Over-exposure to the sun. Being obese. APPENDIX D FACILITATOR'S GUIDE OF THE AMERICAN CANCER SOCIETY TAKING CONTROL PROGRAM 100 APPENDIX D Facilitator's Guide of the American Cancer Society Taking Control Program PURPOSE The purpose oi the American Cancer Society's TAK- NG CONTROL program is to provide an htroduction b a healthy. enjoyable lilestyle that may reduce one's risk ol developing cancer. The program gives an overview oi live " otective iactors" against can- cer and live preventa “risk iactors" as well. A cancer risk iactor is anything that makes it more likely one may develop cancer. A pram lactor onthectherhand. lsanythingknoumblassanthe Ikeiihood oi develong the disease. The TAKING CONTROL program helps people think about health in terms oi their overall lilestyies. TAK- INGCONTROLispositive. optirnisticandactim oriented. CONTENT BIKING CONTROLgoescnrecordwiththegcod news that scientists now believe most cancers may be related to liiestyle and environment. The TAKING CONTROL program aims also to debmk the myth that “everything causes cancer. " TAKING CONTROL specifically addresses live pro- lective iactors: a Cabbage-tamiiy vegetables (cnrciierous) a Fiber a Vitamin A a Vitamin C a Weight control md five risk iactors: a High-tat diets s Salt-cured. smoked and nitrite-cured ioods aTobacco a Alcohol s Excessivesmexpcsure The program also covers other cancer-related tac- . brs.includingexcessiveexpcsureterays.astro- gensandknoancarcinogensinthewcrkplace. Theconceptolrisk-isctorreducticnukeshtoccn- sidel’aiionlheMowing: aNotalicancerrhkiactorsareequalhknpor- tance. sSome risk iactors make each other worse whencombined. anterearecertahriskiacicrsyoucannotcon- trol. such as imniiy history. prior history oi cancer. etc. Program participants are encouraged to take control by adding to their everyday lifestyles the protective habits. and eliminating the other habits that increase their risk at developing certain icnns oi cancer. They areremindedoitheimportanceoibeingmoreeware oi their bodies and havhg regular cancer-related checkups tor early detection oi the disease. The TAKING CONTROL coordinated packagepi ma- mrials includes: a film a Videotape (‘6- and tin-inch) a Slide-tape presentation a Flip chart a Sell-help booklet a Foster a Take-home iolder s Administrator's guide From: Taking Control - Public Education Program on Cancer Prevention-Risk Reduction. Facil itator' s Guide. American Cancer Society. inc. lit. 1985. APPENDIX E CONTENT REVIEW - ACS VOLUNTEERS 101 APPENDIX E CONTENT REVIEW - ACS VOLUNTEERS August 12, 1987 Dear American Cancer Society Program Presenter, Your name was given to me as someone who is familiar with the "Taking Control” program of the American Cancer Society (AC8). I am doing an evaluation of this program for my thesis research at Hichigan State University. As part of this evaluation, 1 have developed a questionnaire to give to people who have seen “Taking Control“. As a part of the testing of this questionnaire, 1 need to get input from presenters like you who have experience with the program. Could you please take a few moments to look over the questionnaire? I've included a list of things to think about while you're reviewing the questions. Thank you for your help. Your assistance in the test development process will improve the evaluation of the "Taking Control" program. With the information that we obtain from this evaluation we hope to continue to improve the programs provided by ACS. Sincerely, Sally R. Anger, AC8 program presenter and MSU graduate assistant PLEASE RETURN OUESTIONNAIRES TO: asses raerauo c/o rue ansatcau camera scorer! rucuan couurr UNIT SUITE 104 416 reasons sass annexes, HI 40912 102 OBJECTIVES Please match each question on the questionnaire with the objective you feel best matches it from the list below. 1. To test the recognition of food sources of the following protective factors: a. cabbage-family vegetables b. fiber c. vitamin A d. vitamin C 2. To test the understanding of the reasons why the following protective factors are important: a. cabbage-family vegetables b. fiber c. vitamin A d. vitamin C e. weight control 3. To test the recognition of the following risk factors: a. high-fat diets b. salt-cured, smoked and nitrite-cured foods c. tobacco d. alcohol e. excessive sun exposure 4. To test understanding of the reasons why the following risk factors are important: a. high-fat diets . b. salt-cured, smoked and nitrite-cured foods c. tobacco d. alcohol e. excessive sun exposure 5. To test the awareness of the 7 warning signs of cancer (from the Cancer Related Checkups brochure used in conjunction with the “Taking Control” program). 103 GUIDELINES FOR TESTING THE QUESTIONNAIRE (1) Do the questions adequately test the concepts presented in the “Taking Control“ program? (2) Does the test cover information actually contained in the program? if so, do you agree with the emphasis placed on the concepts presented? (Please explain). (3) Does the test cover all of the concepts you feel are important? if no, which conceptis) do you feel was not adequately represented? Were any concepts over-tested? (4) Please match each question to the objective you think it is trying to measure. (See attached list of objectives). Do you think these objectives correctly represent the “Taking Control" program? Please feel free to give any additional comments: APPENDIX F CHECKLIST FOR MULTIPLE CHOICE ITEMS 104 APPENDIX F CHECKLIST FOR MULTIPLE CHOICE ITEMS Questions Concerning Multiple Choice Items 0 Is a multiple choice format inappropriate to the type of information desired? 0 I: the language of the item complex or unclear. perhaps containing unfamiliar vocabulary? Do items have more than one correct answer? Do item: concern several central issues at once rather than a single one? Does the stem contain negative statements which have not been emphasized? Do responses or choices come at the beginning or the middle of the stern rather than at the end where they belong? Are response alternatives grammatically inconsistent with the stem? Are response alternatives inconsistent with one another? 0 Are some response alternatives implausible or easily eliminated by students who lack the information tested by the item? Are the response alternatives arranged in an illogical or confusing order? 0 Are the response alternatives interdependent or logically overlapping? Do the response alternatives include extraneous clues due to yammatical inconsistencies or length of response? Are the correct answers. for instance. generally longer than the incorrect ones? 0 is the none of these option used inappropriately? Across items, is the correct answer among alternatives likely to be found in the same position much of the time? For instance. is the correct alternative nearly always in the middle position? From: How to Measure Achievement. by Lynn Lyons Norris and Carol Taylor Fitz-GiEBEn. Sage Publications, Beverly Hills/London l978. p. 154. APPENDIX C CONTENT REVIEW — MSU NUTRITION FACULTY 105 APPENDIX C CONTENT REVIEW - MSU NUTRITION FACULTY MICHIGAN STATE UNIVERSITY ovumamuxnnouana«woumuuumummu umwnumcaflxngm.‘mwuu November 20, 1987 HEHORANDUN 70: Nutrition Faculty FROH: Sally Anger RE: Nutrition/Cancer Test Validation As most of you know, I am still in the stages of validating a nutrition/cancer risk reduction test that l have developed to measure the change in knowledge of subjects in my thesis study. I have enclosed a copy of the most recent draft, along with a brochure from the ”Taking Control“ program which describes most of the concepts put forth in the program. The recommendations listed in the enclosed brochure will be provided to subjects during a slide/tape presentation and question/answer session. 1 am asking you to give me feedback as to whether you feel the questions will adequately assess knowledge of nutrition's role in cancer. in other words, are the questions valid as pertains to this program? Does the test actually test these concepts? Any comments or suggestions you may have would be greatly appreciated. Thank you for your help. USUt-MMMW} yum-u.- 106 Draft of NCKT accompanying Faculty Letter (prior to first pilot test) and accompanying letter to the ACS volunteers. QUESTIONNAIRE OF KNOWLEDGE OF NUTRITION/CANCER RISK REDUCTION DIRECTIONS: Choose the one answer which best completes the question. Then, with a No. 2 pencil, blacken the circle on your answer- sheet which corresponds to your answer. DO NOT USE ink, ballpoint, or felt-tip pens. lT is lHPORTANT THAT you ANSWER ALL QUESTIONS. THANK YOU FOR YOUR HELP IN OUR STUDY! 1. Which of the following is thought to INCREASE one's chance of getting cancer? a. vitamin A b. salty diets c. diets high in refined sugar * d. high-fat diets e. low-fat diets 2. What is one advantage to eating foods like carrots and sweet potatoes? a. they are low in dietary fiber and therefore protect you from cancer. * b. they are high in beta carotene. c. they help you to tan and therefore protect you from skin cancer. d. they are cruciferous vegetables. e. they are good sources of energy. 3. Which of the following foods are good sources of fiber? * a. vegetables. b. fish and poultry. c. lowfat dairy products. d. red meats. e. vegetable oils. 4. Which of the following vegetables belong to the cabbage- family (cruciferous) vegetables? a. spinach b. parsley * c. cauliflower d. lettuce e. all of the above. S. 107 Which of the following foods (is/are) high in beta carotene? a. apricots b. broccoli c. peaches d. squash e. All of the above. Of the following foods, which is a the best source of both vitamin A and vitamin C? a. orange b. apricot c. broccoli d. carrot e. cauliflower Eating vegetables from the cabbage (cruciferous) family is important because eating these vegetables may: a. reduce the risk of cancer. b. contain substances which help you ”burn off" fat. c. increase the amount of iron absorbed from the diet. 6. increase the amount of protein available from the diet. e. all of the above. whole grain breads, whole grain flours and cereals, dried peas, beans and most fruits and vegetables are all good sources of: a. fiber b. beta carotene c. ascorbic acid d. vitamin A e. protein Which of the following foods (is/are) a good source of vitamin c ? a. oranges b. tomatoes c. broccoli d. green peppers e. all of the above 10. 11. 12. 13. 14. 108 An increased risk of cancer of the breast and colon is linked to: a. being overweight. b. a hectic lifestyle. c. being ”out of control” of yourself. d. emotional stress. e. high blood pressure. Of the possible cancer risk factors that can be controlled by an individual, the most harmful one is: . excessive sun exposure. eating too much fat. smoking tobacco. eating nitrite-cured foods. all risk factors are equal. (ID-00’“ Which of the following statements is true? a. a low-fat diet increases your chances of getting cancer. . b. the most common sources of fat in the American diet are meat, dairy products, pastry and chocolate. c. Americans get about 30 percent of their calories from fat, whereas a healthy diet would include a maximum of no more than 10 percent fat. d. being underweight increases your risk of getting cancer. Which of the following statements {a true? a. Excess sun exposure rarely causes skin cancer. b. Indoor sunlamps and tanning parlors are much less harmful to the skin than the sun. c. if you see changes in a mole or a sore that does not heal, it may be a sign of skin cancer. d. A sunscreen with an SPF of 8 is best for protection from the sun's harmful rays. e. People who tan easily are at no risk of skin cancer. It is recommended that you eat LESS smoked, salt-cured, nitrite-cured foods because a. their high sodium content is a risk factor for most cancers. b. Americans consume too much of them.‘ c. they are foods which are usually eaten by overweight people. d. they are foods which are associated with an increased risk of cancer. e. they lower your resistance to infection. and 15. 16. 17. 18. 109 Approximately 30\ of all cancer is related to: I. b. c. d. e. low fiber intake. bumps and bruises cigarette smoking. obesity. excessive nitrate consumption. Compared to pregnant women who don't smoke, pregnant women who smoke are: a. b. c. d. more able to keep.their weight down to a reasonable level. more prone to miscarriages and their babies are more likely to have birth defects. more likely to die at time of delivery than are nonsmokers. greatly protected by using cigarettes with filter tips. more likely to have twins. If you are a smoker, a major health priority should be to: stop smoking. increase your consumption of foods high in vitamins A and C to help protect against the effects of smoking on your lungs. look for signs of cancer to detect it early. strictly control other cancer risk factors. control your weight. Which of the following statements is true for alcohol? a. b. c. d. To avoid getting cancer of the liver, alcohol must. be removed from the diet. Drinking alcohol and smoking cigarettes is worse when combined than either is by itself. Moderate drinking lowers your risk of getting cancer of the liver and pancreas. Drinking alot of water after a night of heavy drinking will greatly decrease your risk of getting liver cancer. None of the above statements are true. APPENDIX H SCORING OFFICE RESULTS - FIRST PILOT TEST 110 APPENDIX H SCORING OFFICE RESULTS - FIRST PILOT TEST . - - ............ .. ..... Mvmmfi. ...... E wsmas-mmuu-ub-fiomflm-mean-1&7.- - - - - - - - - - - - - - - - - - - - r uuuuuuuuuuuuuuuuuuuu n-o-. w.- - - - bma- wt“. pout-um gMomc-zmwc- «new: uuuuuuuuuuuuuuuuuu - - .m- zoacazaxaaumaa zaps 2mm“ s ................ .- - - - - - - em ........... capstan 3c .5... .. - - - ............... f ........................................... - - - - - - - - ' ....... - - ' o o 00 z«:» anus o . o. . oo o o o. - oo. . ...... mu ........ m ..... o m---.-. lllllllllllllllllllll @----o.---0m--.-.-- .. a on . .u a. « ea . .u . ...... mo ........ m lllll o N-J.-o uuuuuuuuuuuuuuuuuuuuu o. ccccc --o.--bm--.-n-- u . on - .c o . on . .c o o oo - .m- o. m as - .n - o . oh - .c u. a o» . .o .--..-m -------- w ----- o .o--.-a. --------------------- m. -------- m---0m--.-~.-- o o om . .u a. n so . .o . ------ o -------- o ----- on ...--00 --------------------- mu -------- v- - - - - bow-b? - - - was». new». moatzaumua to names: uuahzuuaua to «more: a - Mme-62.738 ram c... ...o-a mo- mo- amino-a Emmo- ------------- m manaf twee-Ewan huh-- mo- was-fl -u- km -o- - r IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII ------‘I--Is'---------‘Uav- ¢¢§w umtoummu tub— ------ .----.-- DISC 33 STUDENTS OIFF 7! LOVE! 2 I1 IIOOLE l0! UPPER 27$ '°”"""'"""'°""‘3flbfifi§?i Ir tress o-ou-on--Lo-o-u--n--t-o-u-~n--L-o-u--a--L-o-u--n--L--ou--n-oL-o-u-o-n-oL-o "Ifti'itfv":""t""’0'"'3""3""!""0"”‘4""'o""S""-'"'bil‘t""0"°f SYINDIID POI. I I oopop I I I I O... 0.. sense I I I I oopnp ' I I oopop I I e at. 000 °eb°b I I oabom I I. I I oomom I I" I I a a.a 0.0 H t I -.:L°:l8 I I 0 .IO NIH fi . CD..- I ’0 0 00 mm ~ “can" oohoh I. I I I a... I" I a o'a a'a I I wInh-nhn I I I I 0 0|0 0|0 I I I I 00.00.0 I I I I O 0.0 0.0 00-0 0-0 I I I I oopop I I oopop I I a ale 0'. camkaefo I" N' I I IIOWDvMs I "I I I 0 0.0 0.0 I I I I O 0 .0 0 .0 0 0I0 0-0 I I I I 0 0,0 0,0 I I oopop I I a Mo a Is I I 8 0.0 0.8 I I flgoksoh- I IQ I I 0 OIO 000 I I I I a a ,s a ,a 0 mm I'I‘O «I- “I I I 0 0I0 0I0 I "I I I 00,0 8,0 "0 «III .I "I0 i u ---?7.----------- 9‘ I (as: 9_O I00 .1! - -... :..L: ...: 112 Draft of NCKT used for first pilot study. QUESTIONNAIRE OF NUTRITION KNOWLEDGE DIRECTIONS: Choose the one answer which best completes the question. Then, with a No. 2 pencil, blacken the circle on your answer- sheet which corresponds to your answer. DO NOT USE ink, ballpoint, or felt-tip pens. IT IS IHPORTANT THAT YOU ANSWER ALL QUESTIONS. THANK YOU FOR YOUR HELP IN OUR STUDY! 1. Which of the following is thought to INCREASE one's chance of getting cancer? a. vitamin A b. salty diets c. diets high in refined sugar d. high-fat diets e. low-fat diets 2. One advantage to eating foods like carrots and sweet potatoes is that they are: a. low in dietary fiber. b. high in beta carotene. c. able to protect you from skin cancer. d. cruciferous vegetables. e. good sources of energy. 3. Which of the following foods is a good source of fiber? a. vegetables. b. fish and poultry. c. low-fat dairy products. d. red meats. e. vegetable oils. 4. Which of the following are cabbage-family (cruciferous) vegetables? a. spinach b. parsley c. cauliflower d. lettuce e. green peppers S. 113 Which of the following foods is LOW in beta carotene? a. apricots b. broccoli c. peaches d. squash e. oranges Of the following foods, which is the best source of both vitamins A and C? a. orange b. apricot c. broccoli d. carrot e. cauliflower . Eating vegetables from the cabbage (cruciferous) family is important because eating these vegetables may: a. reduce the risk of cancer. b. help one to ”burn off“ fat due to compounds they contain. c. increase the amount of iron absorbed from the diet. d. increase the amount of protein available from the diet. e. control the amount of toxins one's body produces. Whole grain breads, whole grain flours and cereals, dried peas, beans and most fruits and vegetables are all good sources of: a. fiber b. beta carotene c. ascorbic acid d. vitamin A e. protein Which of the following foods is a poor source of vitamin c ? a. oranges b. tomatoes c. broccoli d. green peppers e. apples 10. ll. 12. 13. 14. 114 An increased risk of cancer of the breast and colon is linked to: a being overweight. b. high dietary cholesterol levels. c. being ”out of control” of yourself. 6. low iron levels in the blood. e high blood pressure. Of the possible cancer risk factors that can be controlled by an individual, the most harmful one is: . excessive sun exposure. eating too much fat. smoking tobacco. eating nitrite-cured foods. all risk factors are equal. (IO-OU'D‘ Which of the following statements is true? a. a low-fat diet increases your chances of getting cancer. b. the most common sources of fat in the American diet are meat, dairy products, pastry and chocolate. c. Americans get about 30 percent of their calories from fat, whereas a healthy diet would include a maximum of no more than 10 percent fat. d. being underweight increases your risk of getting cancer. Which of the following statements is true? a. Excess sun exposure rarely causes skin cancer. b. Indoor sunlamps and tanning parlors are much less harmful to the skin than the sun. c. If you see changes in a mole or a sore that does not heal, it may be a sign of skin cancer. d. A sunscreen with an SPF of S is best for protection from the sun's harmful rays. e. People who tan easily are at little risk of skin cancer. The American Cancer Society recommendeds that one restrict consumption of smoked, salt-cured, and nitrite-cured foods (such as bacon, ham and hotdogs) because they are a. low in fiber. b. responsible for most of the fat in the American diet. c. usually eaten by unhealthy people. d. associated with an increased risk of cancer. e. believed to lower one's resistance to infection. 115 15. Which of the following statements is true for alcohol? To avoid getting cancer of the liver, alcohol must be removed from the diet. b. Drinking alcohol and smoking cigarettes is worse when one does both than either one alone. c. Moderate drinking lowers your risk of getting cancer of the liver and pancreas. d. Drinking alot of water after a night of heavy drinking will greatly decrease your risk of getting liver cancer. e. Alcohol is linked to more than 30% of cancer in Americans. 16. Compared to pregnant women who don't smoke, pregnant women who smoke are: a. better able to keep their weight down to a reasonable level. b. more prone to miscarriages and their babies are more likely to have birth defects. c. more likely to die at time of delivery. d. greatly protected by using cigarettes with filter tips. e. more likely to have babies who will grow up to get cancer. 17. If you are a smoker, a major health priority should be to: a. stop smoking. b. increase your consumption of foods high in vitamins A and C to help protect against the effects of smoking on your lungs. c. look for signs of cancer to detect it early. d. strictly control dietary cancer risk factors. e. control your weight. Please STOP here. Do NOT continue with the second half of the questionnaire until instructed; after today's presentation. It is very important that you NOT look at the next section. Thank you for your cooperation! APPENDIX I SCORING OFFICE RESULTS - SECOND PILOT TEST 116 APPENDIX I SCORING OFFICE RESULTS - SECOND PILOT TEST ............................ ES-----.Eamemma-e.sae.335------------.-----..----------- ............................ 5e----RCEEE.cans-35.3-2--------------------------- B E _ ................................ writ-----.cfiaes3:33}-.-.-.-..---.-..-------.---- o o E3 . . O o o. I 8 a O o. I a ............... ell----...----.om......-------------------lolllla.---.om..-...-.------------- a. a on I pm on e an I a ............... e.-------m----.8...s.-------------...--lb...it---».---.3..-.....------------- n. a an" a n ma... 0n n 2. I o. 0 0 Oh I .0 ........... ..--e.-------m----.8.......--------------------n-l-----p.---3..-..r.--.---------- o .. 00- a. 0 6 SI .- .............. s-.------s- --.853------:------------c--------a----3...-s.-------------- may: «as: .... as: E ......... neoprsafiwaasfih3_§eaerp------------heme.t...B§a.-¥..$3§.cxp------------ ............................ mh‘mwm-I--M---m..mu'~m'----'-'-------.-------' 40 STUDENTS r’”6fs<’:"“” 01-? I LOVE! 27’ 2O IIDDLE 40% IO UPPER 27$ """""'Sftibenrs Pea scour? "‘ifr‘i'kfv”-'""I'""o""‘i""'3""5'"'e""'4""‘o""S"",'o'",bit'r""o”‘in'h6 PEICENTAOES I1 ITENS STANDARD EDI. 117 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I9 93 8:2 9:3 8:8 2:8 2:8 IS 2; I I Eh! I I I I I I :fl:’ I I I I I I hula-I I I I I “I In l I I I I we cs.e 322:3 press a I l I I I I I I l I I I I I I I I I I I I I I I I I I . I e a.e e.e e e eIe e-e a a e-a e-e a e 0.0 I CICNCIIDTD CICYCICVC> CICNCICNCD ¢>¢>'° I I I l I I Id I I I I I I I: CDCMCICMCD IDEDMDCMC> CICMCICMCI CICNI I I I I I I N! I I I I I I II CIChCICMCI CIGDJDIMCI CICMC>CMC> IICMI : I I I I I I Is 0 0:. 0:. O 0 0:0 0:0 0 0 0:0 0:. C C 0:; CICMCICMCD EDEDKDCDKD GDCDNDEDI CI -0 I I I I' 'I 'fia "stf Edodo bobob oobob 00% I I I I I I I: oobob oobob ododo od- I I . I I I I u: I I I I I I I' I a ale ale a a ale aIe a e eIe ale 0 a eIe E RSPEB oobgb go'oos ooh I I wk! Iv 'f I I I I l I l I I ner°e 2°92? esseL sea I I I I I I I, CI 'C) CD! Ci I I 2.°s D-eee sets? as a 0:0 0:. a e e:a e:e e a e:e e:e a a 0.: 00200 0008b moeog 00“ I VT“ " r' I I" I ..I I I I I I I 'I: C’E’N’Ifldfl CICMCICMII CICM‘ICMCD C’E’II ' I I" I '3" I I”’ , I I I I I I II oqogo ooegp ooeoe °°6 I l I I I n- II a 0's 0:. e a .I. e:a e a 0:0 a:a a e 0:; I 9°93? 82P°£ 89??? 999 I I I I I I I' agnag 'npgp capgp 8% I I I I I I I: I I I I I I I C)“f’:2§> ‘DE’EBfD ChEI‘BfD Ii! I I I I I I I: DOI. CID. OCIO OIOODOII OICOOCI: I I I I I I 339°: °°.°e.s sees as I I I I I' O I I I “I I I 3 3.3 °3 °-° '13 v.3 ”-9 3.? I I I I I I $8000 omoo oggmo m: I I I 'I - u I I I I I I I. .... .... ......C . .... .... ...: moo o-oo om mo .- Sasafl‘ I" 13;: f’ Ii;"I Ff I I I I I I ¢M£BSBJ> C’:p€g‘>fi9 C’Shszsad” 'N' I I I- I I I II 2°32? 8&9? oobob o I I I I II a e:a a'a a a a'e 0:0 0 a a:e a:a a a a'e I onu-«II-o-Luoo-uu-II--I.-o-u--II--L-o-u--II--I.-o-u--II--L-o-u--n--I.-.o-u--II-urn- «uknh n—Fm' mwmwe Nd I I I I I -~msm pep-9 : ye 3g #2- I I I I I I I . I I I I I I I I I I I I I I I I I I I I I - .-------------------------------------------------------------—-------------------- - - .-------------------------------------d-------fi---------------q-------------------. ------ .---COCO-----------------—-------------------------------------------—----—-------------. - . -- .---------------------------------------------------------C--C‘-----------------------QDOO. - - - .--—------C----------------------------------------------------Q-----n------------------. - C - 118 Draft of NCKT (used in pilot test 2). QUESTIONNAIRE OF NUTRITION KNOWLEDGE DIRECTIONS: Choose the one answer which best completes the question. Then, with a No. 2 pencil, blacken the circle on your answer- sheet which corresponds to your answer. PLEASE DO NOT USE ink, ballpoint, or felt-tip pens. IT 18 IHPORTANT TO GIVE ONE ANSWER TO EACH QUESTION. THANK YOU FOR YOUR HELP IN OUR STUDY! 1. Which one of the following statements is true ? a. It has NOT been proven that cigarette smoking is the main cause of lung cancer. * b. For American women, lung cancer has overtaken 6 breast cancer as the number one cause of cancer death. c. Ninety percent of all cancer is smoking related. d. In the United States, the rate of cigarette smoking in adults has been steadily rising. e. Compared to pregnant women who don't smoke, pregnant women who smoke are more likely to have babies who will grow up to get cancer. 2. One advantage to eating foods like broccoli and sweet potatoes is that they are: a. a concentrated source of calories for energy. b. hi h in beta carotene, a form of vitamin A. (::> c. ab e to protect you from skin cancer. d. able to lower your risk of leukemia. e. cruciferous (cabbage-family) vegetables. 3. Which one of the following foods is a poor source of fiber? a. fish b. tomatoes (::> c. potatoes d. beans e. oatmeal I. which one of the following is a cabbage-family (cruciferous) vegetable? a. spinach b. parsley c. cauliflower d. lettuce e. green pepper * Table of Specifications Code (please see Table} for explanation). S. which one b. =- a. 119 of the following foods is now in beta carotene? apricots broccoli peaches squash oranges 6. Of the following foods, which one is the best source of both vitamins A and C? O 23 0 2: orange apricot broccoli carrot cauliflower 7. Rating vegetables from the cabbage (cruciferous) family is important b. @ «=- d. because eating these vegetables may: reduce the risk of cancer of the colon and stomach. reduce the risk of breast cancer. increase the amount of iron absorbed from the diet. be the best source of herbicide and pesticide free vegetables. control the amount of toxins one's body produces. 8. Eating foods which are high in fiber is important because dietary fiber: ‘0 b. @ C. de e. may decrease your risk of getting cancer of the colon. may decrease your risk of getting cancer of the lungs and larynx. is one of the highest quality, naturally potent and balanced food components available today. is an essential nutrient. contains live enzymes necessary to trigger bodily functions which protect you from cancer. 9. Which one of the following foods is a poor source of vitamin C? a. oranges b. tomatoes c. broccoli d. green peppers apples 120 10. An increased risk of cancer of the breast and colon is linked to: b. @ c. 6. being overweight. . high dietary cholesterol levels. being “out of control“ of yourself. low triglyceride levels in the blood. over-consumption of sugar. 11. Of the possible cancer risk factors that can be controlled by an individual, the most harmful one is: I. b. =- d. e. 12. Which C. b. ‘. c. excessive sun exposure. eating too much fat. smoking tobacco. eating nitrite-cured foods. all risk factors are equal. one of the following statements is true? A high intake of fat has been linked with increased risk of lung cancer. A high intake of fat has been linked with increased risk of colon cancer. Americans get about 20 percent of their calories from fat, whereas a healthy diet would include a maximum of no more than 10 percent fat. Being underweight increases your risk of getting cancer. Eating more saturated fats than unsaturated fats increases your risk of getting cancer. 13. Which one of the following statements is true? b. s: The sun's rays are most damaging in the late afternoon. The only way to lower your risk of getting skin cancer is to stay out of the sun. People who tan easily do not get skin cancer. A sunscreen with an SPF of 8 is best for protection from the sun's harmful rays. some exposure to sunlight is necessary for good health. ' 14. Uhich one of the following would most likely decrease your risk of 69 ES getting cancers of the esophagus, larynx and lung ? Jog three miles every day. Eat plenty of whole-grain, high-fiber foods. Eat plenty of foods containing beta carotene. Avoid over-exposure to the sun. Abstain from alcohol. 121 15. The American Cancer Society recommends that one restrict consumption of smoked, salt-cured, and nitrite-cured foods (such as bacon, ham and hotdogs) because they are a. sesgonsible for most of the fat in the American e . b. used as preservatives at an alarmingly high rate in the United States. (::) c. associated with an increased risk of lung cancer. d. associated with an increased risk of stomach and esophageal cancer. - e. high in salt. 16. Which one of the following statements is true for alcohol? a. To avoid getting cancer of the liver, alcohol must be removed from the diet. b. Drinking alcohol and smoking cigarettes is worse when one does both than either one alone. c. Moderate drinking of alcohol lowers your risk of getting cancer of the liver and pancreas. d. Drinking alcohol is the single most preventable cause of cancer in the United States today. e. Alcohol is linked to more than 30‘ of cancer in Americans. 17. Which one of the following risk factors has been associated with an bladder @ ES increased risk of cancers of the uterus, gall and colon? Not eating enough foods high in beta carotene. Consumption of alcohol. Eating to many foods high in nitrates. Over-exposure to the sun. Being obese. Please STOP here. Do NOT continue with the second half of the questionnaire until instructed; after today's presentation. It is very important that you NOT look at the next section. Thank you for your cooperation! APPENDIX J QUESTIONNAIRE ON HEALTH BELIEFS 122 APPENDIX J QUESTIONNAIRE ON HEALTH BELIEFS (Includes Wallston's HLC, Cancer deus of Control Test, and Demographic Questions). We lakine Control Droeram of The American Cancer Society and M'chigan State University Program Presenter: Sally Anger I... when [I]: This questionnaire contains a number of questions that ask you about your beliefs about health and cancer issues. Please follow the instructions carefully. It is important that you answer all of the questions as honestly and as accurately as you can. Some of the questions ask you for personal information; it is important that you remember that this questionnaire is completely confidential. Please do not write your name anywhere on the questionnaire. However it would be helpful to us if all of the information we receive from you is marked with one code number. This code number will be known only to you. Please fill in the code number section in the upper right hand corner of the page by designing a personal code using the 2 digit number which represents the day of your birth, the 2 digit number which represents the month of your birth, and the last 2 digits of your telephone number. For example: If your birthday is March 11, and your telephone number ends in 89, then your code number is: [11 To: 189] r i ~. day}7 month last 2 digits of birth of birth of telephone number On the day of the American Cancer Society program, everyone will be given a personal packet according to their code number. Thank you for participating! 123 Code Number _/_/_ oussronIIIaxes oII IIIALTII semen * Directions: Please write your COD! NUHIIR on each sheet. Please do not write your name on this questionnaire. This is a questionnaire designed to determine the way in which different people view certain important health related issues. Each item is a belief statement with which you may agree or disagree. beside each statement is a scale which ranges from strongly disagree (I) to strongly agree III. for each item we would like you to circle the number that represents the extent to which you disagree or agree with the statement. The more strongly you agree with a statement, then the higher will be the number you circle. The more strongly you disagree with a statement, then the lower will be the number you circle. Please make sure that you answer every itemiand that you circle gal! gng,number per item. This is a measure of your personal beliefs; obviously, there are no right or wrong answers. Please answer these items carefully, but do not spend too much time on any one item. As much as you can, try to respond to each item independently. Ihen making your choice, do not be influenced by your previous choices. It is important that you respond according to your actual beliefs and not according to how’you feel you should believe or how you think we want you to believe. .5. 0‘0 0 5:5.5. 3 II 3 3 Is 3 3.32.: o-I e-I b.5505 dads-IBM .flfluafl Has-.3: 33:32:: I. If I get sick, it is my own behavior which determines how soon I get well again. I 2 3 I s 6 2. No matter what I do, if I am going to get sick, . I will get sick. 1 2 3 I 5 f 3. having regular contact with my physician is the best way for me to avoid illness. 1 2 3 I 5 6 I. lost things that affect my health happen to me . by-accident. l 2 3 I 5 6 5. Whenever I don't feel well, I should consult a medically trained professional. 1 2 J I 5 6 f. I am.in control of my health. I 2 2 I 5 I * Questions 1 - 18 from Wallston & Wallston Questions 19 - 24 = Cancer Locus of Control Questions 25 - 33 = Demographic Questions 124 Code Number __/ L l :i: 539;. 3 :I e g? 'g 2:: a aaaaga Stfo‘éos agar”: a. «mi!» 1. fly family has a lot to do with my becoming sick or staying healthy. I 2 3 I 5 6 I. When I get sick, I am to blame. I 2 3 I 5 6 3. Luck plays a big part in determining how soon I will recover from an illness. 1 2 3 I 5 6 10. health professionals control my health. 1 2 3 I 5 6 ll. Ny good health is largely a matter of good fortune.l 2 3 I 5 6 12. The main thing which affects my health is what I myself do. 1 2 3 I S 6 13. If I take care of myself, I can avoid illness. 1 2 3 I 5 6 Id. When I recover from an illness, it's usually because other people (for example, doctors, nurses, family, friends) have been taking good care of me. 1 2 3 d 5 6 15. No matter what I do, I'm likely to get sick. 1 2 3 I 5 6 16. If it's meant to be, I will stay healthy. 1 2 3 I 5 6 11. If I take the right actions, I can stay healthy. 1 2 3 I 5 6 II. Regarding my health, I can only'do what my doctor tells me to do. I 2 3 I S S 13. Whether or not I get cancer is entirely up to ma. 1 2 3 I 5 6 20. Avoiding cancer is mostly a matter of good fortune.l 2 3 I S 6 21. No matter what I intend to do, if I get cancer or not, it is just going to happen. 1 2 3 I 5 6 22. If I eat properly, get enough exercise, rest and avoid cancer-causing substances, I can prevent myself from getting cancer. 1 2 3 I 5 I 23. If I get cancer or not is a matter of fate. I 2 3 I 5 I 20. I can avoid cancer by taking good care of myself. 1 2 3 I 3 I 25. 26. 27. 28. 29. 30. 31. 32. 33. 125 What is your age? Code Number—LL I. 20'. b. 30'. c. IO's d. 50's a. 60's What is your sex? a. female b. male What is your level of education? a. 6th grade or lower b. high school diploma or equivalent c. bachelor's degree from college d. graduate degree a. other, please specify Do you have cancer? a. yes b. no c. I don't know Does someone in your extended family have cancer? a. yes b. no c. I don't know Have you ever seen a program done by the American Cancer society before? a. yes (if yes, which program? ) b. no Have you attempted to change your diet in any way in the last two years? a. yes b. no If your answer to the previous question was 'yes', how have you attempted to change your diet? a. weight reduction b. lower cholesterol c. both a and b d. other, please specify: a. my answer to the previous question was “no“. What is your job title? APPENDIX K VALUES, ATTITUDES AND INTENTION TO CHANGE BEHAVIOR SURVEY 126 APPENDIX K VALUES, ATTITUDES AND INTENTION TO CHANGE BEHAVIOR SURVEY. VAL". SUNVIY below you will find a list of ten values listed in alphabetical order. we would like for you to arrange them in order of their importance to YOU, as guiding principles in your life. ‘ study the list carefully and pick out the one value which is the most important to you. Write the number '1' in the space to the left of the important value. Then pick out the value which is second most important to you. Write the number '2' in the space to the left. Then continue in the sane nanner for the remaining values (3 through 10) until you have included all ranks from i to lo. RACK VALUE SHOULD RAVI A DIFFERENT RANK. Only one value is ranked “i“, only ON! value is ranked '2', only 08! value is ranked '3', and so on. we realize that some people find it difficult to distinguish the importance of some of these~values. Do the best you can, but PLEASE RANK ALL 10 OF THEN. The end result should truly show how you really feel. » A COHFORIABLB LIFI (a prosperous life) Al EXCITING LIFI (a stimulating, active life) FREEDOM (independence, free choice) RAPPIles (contentedness) hIALIfl (physical and mental well-being) qull RARHONI (freedom from inner conflict) PLFABURI (an enjoyable leisurely life) SILF-RISPICT (self-esteem) A SENS! OF ACCOHPLISHHFN? (lasting contribution) SOCIAL RICOONITIOW (respect, admiration) 127 Code Number _/_/_ This questionnaire contains a number of questions that ask you about your attitudes and beliefs about nutrition. Please answer all of the questions as honestly and as accurately as you can. DIRECTIONS: Bach statement below is a belief statement with which you may agree or disagree. Beside each statement is a scale which ranges from strongly disagree (1) to strongly agree (6). For each item please darken the circle on your answer sheet which corresponds to the number that represents the extent to which you disagree or agree with the statement. The more strongly you agree with a statement, then the higher will be the number you darken on your answer sheet UP TO SIX. The more strongly you disagree with a statement, then the lower will be the number that you darken on your answer sheet. Please make sure that you answer every item and that you choose ONLY ONE number per statement. This is a measure of your personal beliefs, so obviously there are no right or wrong answers. It is important that you respond according to your actual beliefs and not according to how you feel you should believe or how you think we want you to believe. 0 I) 0H0 0&0 I) hilt-I a :22: v m a m .. 2’ 3 353:»: .21....7. HUI-1H8: file-DUI” Ch££h own-cog 33:12”: In wm£m 1. Nutrition is not so important 1 to me as long as I am eating plenty of food. a: I» a» (n o! A21 2. I intend to start eating a more 1 2 3 I 5 6 Intent healthy diet in the near future. 3. My doctor thinks it would be 1 2 3 I S 6 good for me to change my diet. SN I. It is very likely that I will 1 2 3 I 5 6 start eating a diet aimed at CBI lowering my risk of getting cancer. 5. Trying new and different foods 1 2 3 I 5 6 B appeals to me. oren * R & S = from Rosander & Sims' Personal Control Over Eating Intent - Intent to change eating habits SN = Social Norm question CBI = Cancer Behavior Intention Boren = from Boren's Attitude Toward Nutrition (—) = question reversed for scoring 10. 11. 12. 13. II. 15. * CA 128 Code Number__/__/__, Other people's opinions about my diet are important to me. I would be willing to try an unfamiliar food at least once. Changing my diet would.help lower my risk of getting cancer. I eat whatever I want and never think about it later. For better health, I would be willing to try a food I hadn't eaten before or several foods over a period of time. In my spouse, fiance, boyfriend or girlfriend's opinion, my changing my diet in the near future would be bad. It's too late for me to change my diet to lower my risk of cancer. I like to make my own decisions about what I eat, but I often eat what everybody else is eating. Ny parents would like it if I changed my eating habits. I've been eating the same way for years and, at this point, it would be impossible to change. = Cancer Attitude ,. Strongly Disagree 0 8o :92 a MMOHO a 2 2:22 #400 DD » a <: a. < e-I H U>I?~Uh 37.33323 “5538 3.31.: £383.» 2 3 I s 6 SN 2 3 4 5 5 Boren 2 3 I 5 6 CA* 2 3 I 5 6 R & S (-) 2 3 4 5 5 Boren 2 3 I S 6 SN (-) 2 3 I 5 6 (Hi (-) l 2 3 I 5 63'5 S (-) 1 2 3 I 5 SSN l 2 3 I 5 6%;9 S 16. 17. 13. 129 Code Number __/__/__ Strongly Disagree Moderately Disagree Slightly Disagree Slightly Agree Moderately Agree I would be willing to spend time in making nutritious foods available for myself and/or family instead of eating convenience foods of low nutritional quality. ..a N U . 0' fly close friends would think 1 2 3 I 5 it was bad if I adopted a diet to lessen my risk of cancer. Unfamiliar foods often 1 2 3 I 5 interest me. Strongly Agree Boren Ci (-) 5 Boren APPENDIX L SCORING OFFICE RESULTS - FINAL VERSION NCKT BO APPENDIX L .IIII..|I. I OwI Iuo 20~h<~>mo Oa<0202w30wuu outclahm mauthouwQ u>~h1438830 20~h2¢~¢hm~o uuoow 3‘8 muu~>¢wm 20~h<=4<>u I " I I I I I I I I Poghbfi%n¢flhv' I I -%n >g§5=u OI. FWD I! Q GIN v-I 131 CONN. p thlmcammut to Roman ocxozmhm I I I I I I I I I IIIIIII I IIIII rommw I I I I8 thuweuwflwm kmmIomeIrmchI 53% I I I IIIIIIII I I I I I I I I cm zoaeezaxaaumcoIMuhe 24mm 3 3433.5uwfiIzkmflIIIIIIIIIIIIIIIIIII o o 8:3: mm“... o o o. .8 o o o. .8 IIIIIIIII .IoIIIIIIIImIIIIIquIIIerIIIIIIIIIIIIIIIIIIImIIIIIIIIeIIIIIomIIICIIII a. . 8-." on n 8.3. IIIIIIII Im.IIIIIIIImIII IowIIIwnIIIIIIIIIIIIIIIIIIIIImIIIIIIIIwIIIIemIII.InIIII R .v 8. 3 an n on. 3 R c 8. 3.. ma c on- .m s . on- .o a. u 2.- .o IIIIIIIII on.IIIIIIImIIIIIomIIIwPIIIIIIIIIIIIIIIIIIIIoIIIIIIIIpIIII.omIILIpIIII o o co. 3 o o cm. 3 IIIIIIIII mIIIIIIIIeIIIIIowwIImmIIIIIIIIIIIIIIIIIIIIImIIIIIIIIeIIIIB.IIIeIoIIII mew». meme. 3123?: go comebz 33233.. no 3952 moo—9: 22:23:90.3 mo 22:52.58 I IIIIIIIIII m M~_9I.I>IG.I6IEIGI..I»I_I moImoIfipommpmonI I¢ha¢883m wmzcnmwc lwhn "II?IILIIII‘I'UIIII