Personality and Dietary Habits

Background: The personality of healthy individuals has not been well studied in relation to health consciousness, dietary habits and actual food intake, simultaneously. Objective: Our objective was to study the association between personality and dietary habits. Design: Information on dietary habits, including taste preferences and the frequency of food consumption, was collected through a questionnaire from 76 male and 394 female students. The personality of students was determined by a modified NEO-FFI test. Health status, height, body weight, body fat percentage and blood pressure were measured by physical examination. Main outcome measures were personality scores as indicators of a healthy dietary pattern. Results: Food intake was influenced by neuroticism (N), extraversion (E), openness (0) and agreeableness (A) of personality. Taste preferences and receptivity to dietary advice were also influenced by personality: the odds ratios (ORs) between the high and low tertiary points of the NEO-FFI scores for salty and sweet taste preferences were significantly higher in the group that scored high for neuroticism (N) (salty taste preference: OR=2.25, NS in males and OR=2.39, 95%CI=1.16-4.93 in females; sweet taste preference: OR=21.00, 95%CI=2.40-183.99 in males and OR=3.33, 95%CI=1.61-6.91 in females). On the other hand, the groups with high scorer for 0 and A did not like salty tastes. The groups with high scores for A and C were receptive to dietary advice. High scores of each N, E, 0, A, and C factor were characterized by distinguishable, dietary habits and lifestyle. For nutritional or health education, group classes are sufficient for high A and O. High C scorer displayed discrepancies between health consciousness and dietary habits, so intervention or a close follow-up by medical professionals would be necessary to improve the health of individuals in this group. High E scorer possessed a confident attitude towards their health, but they were not interested in developing healthy habits. High N scorer was adverse to receiving health information and learning healthy dietary habits. Conclusion: Personality determined by NEO-FFI test was related to dietary habits and health attitude. Effective health education methods must take the personality of the targeted individuals into consideration. J Epidemiol, 2000 ; 10 : 191-198


INTRODUCTION
Dietary habits and lifestyle are closely correlated with chronic diseases such as diabetes mellitus, cardiovascular disease and cancer. Modification of dietary habits, however, is often difficult and seems to depend on health consciousness. Selfmanagement of health, like body weight control, requires a high level of consciousness and much effort . The personality of an individual influences their health consciousness and thus has an effect on the efficacy of nutritional education . However dietary habits have not been sufficiently studied in relation to personality. Previously, we developed a modified Japanese version of the NEO-FFI questionnaire and found that personality was correlated with individual health consciousness " . The purpose of this study was to clarify the relationship between personality, as determined by the NEO-FFI score, and dietary habits, including taste preferences. Information on dietary habits were collected through the use of a questionnaire of food consumption 3), and the effects of personality on health consciousness and dietary practices were analyzed.

SUBJECTS AND METHODS
A total of 76 male and 394 female junior students in the Department of Nutritional Science at the Tokyo University of Agriculture participated to in the study, which was conducted every October from 1996 to 1998. The subjects were questioned about their dietary habits, frequency of food consumption, and personality in addition to receiving a physical examination. Dietary patterns were identified using a self-administered questionnaire developed by a JPHC study 3). The questions on the questionnaire were designed to determine the consumption of 40 items, covering at least 80 % of the individual's energy intake. The students were asked to note the number of regular sized bowls of rice that they consumed per day. The frequency of miso soup, mayonnaise, dressing, ketchup and fried and deep fried dishes consumed per week was determined by asking the students to selected one of four categories: rarely, 1-2 times/week, 3-4 times/week, and almost everyday. The consumption of a wide variety of food items (chicken, beef, pork, ham, bacon and sausages, liver, fresh fish, dried and salted fish, fish paste products, salted roe and salted preserves, small fish, seaweed, green vegetables, carrots, tomatoes, tofu, natto, nozawana-zuke, takana-zuke and other pickled vegetables, apples, citrus, eggs, milk, cheese, yogurt, butter, margarine, bread, noodles, instant noodles, western confectioneries, and Japanese confectioneries) were determined by selecting one of five categories: never, rarely, 1-2 days/week, 3-4 days/week, and almost everyday. Taste preferences for oily, salty, sweet, sour, or sweet snacks, and hot (temperature) dishes were also queried. Questions concerning dietary precautions, such as avoiding burnt fish or meat, salty foods, cholesterol-rich foods, and animal fats as well as good dietary practices, such as the frequent blood pressure and diastolic blood pressure are summarized in Table 1.
The NEO-FFI scores of the students are shown in Table 2. The scores for both males and females coincided well, but female students showed significantly higher scores for neuroticism and agreeableness. Answers concerning food frequency were similarly distributed for males and females. The intake of rice and bread, however, was 3 bowls/day for male students and 2 bowls/day for female students. Furthermore, the males consumed bread 1-2 days/week on average while the females consumed bread almost everyday. The consumption of liver, natto, and pickles was lower in females, but the consumption of yogurt and milk was higher.
The results of the principal component analysis with a varimax rotation for food items are shown in Table 3. The Kaiser-Meyer-Olkin value was 0.869, and the attributable cumulative percentage of the 10 factors was 58.6%. The factors were characterized as follows: Factor 1= healthy food, Factor 2= salty food, Factor 3= meat, Factor 4= fruit, Factor 5= dairy products, Factor 6= staple foods, Factor 7= western foods, Factor 8= fish products, Factor 9= confectioneries, and Factor 10; noodles.
The Pearson's correlation coefficients between principal component score by the ten factors made from the food frequency answers and the NEO-FFI scores are shown in Table 4. C did not associate with any factor of food frequency in both sexes. N scores had a positive association with the frequent intake of dairy products in males. Female students with the frequent intake of fish products and a smaller consumption of fruits. Male students with high scores for 0 displayed a negative correlation with western meals. Female students with high scores for 0 showed a positive correlation with healthy foods and fruits. A scores had a positive association with the frequent intake of staple foods and confectioneries, and a negative association with the frequent intake of salty foods in male students, while female students with high scores for A showed a positive correlation with fruits.
The odds ratios (ORs) for the personality scores and dietary habits are shown in Table 5. The N, E, 0, A and C scores all showed significant ORs for dietary habits and food preferences.
Male students that scored high for N did not avoid burnt fish or meat (OR=0.28, 95%CI=0.08-0.96) but did avoid choles- for females). Female students with high scores also ate an appropriate amount, ate a balanced diet, and ate breakfast regularly. They also showed a high OR for the avoidance of salty foods, the avoidance of cholesterol-rich foods, and the avoidance of animal fat (OR=2.05, 1.79, 2.29, 2.69, respectively).

DISCUSSION
Dietary habits are an important factor in the maintenance of health. The influence of personality on dietary habits, however , has not been sufficiently studied. Our previous study clarified in females). Conversely, students with high scores for N displayed a low level of health consciousness. Neuroticism is known to be related to somatic complaints of all kinds. N scores were reported to be useful in detecting false positive symptom, when patients had complaints without real disease 5). Prediction of health behaviors and of compliance with medical instructions was important in case of alcohol abuse program, because it made possible to identify individuals who were more likely to benefit from program, or who were in need of more intensive intervention efforts 6).
The present study attempted to clarify the influence of personality on dietary habits, taste preferences and the actual intake of food using data obtained from a questionnaire on food consumption. High scores of each personality factor N, E, 0, A and C showed characteristic feature, as follows. Table 3. Rotated factor loadings based on rank correlations of food frequency.
Principal components analysis with varimax rotation      Students who scored high for A considered themselves to be healthy and were characterized by a desire to be healthy, regular dietary habits, a dislike of salty and hot (temperature) dishes, and an avoidance of salty food, burnt fish or meat, and animal fat. They also frequently consumed fruits and confectioneries, and they limited their intake of salty food.
Students who scored high for 0 had similar characters to those that scored high for A, except for considering themselves not to be healthy. They were interested in trying to improve their health and practice good dietary habits.
Students with high scores for C had a high level of health consciousness, considered a dietary balance, and were active in sports. They avoided salty foods and animal fats. No significant relationships, however, were found between their personality characteristics and taste preferences or actual food consumption detected from the food frequency answers. Students who scored high for E also considered themselves to be healthy and displayed a high level of health consciousness, although they only ate breakfast once or twice a week.
They preferred salty foods, oily dishes, and frequently consumed fish products.
Students with high scores for N did not consider themselves to be healthy, but neither did they follow any healthy practices nor avoid eating burnt fish or meat. For nutritional and health education, high scorer of A and 0 would be receptive to group education. High scorer of C displayed discrepancies between health consciousness and dietary habits, so intervention or a close follow-up by medical professionals would probably be necessary to improve the health of individuals in this group. High scorer of E and N are healthy so far, but they may be at a high risk for developing health-related problem 7). High scorer of E was confident about their health, so they did not practice any healthy habits. High scorer of N seemed adverse to receiving health information and learning healthy dietary habits. These people are necessary to be introduced motivation for keeping their health.
A variety of scales to measure constructs as coronary prone behavior, abnormal illness behavior, and health locus of controls have developed in recent years 8). These scales, however, addressed narrow and focused issues in health psychology. NEO-FFI seemed to be useful for many purposes to obtain more general measures of personality. The present subjects were all young healthy students, but they had already shown some trends in health habits, including dietary habits.
Our findings suggest that effective health education methods must take the personality of the education recipients into consideration.