This cross-sectional, observational study was completed at Andrews University which is a Seventh-day Adventist (SDA) institution of higher learning. SDA represent a unique population known for their wide range of dietary habits. This conservative religious group prohibits the use of alcohol, tobacco, and pork and recommends that members adhere to lacto-ovo-vegetarian diet [11,12]. The study was approved by the University’s Institutional Review Board (IRB protocol # 07-122). Participants were drawn from a large undergraduate introductory-level nutrition class that is open to students from all academic directions. Students were recruited by the instructor and assured that anonymity and confidentiality would be maintained. Participation in the study was voluntary. Those who choose to participate received ten bonus points which were counted toward their final grade. Data collection took place over the Thanksgiving holiday in 2007. Students were asked to recruit their parents and grandparents for participation in this survey.

Each participant was asked to complete a four-page Lifestyle Practices Survey which consisted of four parts. Section one had 11 basic census questions (gender, ethnicity, marital status, education, occupation, age, etc.). In section two a 29-item Food Frequency Questionnaire (FFQ) was used to accurately ascertain the vegetarian status of the participants. In section three, questions addressed the use of herbs and supplements. In section four participants were asked to describe which lifestyle they practice (non-vegetarian, pesco-vegetarian, lacto-ovo-vegetarian, or vegan). Using a Likert Scale from 1 to 5 (strongly disagree [1]–agree[2]–no opinion[3]–agree[4]–strongly agree[5]) participants answered questions concerning their attitudes, beliefs, knowledge, and misconceptions about vegetarian lifestyles (Table 1).

Data were analyzed using One-Way-ANOVA, Pearson correlation and Structural Equation Modeling (SEM) [13] techniques with SPSS (version 15.0) and AMOS 7.0 statistical software. One-Way-ANOVA was used to test comparison between age groups. Descriptive data was tested for normality. Pearson correlations were examined to check for the internal validity of the data. In the development of the lifestyle questionnaire cluster analysis was used to group the questions into four separate areas labeled: knowledge about nutrition; beliefs in health foods; attitudes about vegetarian lifestyle; and nutritional misconceptions. P ≤ 0.05 was considered statistically significant.

Overall there were 609 participants who completed the survey. Descriptive data are shown in Table 2. Out of the 609 participants, 215 (35%) were male and 394 (65%) were female. The mean age was 32.0 years for males and 30.6 years for females. Body Mass Index (BMI) was calculated for all participants. The mean BMI was 25.0 for males and 24.4 for females. Using the Food Frequency Questionnaire (FFQ), we identified that 4% of the participants were vegans, 25% lacto-ovo vegetarians, 4% pesco-vegetarians and 67% non-vegetarians.

The lifespan of a generation is not clearly defined. Depending on the cultural norms for marrying age it is generally 20 to 30 years per generation. The age distribution of the population did create four clusters of similar age groups, however, with not very clearly defined beginnings and ends. We have done several statistical analyses defining generation between 20 to 25 years. They all provided somewhat similar results, therefore we are reporting the results using following generational categories: 11–20 years, 21–40 years, 41–60 years, and 61 and older. We asked four questions concerning reasons why they choose a vegetarian lifestyle—the moral reason (it is wrong to kill animals), the health reason (vegetarians live longer and are less sick), the environmental reason (vegetarian lifestyle is much more protective against the environment) and (because 80% of our respondents were Seventh-day Adventists) the faith reason (being vegetarian is part of Adventist lifestyle). The results (Figure 1) showed that the younger people (11–20 years) significantly agreed more with the moral reason (p = 0.003). People ages 41–60 significantly agreed more with the health reason (p = 0.010). Finally, younger people (11–20 years) also significantly agreed more with the environmental reason (p = 0.025). There were no significant differences concerning the faith reason (p = 0.715).

Previous studies raised concerns that self-defined vegetarian status can be an unreliable indicator of true dietary preferences [14,15,16,17]. Table 3 represents the self-defined versus verified vegetarian status of the subjects. The bolded numbers represent those that defined their dietary preferences correctly. In order to make the self-identification process easier when asking participants what lifestyle they follow, the different vegetarian lifestyles were defined, e.g., vegan was defined as—eats vegetables, fruits, legumes, grains; lacto-ovo-vegetarian as - eats dairy products, eggs, vegetables, fruits, legumes, grains, etc. The results vary greatly according to the group. In non-vegetarians 97% of females and males identified themselves correctly, in pesco-vegetarians only 32% of females and 17% of males identified themselves correctly. In lacto-ovo-vegetarians 82% of females and 78% of males identified themselves correctly, and in vegans 48% of females and 57% males identified themselves correctly. The data for vegans however need to be interpreted carefully because of small numbers.

This study examined the way underlying health concepts could explain why people chose vegetarian lifestyle using the SEM statistical method. SEM is a powerful multivariate statistical method being used in social sciences, and with increasing frequency in health behavior research. SEM examines underlying relationships among variables in the model and helps to explain social or behavioral phenomena [13]. Our model (Figure 2) was constituted by four sets of concepts: the Attitudes toward vegetarian lifestyle, Nutritional knowledge, Nutritional misconceptions and Health food beliefs. How are Attitudes toward vegetarian lifestyle related to Nutritional knowledge, Health food beliefs and Nutritional misconceptions?

The hypothesized model was assessed by AMOS version 7.0 using the maximum likelihood method. The model was evaluated by four fit measures: a, the chi square b, the Comparative Fit Index (CFI) c, the Good-of-Fit-Index (GFI) and d, the Root Mean Square of Approximation (RMSEA). The results for three out of the four indices support the proposed model. The chi square had a value of 165.057 (Df = 82, n = 609), p=0.000, indicating a non-acceptable match between the proposed model and the observed data. However due to the size of the sample additional fitted indices were considered. The CFI = 0.926, GFI = 0.965, both of them indicating an excellent fit of the model. The RMSEA measures the discrepancy between the sample coefficients and the population coefficients equals 0.041 (confidence interval 0.023 – 0.050) indicating an acceptable fitting [18].

Findings support model that suggests that the Attitudes toward vegetarian lifestyle are significantly correlated with Nutritional knowledge (r = 0.43, p = 0.000) and have negative effect on Health food beliefs (r = −0.21, p = 0.034) and Nutritional misconceptions (r = −0.46, p = 0.000). The observed effect between Nutritional knowledge and Attitudes toward vegetarian lifestyle (r = 0.43, p = 0.000) is bilateral suggesting that these variables influence each other. Increased nutritional knowledge might lead to positive attitude toward vegetarian lifestyle, and vice versa vegetarian lifestyle may promote increase in nutritional knowledge. Further, Nutritional knowledge has negative effect on Nutritional misconceptions (r = −0.32, p = 0.000) and positive effect on Health food beliefs (r = 0.28, p = 0.012). The model seems to indicate that in our population positive attitude toward vegetarian lifestyle is knowledge base instead of just being nurtured by some traditional nutritional beliefs or misconceptions. Positive attitudes toward vegetarian lifestyle contribute to the reduction of nutritional misconception and non scientific beliefs about health foods.

The data presented in this paper indicate that there are significant differences across generations as to why people choose to live a vegetarian lifestyle. Young people under the age of 20 seem to choose vegetarian lifestyle for moral and environmental reasons, while the middle age group of people between ages 41 to 60 seem to choose this lifestyle for health reasons. This trend seems understandable, given the wealth of publications documenting the health benefits of vegetarian and low-meat diets [11,19,20,21,22]. For younger people health issues are not priority, however as they age this increases in importance as shown by the data. The health reason to be vegetarian in our population produced the strongest attitudes on the Likert scale, confirming previous finding that Adventist traditionally chose vegetarian lifestyle for health reasons.

Several potential limitations to this study should be considered. This is a population-based cross-sectional study, which included both genders and all age groups. The study was conducted on a campus of a American private university which may limit the generalizibility of the results. Although the sample size of the population was large enough, some groups such as the over sixty or vegans were underrepresented so the results should be interpreted with caution. Although SEM is a sophisticated analytic tool for testing theoretical models in behavioral or social science, the analyses are correlational which makes it difficult to establish causality. Because the isolation of variables in the model are impossible, all models must be looked at only as estimation of reality [23].