Multivariate analyses of dietary-related patterns as an improvement of methodological tools in medical anthropology
Saša Missoni (Institute for Anthropological Research)
Olgica Martinis (Education and Teacher Traininig Agency)
Paper short abstract:
Analyzing dietary habits is often reduced to consideration of daily intakes or food frequency. Holistic interdisciplinary approach can improve the inclusion of other nutrition-related variables, such as health locus of control and beliefs about education on nutrition, using multivariate methodology.
Paper long abstract:
Applying multivariate statistical methods can reduce the number of variables in the research of dietary habits, often considered as daily nutrients intakes or included in food frequency questionnaires. On the other hand, reducing the number of manifest variables with latent ones could enable the use of other nutrition-related variables in holistic interdisciplinary anthropological approach. Variety of manifest variables could be comprised in latent variables, such as health locus of control and beliefs about education on nutrition. In this study, conducted on two samples of Croatian adolescents who live in two different regions, we are showing the example using Principal Components Analysis (PCA), where few sets of manifest variables are analyzed: multidimensional health locus of control, dietary habits, healthcare style and beliefs about education on nutrition. Cross-sectional study included a total of 165 female high school students (110 from island Hvar and 55 from Beli Manastir), as well as 47 male students (37 from Hvar and 10 from Beli Manastir) ages 15 to 18. For all sets of variables, by using exploratory strategy of PCA, the authors have constructed specially designed questionnaires (i.e. measuring instruments), with satisfactory reliability and construct validity, which were adjusted to the research samples. Except revealing the insight in structure of certain sets of manifest variables, which represent characteristic dietary-related patterns, regression factor scores could enable further statistical analyses among two samples of participants.
Medical anthropology into the future: aspirations and challenges (Commission on Medical Anthropology and Epidemiology)