Young adolescents' engagement in dietary behaviour - the impact of gender, socio-economic status, self-efficacy and scientific literacy. Methodological aspects of constructing measures in nutrition literacy research using the Rasch model

Abstract

Objective: The present study validates a revised scale measuring individuals' level of the 'engagement in dietary behaviour' aspect of 'critical nutrition literacy' and describes how background factors affect this aspect of Norwegian tenth-grade students' nutrition literacy.

Design: Data were gathered electronically during a field trial of a standardised sample test in science. Test items and questionnaire constructs were distributed evenly across four electronic field-test booklets. Data management and analysis were performed using the RUMM2030 item analysis package and the IBM SPSS Statistics 20 statistical software package.

Setting: Students responded on computers at school.

Subjects: Seven hundred and forty tenth-grade students at twenty-seven randomly sampled public schools were enrolled in the field-test study. The engagement in dietary behaviour scale and the self-efficacy in science scale were distributed to 178 of these students.

Results: The dietary behaviour scale and the self-efficacy in science scale came out as valid, reliable and well-targeted instruments usable for the construction of measurements.

Conclusions: Girls and students with high self-efficacy reported higher engagement in dietary behaviour than other students. Socio-economic status and scientific literacy - measured as ability in science by applying an achievement test - did not correlate significantly different from zero with students' engagement in dietary behaviour.

Keywords: Critical nutrition literacy; Dietary behaviours; Quantitative research; Rasch modelling.

Fig. 1

(colour online) The response category probability curves for item 69 of the engagement in dietary behaviour (EDB) scale. The curves marked 0–5 illustrate the probability of ticking off in each of the six response categories on the rating scale applied in the questionnaire as a function of the estimated attitude levels on the EDB scale. The line - - - - - is an asymptote (probability equals 1, i.e. 100 %)

Fig. 2

(colour online) The item characteristic curve for item 73 of the engagement in dietary behaviour (EDB) scale, plotting the expected value of response on the six-point rating scale applied in the questionnaire v. person attitude level on the EDB scale. The line - - - - - is an asymptote. The mean person attitude level of each of three class intervals is marked on the x-axis. The observed mean response category value for each class interval (●) is compared with the expected values described by the theoretical graphical representation of the polytomous Rasch model (———)