This process requires us to reflect on individual food items and relative differences in their sensory properties.
Gender specific differential item functioning and differences in predictive power deserve further attention.
fMagis and De Boeck ([in press]) recently found inflated Type I error rates in Mantel-Haenszel approach, and similarly attribute the inflation to the item impact and difference in item discriminations.
In this study, several factors were directly examined with respect to detection of mode-of-administration DIF, including DIF size, percentage of DIF items, and mean difference in trait level between modes, item response model, and analytic procedure.
Like Vreugdenhil et al. (2002), we calculated a masculine score (the sum of boy-typical frequency-weighted items), a feminine score (same for girl-typical items), and a difference score (feminine score – masculine score).
Subjects in the RA group had poorer scores on each item, and these differences were statistically significant for all items except for mental health.
For each CONSORT and STRICTA item, the number and percentage of trials that completely reported the item and mean differences of percentages between two time periods with binomial 95% CI were reported.
If ethnic DIF is present in quality of life, it might imply that the existence of ethnic differences in QoL is a result of biased items, and not real differences in the population.
For all SF-36 domains, all SAT-P items and factors, the differences between baseline and follow up scores resulted statistically significant (p < 0.001), enlighting an improvement both in health status and in subjective HRQoL.
Construct validity analysis demonstrated a significant difference (p ≤ 0.02) between production and office workers on answers to 10 questionnaire items, and a significant difference (p ≤ 0.01) on the scores of the adapted questionnaire between subjects that indicated presence or absence of musculoskeletal symptoms.
