Questionnaire Reliability Under the Rasch Model

Objective To examine test-retest reliability and internal consistency of a 5-question food insecurity survey used in the National Health and Nutrition Examination Survey. Methods Researchers administered the National Health and Nutrition Examination Survey in the classroom to 92 fourth-grade children (74 African American; 48 girls) in 2 sessions 27–30 days apart in spring, 2011. Each classroom administration lasted 5–10 minutes. Results Test-retest reliability was 0.66 (Kendall tau), which is modest. Internal consistency (Cronbach alpha) was .67 and .70 for respective administrations. Food insecurity scores were related to gender (adjusted P = .05) and academic achievement (adjusted P = .004) but not to socioeconomic status or body mass index percentile (binomial regression). On average, boys reported higher food insecurity than girls. Children with lower academic achievement scores reported higher food insecurity than children with higher academic achievement scores. Conclusions and Implications The National Health and Nutrition Examination 5-question survey may be group administered to assess food insecurity efficiently as reported by individual fourth-grade children.

show abstract

“…20 Kendall tau coefficient quantified test-retest reliability. 21 A paired t test compared scores between administrations.…”

Section: Methodsmentioning

confidence: 99%

Test-Retest Reliability of the National Health and Nutrition Examination Survey's 5-Item Food Insecurity Questionnaire Completed by Fourth-Grade Children

Baxter

Smith

Hitchcock

et al. 2015

Journal of Nutrition Education and Behavior

View full text Add to dashboard Cite

show abstract

“…In Hamon and Mesbah (2002), the conclusion of intensive simulations shows that "the classical Cronbach alpha seems nevertheless to remain a good reliability coefficient under Rasch model assumptions. "…”

Section: Reliability Of An Instrument: Cronbach Alpha Coefficientmentioning

confidence: 99%

Applied Statistics in Biomedicine and Clinical Trials Design

Chen¹,

Liu²,

Qu³

et al. 2015

ICSA Book Series in Statistics

View full text Add to dashboard Cite

“…After an initial rough clustering of items based on previous forced PCA with Varimax rotation, Cronbach alpha coefficient curves [49,50] were used to identify a precisely one-dimensional cluster of items. One single dimension (all items of the cluster measuring the same underlying construct) means that the above curves are increasing monotonically.…”

Section: Assessment Of Unidimensionalitymentioning

confidence: 99%