Creation and Psychometric Evaluation of the Teen Moms Child Feeding Readiness to Change Questionnaire for Teenage Mothers in Rural Suburbs of Nigeria

Objective: To develop and validate the Teen Moms Child Feeding Readiness to Change Questionnaire for Sub-Saharan Africa for evaluating the transtheoretical stages of behavioral change related to infant and young child feeding. Design: Cross-sectional study. Setting: Communities with populations with low income in rural suburban areas of Abuja, Nigeria. Participants: Ten nutrition experts and 156 teenage mothers. Methods: Nineteen items were generated on a 5-point Likert scale for responses. Two questions were eliminated after content evaluation by a panel of 10 nutrition experts and comprehensibility assessment by 6 teenage mothers. Unidimensionality and internal consistency reliability of the subscales were determined by factor loadings, Loevinger's scalability coefficient H and Molenaar Sijtsma rho coefficients. Construct validity was established via polytomous item response theory in 150 teenage mothers in rural suburbs of Abuja, Nigeria. Results: The final questionnaire contained 11 questions in 3 subscales. All retained items had factor loadings > 0.4, Loevinger's coefficient H > 0.3, and rho coefficients > 0.6. All items displayed good discrimination (≥ 1) and difficulty parameters. Response characteristic curves indicated that the Likert option neither agree nor disagree was not frequently selected; thus, it was deleted. From scale information curves, it was observed that the contemplation, action, and maintenance subscales precisely estimated information for respondents with average ability to perform the latent construct. The precontemplation subscale contained more information for respondents with less than average abilities. Conclusions and Implications: The Teen Moms Child Feeding Readiness to Change Questionnaire for Sub-Saharan Africa demonstrated promising preliminary evidence of validity and reliability for assessing Transtheoretical stages of change in teenage mothers. Further validation in larger and more diverse samples is warranted.