Abstract
BACKGROUND: The Theory of Planned Behavior (TPB) has guided the investigation of breastfeeding since the 1980's, incorporating the major constructs of attitudes, subjective norms/normative beliefs, perceived behavioral control, and intentions. The purpose of this research study was to define a TPB-based structural latent variable model so as to explain variance in breastfeeding intentions and behaviors among a cohort of Midwest breastfeeding mothers. METHODS: The longitudinal descriptive study utilized questionnaire data collected from a convenience sample of 100 women with low-risk pregnancies with the intention to breastfeed at three separate time points (> 30 weeks antepartum, 10 and 60 days postpartum). Data were coded and analyzed using IBM SPSS, SAS and the lavaan package in R. RESULTS: Participants were predominantly White (94%, n = 94), married (95%, n = 95), college-educated (96%, n = 96), and had previous breastfeeding experience (75%, n = 75). The majority gave birth vaginally (79%, n = 75). Varimax analysis revealed a plurality of factors within each domain. Attempts to fit a structural model, including both hierarchical and bi-factor latent variables, failed, revealing a lack of statistical significance and poor fit statistics. CONCLUSION(S): These findings illustrate the importance of using methods that fit the phenomena explained. Contributors to poor model fit may include outdated tools lacking cultural relevance, a change in social norms, or a failure to capture the possible influence of social media and formula marketing on breastfeeding behaviors. The null finding is a significant finding, indicating the need to revisit and refine the operationalization and conceptual underpinnings of the TPB through qualitative methods such as exploring the lived experiences of breastfeeding women in the Midwest region.