The 3P Model of Mathematical Problem Solving: Epistemological and Psychological Associations with Preservice Teachers’ Performance 

Source abstract

This article reappraises the 3P model developed in a 2012 doctoral study of 118 preservice mathematics teachers at 10 institutions in the southern Philippines. Drawing on the archived dissertation, it examines the original inventories, a 20-item problem-solving test, regression and interaction analyses, full and reduced structural equation models (SEM), and six participants' think-aloud work. It also develops a construct-measure-inference framework that distinguishes theoretical labels, observed scores, statistical relations, and substantive claims. In the reduced SEM, need for cognition had a positive conditional path to test performance (β = .864), whereas metacognitive awareness had a negative path (β =-.164); reported fit was χ²(112) = 132.11, p = .094, CFI = .981, and RMSEA = .039. The cognition factor was a need-for-cognition inventory score rather than observed mathematical reasoning, and awareness scores did not establish regulation enacted during problem solving. Comparing these measures with the separately collected process evidence clarifies what the model can support and identifies requirements for independent replication. The article reports no new participant-level analysis or causal effects.