Measuring Infinite Mindset: A Pilot Comparison of Lower- and Higher-Order Models With Mixed-Format Data

Abstract

Infinite Mindset offers a long-term leadership framework, but its measurement hierarchy requires empirical scrutiny. This study assesses measurement feasibility and specification sensitivity using a documented 30-respondent pilot in a plantation-management research context. Reported partial least squares structural equation modeling results were compared across single-construct and repeated-indicator higher-order specifications, without respondent-level re-estimation. Composite reliability was 0.982 in both models, while average variance extracted was 0.689 and 0.690. Direct associations with administrative performance scores were positive, with coefficients of 0.404 and 0.397; neither proposed indirect pathway was statistically supported. An item loading of 0.694 on the overall construct coexisted with a loading of 0.855 on its designated dimension. Reported repairs produced full-rank reduced matrices, but their mapping to fitted item-level models remains unclear. The contribution is a conditional validation pathway separating data feasibility, measurement hierarchy, and external associations. Findings support further instrument development, not definitive validation, causal inference, or higher-order model superiority.