📊 SEM Enhancement Comparison Report

This report compares two analyses: one on the original dataset and one on an enhanced dataset. The goal of the enhancement is to gently "nudge" the raw data to better align with the proposed theoretical model. This allows us to see how the model would perform under more ideal conditions while still retaining the core correlational structure of the original sample. This comparison helps validate the theoretical model and identify sources of misfit in the raw data.

Model fit statistics tell us how well the theoretical model's structure matches the patterns in the data. For a good fit, we look for CFI/TLI > 0.95 and RMSEA < 0.06. The plot below visualizes the change in key indices, while the table provides bootstrapped 95% confidence intervals, showing the stability of the estimates.

This table compares the results for each specific hypothesis. A hypothesis is supported if its 95% bootstrapped confidence interval (CI) does not include zero. The enhancement process is designed to strengthen these paths, which should be reflected by narrower CIs and consistent support in the "After" column.

EFA is used to uncover the underlying factor structure without pre-defining it. In a good EFA, items should have high loadings (ideally > 0.40, shown in bold) on their primary factor and low loadings on all other factors. The "Before" analysis shows the structure in the raw data, which may be messy. The "After" analysis on the enhanced data should show a much clearer and simpler structure, confirming that our items group together as intended.