Description

This Master Prompt focuses on preparing empirical data for defensible AMOS analysis before estimating the measurement or structural model. It guides researchers through questionnaire coding, data cleaning, missing-data assessment, response-quality screening, outlier examination, descriptive statistics, sample adequacy, distributional assessment, multicollinearity diagnostics, preliminary reliability analysis, factorability assessment and other procedures needed before covariance-based structural equation modeling.

A major emphasis is placed on preventing researchers from proceeding directly from raw questionnaire responses to CFA or SEM without first establishing the analytical suitability and integrity of the dataset. The uploaded studies demonstrate a clear division of labor between SPSS and AMOS: preliminary screening and descriptive procedures are commonly performed in SPSS before CFA, measurement-model assessment and structural modeling in AMOS. One AMOS application explicitly examines normality through skewness and kurtosis, outliers using Mahalanobis and Cook’s distances, and multicollinearity through VIF and tolerance before inferential analysis. Another uploaded study demonstrates explicit response-quality cleaning, removal of straight-line responses, sample-size assessment and separation of SPSS preliminary analyses from subsequent CFA in AMOS.

The resulting workflow should provide an auditable transition from raw data to an analysis-ready dataset. Researchers should document every major decision, avoid mechanically deleting cases simply to improve later model fit, distinguish empirical evidence from arbitrary threshold rules, and ensure that any treatment of missing data, outliers, non-normality, problematic responses or inadequate sample size is theoretically and statistically defensible.