Purpose and applicability
Which of three organizational-identification models should be retained when equation-level predictive evidence is the primary criterion?
Prerequisites
- Use the prepared project because it retains three compatible candidate models, their ordinary PLS results, and the qualified POMS result
Do not use it merely because it is available
Use Prediction-Oriented Model Selection only when its estimand, data roles and assumptions answer the stated research question. Choose a related estimator or diagnostic when the intended outcome is not among the documented outputs below.
Study preparation
- Teaching data
- organizational-identification-model-comparison.csv
- Observations
- 305
- Study type
- Deterministic synthetic teaching study
- Calculate route
- Calculate → Prediction-Oriented Model Selection
Open the data dictionary and variable-role reference
Model
Organizational Prestige (8 indicators), Organizational Identification (6), Affective Commitment (Joy) (4), and Affective Commitment (Love) (3), all reflective
Baseline: Prestige → Identification; Identification → Joy and Love. Candidate 2 adds Prestige → Joy. Candidate 3 instead adds Prestige → Love
Exact workflow
Download organizational-identification-model-comparison.csv and the prepared organizational-identification-poms.qpls project. Keep both files in a writable study folder.
Start QuickPLS, choose Open Project, and open the prepared project. Its saved model is already arranged and fitted.
Open Data and confirm that the study contains 305 observations. Return to the model or calculation workspace.
Confirm the saved model specification: Organizational Prestige (8 indicators), Organizational Identification (6), Affective Commitment (Joy) (4), and Affective Commitment (Love) (3), all reflective. Structural specification: Baseline: Prestige → Identification; Identification → Joy and Love. Candidate 2 adds Prestige → Joy. Candidate 3 instead adds Prestige → Love.
Confirm that construct labels, indicators, and arrows are readable. The supplied project is saved after Arrange and Fit; do not rearrange it before the tutorial run.
Choose Calculate, then use Calculate → Prediction-Oriented Model Selection.
Review the essential settings listed below. Open Advanced settings only when the design requires a non-default option.
Resolve actionable blockers, review applicability warnings, and leave Open Results when finished selected.
Choose Start calculation. Wait for Completed, then confirm that Results opens for this method.
Inspect the named outputs below in order. Use Save Report, table Copy, or Export as required, then save and reopen the project to confirm the result remains available.
Essential and advanced settings
- Select all three compatible saved models
- Retain the common Organizational Identification, Affective Commitment (Joy), and Affective Commitment (Love) outcomes
Keep other advanced controls at their documented defaults unless the study design requires a justified change. Record every non-default value in the research log.
Results to inspect
- Candidate ranking
- Equation BIC
- Predictive-loss evidence
- Ties or mixed-evidence decision table
Interpretation guidance
For Prediction-Oriented Model Selection, interpret the listed outputs together with the prerequisite result, the selected settings, model assumptions and data quality. The supplied values are instructional; they are not validation against an external paper or another software package.
- The installed calculation and canonical tables passed; Save Report returned operating-system error 3 in the isolated disposable qualification path, so report persistence is not claimed for this fixture
Reporting guidance
Report the QuickPLS version, Prediction-Oriented Model Selection route, sample size, model or variable roles, preprocessing, essential settings, non-default advanced settings, and the named primary outputs. Retain the project, data checksum and exported table used in the manuscript.
Common mistakes and recovery
- Running the method before completing its prerequisite calculation.
- Changing the data, model or variable roles after the prerequisite result was saved.
- Treating an unavailable or not-applicable value as zero.
- Reporting an estimate without its method-appropriate uncertainty or diagnostic context.
Screen-by-screen evidence
Images shown here are mapped to this tutorial’s required installed-application evidence. Missing captures are labelled explicitly and are not replaced with generic screenshots.