# 27. Mediation

## Purpose

How much of training quality's association with job performance operates through self-efficacy and training transfer?

## Practice study

- File: [digital-training-mediation.csv](../samples/digital-training-mediation.csv)
- Prepared project: [digital-training-mediation.qpls](../projects/digital-training-mediation.qpls)
- Observations: 420
- Data type: deterministic synthetic instructional data
- Complexity: a realistic multi-construct model with multi-item measurement blocks

**Prerequisites**

- Run **PLS-SEM Algorithm** first when starting from a newly authored or changed model.
- The supplied prepared project already contains a qualified bootstrap result for inspection. Create a revision before rerunning it.

## Exact procedure

1. Start QuickPLS and choose **Open Project**. Select [digital-training-mediation.qpls](../projects/digital-training-mediation.qpls). This prepared teaching project contains the exact dataset, four-construct mediation model, bindings, saved layout, and installed-binary-qualified mediation result.

2. Open **Data** and confirm **digital-training-mediation.csv** contains 420 observations. Return to **Model**. The linked CSV remains available separately for inspection and re-import practice.

   ![Mediation sample data](../screenshots/mediation/01-data-ready.png)

3. Confirm that the prepared model contains these constructs and measurement blocks: Training Quality (training_quality_1–training_quality_4), Self-efficacy (self_efficacy_1–self_efficacy_4), Training Transfer (training_transfer_1–training_transfer_4), and Job Performance (job_performance_1–job_performance_4), all reflective.

   Confirm these saved structural paths: Training Quality → Self-efficacy → Training Transfer → Job Performance, plus Training Quality → Training Transfer, Training Quality → Job Performance, and Self-efficacy → Job Performance.

4. The prepared project already contains a saved, neatly arranged model layout. Do not choose **Arrange** or otherwise change the canvas before this tutorial calculation. If you later edit the model, arrange it once and save the edited project before calculating.

   ![Mediation prepared model or variable roles](../screenshots/mediation/02-arranged-model.png)

5. Choose **Validate** and confirm there are no blockers. Because the prepared project contains a retained result, the model bar may show **Revision required**. Choose **Create revision**, then **Edit model**. This prepares an editable scientific revision without requiring you to rearrange or redefine the model.

6. Use **Calculate → PLS-SEM Bootstrapping**.

   ![Mediation calculation setup](../screenshots/mediation/03-calculation-setup.png)

7. Complete the method-specific settings:

- Use 1,000 subsamples, a fixed seed, and two-tailed 95% intervals

   The installed documentation journey shown here uses QuickPLS's valid 100-subsample minimum to keep evidence capture bounded. Its estimates are illustrative workflow evidence, not recommended final research inference. Use the 1,000-subsample tutorial setting above—or a larger design-appropriate value—for substantive analysis.

8. Read **Readiness**, leave **Open Results when finished** selected, and choose **Start calculation** once. Wait for **Completed**; do not close the project while the result is being saved.

9. In **Results**, confirm **Mediation** is selected and inspect these outputs in order:

1. Specific indirect effects
2. Aggregate indirect effects
3. Direct effects
4. Total effects
5. Bootstrap confidence intervals

   ![Mediation primary result](../screenshots/mediation/04-results-primary.png)

   ![Mediation secondary result](../screenshots/mediation/05-results-secondary.png)

10. Choose **Save Report** to preserve the exact result. Use **Export** for the needed table/report format. Close and reopen the project once and confirm the saved result remains selectable.

## Reading guidance

- Interpret a specific indirect effect using its signed estimate together with the bootstrap confidence interval. An interval that does not contain zero supports an indirect association under the selected bootstrap procedure; report the path, estimate, interval, resample count, confidence level, and seed.
- **Aggregate effects** combine the applicable direct and indirect paths between a source and target. Do not substitute them for the named specific indirect path when the research question concerns a particular mediator sequence.
- The data are deterministic synthetic teaching data. They are suitable for reproducing the workflow, not for substantive publication claims.
- Website evidence was last verified against the installed QuickPLS 2.62.8 binary on 12 September 2026.