# 19. MICOM

## Purpose and applicability

Use MICOM to assess whether a composite measurement model is sufficiently invariant before comparing structural relationships across groups. This tutorial asks whether the Organizational Prestige, Organizational Identification, and Affective Commitment composites are comparable across the two recorded gender groups.

MICOM is appropriate when the same PLS measurement model is estimated in two or more meaningful groups. It is not a substitute for establishing sensible groups, adequate group sizes, or equivalent model specification.

## Practice study

- Dataset: [organizational-identification-model-comparison.csv](../samples/organizational-identification-model-comparison.csv)
- Prepared project: [organizational-identification-multimod.qpls](../projects/organizational-identification-multimod.qpls)
- Observations: 305 complete cases
- Groups: `gender = 1` (157 cases) and `gender = 2` (148 cases)
- Data type: deterministic synthetic instructional data
- Research question: Is the measurement model invariant across the two recorded groups?

The project contains four reflective constructs: Organizational Prestige (8 indicators), Organizational Identification (6), Affective Commitment (Joy) (4), and Affective Commitment (Love) (3). Its paths are Organizational Prestige → Organizational Identification and Organizational Identification → both Affective Commitment dimensions.

## Preparation

1. Start QuickPLS, choose **Open Project**, and select the prepared project.
2. Open **Data**. Confirm 305 observations and the `gender` grouping column.

   ![Organizational-identification data prepared for MICOM](../screenshots/micom/01-data-ready.png)

3. Return to **Model** and confirm the four constructs and three saved structural paths. The project opens with saved, fitted coordinates; do not press **Arrange** for this tutorial.

   ![Saved arranged organizational-identification model](../screenshots/micom/02-arranged-model.png)

4. Choose **Validate** and resolve any blocker before continuing. The measurement specification must remain identical across the two groups.

## Calculation

5. Choose **Calculate → MICOM and Multigroup Analysis (PLS / PLSc)**.
6. Use `gender` as the grouping column, retain groups 1 and 2, and keep MICOM enabled. Review each saved configural checklist item rather than treating it as a decorative confirmation.
7. Retain the ordinary PLS profile, deterministic seed, two-tailed comparison, and 5,000 permutation draws used by the qualified project.

   ![MICOM setup with grouping column and configural checklist](../screenshots/micom/03-calculation-setup.png)

8. Leave **Open Results when finished** selected and choose **Start calculation** once. Wait for **Completed** and for the saved project to reopen in Results.

## Results and interpretation

9. Open **Pairwise MICOM results**. Read configural/compositional evidence first, followed by equality of composite means and variances. Partial measurement invariance requires the reviewed configural specification and compositional invariance; equality of means and variances is a stricter, subsequent assessment.

   ![Pairwise MICOM result table](../screenshots/micom/04-results-primary.png)

10. Open **Group eligibility** and confirm both groups were admitted with their complete-case counts. Review resampling accounting under Advanced when documenting the run.

   ![MICOM group eligibility result](../screenshots/micom/05-results-secondary.png)

Do not interpret a non-significant permutation result as proof that groups are identical. Report the exact MICOM step, permutation count, seed, group sizes, estimate, interval or probability, and the measurement-invariance conclusion supported by the table.

## Reporting and reproducibility

Use **Copy** for the visible researcher-facing table or **Export** for a file. Choose **Save Report** when a durable report is needed. Close and reopen the project once and confirm that the exact MICOM result remains selectable.

Common mistakes include changing the measurement model between groups, using a grouping variable with very small levels, skipping the configural checklist, and interpreting structural differences before measurement comparability is established.

## Limitations and related methods

The data are deterministic teaching data and do not support substantive claims about real populations. MICOM supports, but does not itself perform, the structural group comparison. Continue with [PLS-MGA](20-pls-mga.md) after establishing the applicable invariance evidence. For PLSc group analysis, see [PLSc-MGA](21-plsc-mga.md).

Last verified with QuickPLS 2.62.8.