Importance and performance checked with independent calculations

What do importance and performance mean in this test?

The target is the construct whose result you want to understand. Its predecessors are constructs that can reach it through the model's paths. In this saved example, Joy is the target and the constructs labelled I and P are its direct or indirect predecessors.

Importance uses a total effect: the relationship through direct and indirect paths, rather than a direct path alone. This historical test uses standardized total effects. Performance rescales scores using the minimum and maximum observed in the complete sample. A reference must use those same definitions to make a meaningful numerical comparison.

Which software and reference were used?

The table identifies the installed QuickPLS version used for this historical example. The supporting PLS results came from the saved cSEM reference. The IPMA values were then calculated independently using those weights and the observed score ranges; this step did not rerun the original model fits.

On a small screen, swipe the table sideways to see every column.

The installed-app IPMA example

There are 281 comparisons: 60 new IPMA checks and 221 reused PLS reference values. The new checks cover importance and performance for the relevant constructs and indicators, while the reused values provide the supporting model estimates.

The reported zero difference applies after the reference values have been adjusted to the precision of the saved output. That adjustment is part of the recorded comparison, so interpret the result using the stated definitions and precision.

• This observed-range example does not test researcher-configured theoretical ranges, combined IPMA (cIPMA), or every possible target. Those choices need their own matching reference.

Construct importance and indicator performance

The target is Affective Commitment (Joy). Organizational Identification and Organizational Prestige are the two predecessor constructs. The first four rows compare their total importance and performance; the remaining rows compare all fourteen predecessor-indicator performance scores.

Organizational Identification has total importance 0.613918 and performance 63.7219 on the historical observed-range 0–100 scale. Organizational Prestige has total importance 0.222394 and performance 76.6575. These values describe the selected model and observed score ranges; the importance coefficients are not percentages or estimates of an intervention.

The reference was aligned to the saved export: six decimal places for importance and four for performance. The recorded zero differences apply after this alignment. The downloads retain the unrounded reference, its adjustment and the difference from the unrounded value. These eighteen displayed comparisons are part of the sixty IPMA checks; the overall count also includes 221 reused PLS checks.

On a small screen, swipe the table sideways to see every column.

The error limits used for this example

abs(native - aligned_reference) <= absolute_tolerance + relative_tolerance * abs(aligned_reference)

Here native is the QuickPLS result and aligned_reference is the reference under the same definition and recorded precision adjustment. The allowed difference has a fixed part and a part proportional to the reference value. The table separates those decimal checks from exact matches. A result expected to be missing must also be missing in the reference.

On a small screen, swipe the table sideways to see every column.

How to use this comparison

Read the IPMA result together with its target, predecessor list, total-effect definition and observed scaling ranges. Agreement under those choices supports that particular importance-performance calculation. Changing the ranges or the definition of importance changes what needs to be checked.

The comparison count includes many values from one example, and some were already checked in the supporting PLS work. It is not a count of independent datasets or studies. Use the report to distinguish the new IPMA calculations from the reused estimates.

• The error table records which values allow a small decimal difference and which require an exact match. Inspect the full recorded precision when comparing a result with the rounded table above.

How to inspect the IPMA files

Use the JSON report to identify the target, settings, versions and division between new IPMA checks and reused PLS results. Open the CSV to inspect the available numerical comparisons in a spreadsheet.

Repeating the calculation requires the original dataset, model, score ranges, settings and software versions. These downloads report completed work; they do not include the original dataset or app source code.

• Read the comparison rules ↗
• Find checks for related methods ↗