| Tag | Value |
|---|---|
| file | Reliability_eur-reliability-104-en_eur-reliability-104-en |
| name | eur-reliability-104-en |
| section | Reliability/Analysis/Parallel forms |
| type | string |
| solution | "" |
| Type | Conceptual |
| Program | |
| Language | English |
| Level | Statistical Reasoning |
Imagine a test containing 16 items that aim to measure the same construct. You want to calculate the split-half reliability of this test. Explain why it is not a good idea to split this test in subtests containing respectively the first 8 items and the last 8 items.
This is not a good idea because the two parts might not be parallel. For example, due to fatigue or decreasing motivation the error variance might be larger in the second part of the items.