Spearman Brown Reliability
Calculator
Results
- Corrected reliability
- 0.809523
- Reliability gain
- 0.129523
- Error variance (%)
- 19.047619
Academic results
| Corrected reliability | 0.809523 |
| Reliability gain | 0.129523 |
| Error variance (%) | 19.047619 |
formula-map diagram
- Corrected reliability
- 0.809523
- Reliability gain
- 0.129523
- Error variance (%)
- 19.047619
Formula breakdown
Formula
r' = k × r ÷ (1 + (k − 1) × r)= 0.80952380952381
Note
This is a simplified model. Grading rules, credit systems and statistical assumptions vary by institution and study design; check your syllabus, registrar or methods guide before relying on these figures.
More in Academic and research
See all →Frequently asked questions
What problem does the Spearman-Brown formula solve?+
It corrects the reliability estimate obtained from split-half testing, where a test is divided in two and correlated with itself, since splitting a test in half inherently shortens each half and underestimates the reliability of the full-length test.
Why does splitting a test in half underestimate its true reliability?+
Reliability generally increases with test length, because more items average out random measurement error. A half-length test is inherently less reliable than the full test, so correlating two halves directly understates how reliable the complete instrument actually is.
What does the corrected reliability coefficient tell me?+
It estimates what the reliability would be for the full-length test, based on the correlation between its two halves, giving a more accurate picture of internal consistency than the raw split-half correlation alone.
Can the Spearman-Brown formula be used to predict reliability for a longer test?+
Yes, the general form of the formula predicts reliability if a test were lengthened (or shortened) by any factor, not just doubled, which is useful for estimating how many additional items would be needed to reach a target reliability level.
What reliability value is generally considered acceptable?+
A coefficient above 0.70 is often considered acceptable for research purposes, 0.80 or higher for more consequential decisions, and 0.90+ for high-stakes individual assessments, though the appropriate threshold depends on how the test results will be used.