CVI (Content Validity Index) is the average of all individual item CVR values, providing a single validity estimate for the entire scale or instrument. While CVR evaluates individual items, CVI gives you a quick overview of overall scale validity by averaging all the CVR values together.
CVI is useful for:
- Getting a quick scale-level validity summary
- Comparing validity across different scales
- Communicating results to non-technical audiences
- Making pass/fail decisions about scale validity
CVI Formula
CVI = Sum of all CVR values / Number of items
How to Calculate CVI
Step 1: Calculate CVR for each item (using the CVR formula)
Step 2: Add all CVR values together
Step 3: Divide by the number of items
Example Calculation
Suppose you have a 4-item scale with these CVR values:
Item 1: CVR = 0.80
Item 2: CVR = 0.60
Item 3: CVR = 0.90
Item 4: CVR = 0.85
CVI = (0.80 + 0.60 + 0.90 + 0.85) / 4 = 3.15 / 4 = 0.79
Acceptable CVI Thresholds
Different researchers recommend different minimum CVI values depending on the number of experts:
| Experts | Min CVI Threshold | Source | Interpretation |
|---|---|---|---|
| 2 | 0.80 | Davis (1992) | Acceptable validity |
| 3-5 | 1.0 | Polit & Beck (2006) | Perfect agreement needed |
| 6-8 | 0.83 | Lynn (1986) | Good validity |
| 9+ | 0.78 | Lynn (1986) | Acceptable validity |
Why Do Thresholds Vary?
With fewer experts (2-5), the standard is stricter because each expert's opinion carries more weight. As your expert panel grows, the threshold becomes more lenient because you have more data to show statistical consensus.
CVI Limitations and Alternatives
Why CVI is being replaced:
- Loss of information: Averaging hides problems with individual items. A 0.79 CVI could hide that one item was rated 0.20 by experts.
- No item details: You don't see which specific items need revision
- Doesn't account for item-by-item variability: Different expert patterns produce same CVI
Preferred alternatives:
- S-CVI/Ave: Average of I-CVI values (more modern, recommended)
- S-CVI/UA: Percentage of items with universal expert agreement
CVI vs S-CVI/Ave
| Metric | Formula | Interpretation | Threshold | Modern Use |
|---|---|---|---|---|
| CVI | Mean of CVR values | Scale validity (rough) | 0.78-1.0 | Legacy/reference |
| S-CVI/Ave | Mean of I-CVI values | Scale validity (precise) | ≥ 0.90 | Current standard |
Practical Interpretation
If CVI = 0.90
✓ Your scale has excellent content validity. Proceed with confidence to the next phase of validation (e.g., pilot testing, reliability analysis).
If CVI = 0.78-0.89
⚠ Your scale has acceptable validity, but there may be room for improvement. Consider revising items with lower individual CVR values before proceeding.
If CVI < 0.78
✗ Your scale needs more work. Review the CVR for each item, revise low-scoring items, and consider another expert panel review.
References
Davis, L. L. (1992). Instrument review: Getting the most from a panel of experts. Applied Nursing Research, 5(4), 194-197.
Lynn, M. R. (1986). Determination and quantification of content validity. Nursing Research, 35(6), 382-385.
Polit, D. F., Beck, C. T., & Owen, S. V. (2007). Is the CVI an acceptable indicator of content validity? Appraisal and recommendations. Research in Nursing & Health, 30(4), 459-467.