| Digit | Count | Observed % | Expected % | Deviation | Z-Score | Flag |
|---|---|---|---|---|---|---|
| 1 | 314 | 25.18% | 30.10% | -4.92% | -3.82 | Flagged |
| 2 | 228 | 18.28% | 17.61% | +0.67% | 0.62 | Normal |
| 3 | 150 | 12.03% | 12.49% | -0.46% | -0.50 | Normal |
| 4 | 127 | 10.18% | 9.69% | +0.49% | 0.59 | Normal |
| 5 | 138 | 11.07% | 7.92% | +3.15% | 4.14 | Flagged |
| 6 | 81 | 6.50% | 6.69% | -0.19% | -0.27 | Normal |
| 7 | 75 | 6.01% | 5.80% | +0.21% | 0.33 | Normal |
| 8 | 60 | 4.81% | 5.12% | -0.31% | -0.51 | Normal |
| 9 | 74 | 5.93% | 4.58% | +1.35% | 1.23 | Normal |
The first-digit distribution of 1,247 numerical disclosures extracted from Vaxart's FY 2024 filings yields a Mean Absolute Deviation (MAD) of 0.018 against Benford's Law expected frequencies. This places the company in the "marginally acceptable" range, above the 0.015 conformity threshold but below the 0.022 nonconformity threshold established by Nigrini (2012).
Two digits show statistically significant deviations at the 95% confidence level. Digit 1 appears at 25.18% versus the expected 30.10%, a deficit of nearly five percentage points with a z-score of -3.82. This under-representation of leading 1s can indicate truncation or selective omission of smaller values. Conversely, digit 5 appears at 11.07% versus the expected 7.92%, an excess of 3.15 percentage points with a z-score of 4.14. Excess leading 5s often correlate with rounded estimates, settlement figures, or threshold-anchored disclosures.
The chi-squared test yields a p-value of 0.041, borderline at the conventional 0.05 significance level. While this does not constitute proof of manipulation, the combination of a marginally acceptable MAD with two individually flagged digits suggests that specific categories of reported figures may benefit from targeted review. Analysts should examine whether the digit 5 excess concentrates in R&D expense estimates, milestone payments, or other areas where round-number estimation is common in pharmaceutical disclosures.