Skip to main content
Swiss Journal of Economics and Statistics

Table 7 Multiple hypothesis testing adjustment

From: Working from home is here to stay, but how does it affect workplace learning?

Statistically significant estimate of “working from home” in “practical knowledge” specifications

Unadjusted probability value (p-value)

Romano and Wolf (2016) adjusted probability value

Holm adjusted probability values

Column (4) Table 2

0.0062

0.0030

0.0180

Column (4) Table 4

0.0021

0.0020

0.0100

Column (4) Table 5

0.0491

0.0410

0.0480

Column (4) Table 8

0.0250

0.0190

0.0350

Column (4) Table 9

0.0147

0.0090

0.0080

  1. We have run ten regressions in order to investigate the effect of working from home on Practical Knowledge and on Theoretical Knowledge. In five of these specifications (listed in the table), we find a significant effect. The second column of this table shows unadjusted p-values obtained through estimations. The overall type 1 error probability exceeds 5% in this column. Therefore, the third column applies strong FWER control to keep the type 1 error probability under or equal to 5%. The third column shows p-values that have been adjusted, with strong FWER control applied (Romano & Wolf, 2016). The fourth column shows Holm-adjusted p-values
Back to article page