Table 3. Probit regression testing the associations between LM training for MH (y/n) and organisational-level sickness absence trends due to mental ill-health.

Outcomes	Results
Presence of sickness absence due to mental ill-health (n = 3385)	β .160 (.0471)
	LR chi2 286.805***
	Log likelihood –493.883
Proportion sickness absence due to mental ill-health (n = 1116)	β -.077 (.0525)
	LR chi2 41.654***
	Log likelihood –414.569
Repeated sickness absence due to mental ill-health (n = 3566)	β .027 (.0803)
	LR chi2 49.490***
	Log likelihood -339.804
Proportion of long-term sickness absence due to mental ill-health (n = 3566)	β -.132* (.0577)
	LR chi2 388.557***
	Log likelihood –390.677

Note 1: Analysis controlled for wave, sector, size, and age of organisation.

Note 2: Standard error placed in brackets.

Note 3: LR chi² = Likelihood ratio chi-square.