Table 2.

The estimated sample sizes required to detect a housing effect using between-subject t tests (with alpha = 0.05, and power = 0.8).

Subjects	Judgment bias	Startle amplitude
All hens (not controlling for personality)	164 per group	5 per group
Exploratory hens	85 per group	3 per group
Non-exploratory hens	N/A	4 per group

Estimates are based on the mean differences between housing treatments, pooled standard deviations and the resulting effect sizes (from the first tests run, before the housing switch; see Supporting Material), and were calculated using online calculators, e.g. http://www.sample-size.net/sample-size-means/. The “N/A” for Non-exploratory hens reflects the lack of predicted direction of the effect, thence lack of validity as an affect indicator, for judgement bias in this sub-group.