Table 4.

Contributions of each moderating variable to account for variance in effect sizes in non‐human mammal studies

Moderator	F	Df1, Df2	p
Training and testing
Log training length	1.121	(1, 6)	.331
Test response	1.262	(1, 6)	.304
Surface‐level properties
Categories in language	0.760	(1, 6)	.418
Number of unique vocabulary items	0.365	(1, 6)	.567
Structural properties
Non‐adjacent dependencies	0.111	(1, 6)	.750