TABLE A1.

Results of multiple regression relating to Figure 4 in the main manuscript, showing the relationship between time spent per image (continuous response) and various covariates: analysis method (categorical predictor), number of corals per image (continuous predictor), their interaction, and other sources of potential error observer ID (categorical predictor), and Method First (categorical predictor). The reduced and then final models were formed after backward selection removal of non‐significant predictors starting from the full model

Model	Predictor	Estimate	SE	T value	p Value
Full	Intercept	2.29	0.39	5.86	<.001
	Method‐SizeExtractR	0.23	0.02	13.15	<.001
	No. corals	−1.13	0.45	−2.51	<.05
	Observer.ID‐ L.L.	−0.24	0.19	−1.28	>.05
	Method. First‐SizeExtractR	0.29	0.17	1.76	>.05
	No. corals*Method‐SizeExtractR	−0.09	0.02	−4.12	<.001
Reduced	Intercept	2.88	0.31	9.17	<.001
	Method‐SizeExtractR	0.20	0.01	14.53	<.001
	No. corals	−1.48	0.38	−3.93	<.001
	Observer.ID‐Lachs	−0.26	0.16	−1.65	>.05
	No. corals*Method‐SizeExtractR	−0.07	0.02	−3.64	<.001
Final	Intercept	2.61	0.27	9.75	<.001
	Method‐SizeExtractR	0.21	0.01	15.96	<.001
	No. corals	−1.48	0.38	−3.91	<.001
	No. corals*Method‐SizeExtractR	−0.07	0.02	−3.62	<.001

The p Values of significant predictors for on alpha level of .05 are shown in bold.