Collecting big data with small screens: Group tests of children’s cognition with touchscreen tablets are reliable and valid

Table 1 Descriptive statistics for tablet cognitive assessments

Task name	N	Predictive validity				Reliability			Time taken
Task name	N	r	LB	UB	b*	ω	LB	UB	10%	50%	90%
Visual Search Speed	526	– .38	– .46	– .30	– .35	.74	.68	.78	1.59	2.18	3.30
Verbal Short-Term Memory	532	.43	.35	.51	.40				2.28	4.36	6.82
Spatial Short-Term Memory	532	.45	.37	.52	.41				1.55	3.05	5.30
G/N-G - D'	362	.18	.06	.30	.19	.53	.43	.62	1.70	4.32	4.68
G/N-G - Omission Errors	362	– .41	– .50	– .30	– .36	.71	.65	.77	1.70	4.32	4.68
G/N-G - Commission Errors	362	– .13	– .24	– .01	– .07	.60	.51	.67	1.70	4.32	4.68
Matrix Reasoning	515	.48	.40	.55	.46	.76	.73	.78	2.55	4.17	6.38
Reading Fluency	535	.63	.57	.69	.66	.87	.85	.89	3.15	3.20	3.32
Rhyme Judgement	488	.27	.18	.36	.26	.68	.59	.76	1.11	3.52	4.25
Phonological Discrimination	275	.26	.12	.39	.23	.72	.67	.78	4.73	5.32	6.54
Arithmetic Fluency	535	.55	.48	.61	.52	.89	.87	.91	3.48	3.53	3.65
Non-Symbolic Num. Discrim.	515	.45	.37	.53	.42	.85	.83	.86	3.32	4.00	4.72
Line Estimation	193	.53	.42	.63	.53	.92	.90	.94	2.20	3.22	4.44
Liquid Equalization	348	.42	.32	.51	.41	.88	.85	.90	3.02	3.53	4.21

Note: this table only includes data from the school cohort. Predictive validity was assessed by the linear Pearson correlation with teacher-rated academic ability, along with the lower and upper bounds of a 95% confidence interval (LB, UB). We also report the standardized regression coefficient (b*) for each task predicting academic ability whilst accounting for age and normalized neighborhood deprivation, in a multivariable regression performed separately for each task. The 10%, 50% (median) and 90% percentiles of time taken in minutes to complete each task are reported in the final three columns.