Table 3.

Training Effects on Proximal and Primary Outcomes^*

	Memory Training		Reasoning Training		Speed Training		Control†
Measure	Net Effect Size (P Value)‡	Showing Reliable Improvement, %§	Net Effect Size (P Value)‡	Showing Reliable Improvement, %§	Net Effect Size (P Value)‡	Showing Reliable Improvement, %§	Showing Reliable Improvement, %§
Proximal Outcome Composites
Memory
Posttest	0.257 (<.001)	26	−0.009	17	−0.012	13	15
A1	0.212 (<.001)	22	−0.011	11	−0.021	12	14
A2	0.174 (<.001)	40	−0.03	27	−0.052	28	29
Reasoning
Posttest	−0.018	34	0.480 (<.001)	74	−0.026	35	39
A1	0.021	34	0.402 (<.001)	63	−0.003	29	31
A2	0.045	36	0.257 (<.001)	53	−0.019	30	35
Speed∥
Posttest	−0.045	34	0.003	33	−1.463(<.001)	87	31
A1	−0.054	35	−0.033	34	−1.212 (<.001)	81	32
A2	−0.034	36	−0.043	35	−0.867 (<.001)	73	37
Primary Outcome Composites
Everyday problem soMng
A1	−0.045	19	0.03	23	0.008	20	21
A2	−0.073	21	−0.027	25	0.031	26	23
ADL and IADL functioning∥
A1	0.02	17	−0.125	18	−0.05	14	16
A2	0.017	17	−0.056	16	−0.07	17	17
Everyday speed∥
Posttest	−0.091 (.02)	31	0.004	22	−0.016	26	27
A1	−0.041	31	0.05	28	0.001	31	30
A2	−0.007	33	0.03	29	−0.009	30	29
Driving habits∥¶
A1	0.017	19	−0.052	19	0.068	16	19
A2	0.085	16	0.079	16	0.077	16	18

^*Only significant P values reported. A1 indicates first annual evaluation; A2, second annual evaluation; ADL, activities of daily living; and IADL, instrumental activities of daily living.

^†Net effect of the control is 0 at all time points, since net effect of group is defined as (group mean-control mean at time point)-(group mean-control mean at baseline).

^‡Net difference divided by intrasubject SD (see “Methods” section).

^§Calculated as the percentage of participants in each group who were ≥1 SEM above baseline.

^∥Favorable response is in the negative direction.

^¶For self-reported drivers only.