BMC Medical Research Methodology

Table 1 Estimated numbers of required subjects for 2, 3, 4 and 5 observations with ICC = 0.1

From: Sample size determination for mediation analysis of longitudinal data

Observations
	2			3			4			5
Effect size^a	Sobel	Product	Bootstrap	Sobel	Product	Bootstrap	Sobel	Product	Bootstrap	Sobel	Product	Bootstrap
SS	365	299	304	257	215	215	207	163	169	169	136	142
SH	272	235	237	194	172	180	151	140	150	133	126	131
SM	248	226	230	188	188	200	146	144	145	132	126	130
SL	238	248	251	176	176	176	148	147	149	124	126	128
HS	238	201	209	161	138	143	123	104	108	99	83	85
HH	109	88	94	78	65	69	61	50	57	51	42	40
HM	87	77	99	58	55	57	49	42	52	42	40	40
HL	74	72	73	57	51	25	47	46	20	40	38	39
MS	215	200	209	138	134	140	110	102	105	86	83	85
MH	79	65	69	54	46	46	42	35	36	35	29	31
MM	51	41	44	38	30	33	29	23	28	24	21	23
ML	40	36	38	29	27	28	24	22	25	20	18	20
LS	204	206	204	139	132	140	105	101	103	83	82	82
LH	65	60	69	44	41	40	34	31	32	28	24	24
LM	36	31	33	25	22	24	19	17	18	17	15	15
LL	24	20	22	17	16	15	14	13	12	12	11	11

^aEffect size: The first letter is the size of β_a, the second letter is the size of β_b; S is small (0.14), M is medium (0.39), L is large (0.59) and H is halfway (0.26) between large and medium effect sizes

Back to article page

ISSN: 1471-2288

Contact us

General enquiries: journalsubmissions@springernature.com