Table 6.

Morgan-Russell subscale E3 (Personal contacts); dose–response analysis in the AN group

	Personal contacts																Analysis for time trend
																	Doubly ordered R × C tables
																	Linear-by-linear association test
	AN Study 2				AN Study 3				AN Study 4				AN Study 5				Test statistic	Exact P
Dose	Response category				Response category				Response category				Response category
	0	4	8	12	0	4	8	12	0	4	8	12	0	4	8	12
Never ASD	2	5	5	22	2	9	6	17	0	6	4	24	0	13	2	15	−0.6536	0.5
ASD × 1–3	2	3	2	3	1	5	1	3	0	5	2	3	2	7	1	0	−1.117	0.3
ASD × 4	2	3	1	0	1	5	0	0	2	2	2	0	2	3	0	1	+0.4708	0.7

Dose–Response analysis	Linear-by-linear association test				Stratified R x C contingency tables
Doubly oredered R x C tables					Cochran–Mantel–Haenszel test
Test statistic	−3.462	−2.881	−3.851	−3.095	Test statistic—6.798
Exact P	0.0004	0.004	0.00009	0.0015	exact p < 0.00001

AN: anorexia nervosa; R × C: rows and columns; AN study 2: 6-year follow-up; AN study 3: 10-year follow-up; AN study 4: 18-year follow-up; AN study 5: 30-year follow-up; ASD × 4: Individuals in the AN group with an ASD diagnosis at AN Study 1 to AN Study 4 (all four examinations) (n = 6); ASD × 1–3: those who had been assigned an ASD diagnosis at least once and at most three times (n = 10); never ASD: those who had never fulfilled criteria for ASD (n = 34)

Raw data are situated in the middle of the table. Results of the initial omnibus test, a stratified Cochran–Mantel–Haenszel test for a doubly ordered R × C table, are placed in the lower right corner. Results of the analyses for time trend, the Linear-by-Linear Association test, are placed to the right of each level of ‘dose’. Below each follow-up study we have placed the results of the ‘dose–response’ analyses, using the Linear-by-Linear Associations test