Language delay aggregates in toddler siblings of children with autism spectrum disorder

Table 4 Group moderation of relationships between social performance and language

	Unstandardized coefficients		Standardized coefficients	T	Sig.	95.0% confidence interval for B
	B	Std. error	Beta	T	Sig.	Lower	Upper
Receptive language model
(Constant)	53.90	2.85		18.90	< .001	48.29	59.51
Sex	3.21	1.05	.16	3.05	.002	1.14	5.28
Maternal education	2.23	1.27	.098	1.76	.080	− .27	4.73
Income	3.60	1.13	.18	3.19	.002	1.38	5.82
Race	− 2.61	1.54	− .087	− 1.70	.090	− 5.63	.41
ASD risk status	− 6.41	1.52	− .30	− 4.21	< .001	− 9.40	− 3.41
Social affect	− 1.35	.41	− .32	− 3.33	.001	− 2.14	− .55
Social affect by risk status	.92	.48	.21	1.92	.056	− .022	1.86
Expressive language model
(Constant)	46.42	3.16		14.68	< .001	40.20	52.64
Sex	3.15	1.17	.14	2.70	.007	.86	5.45
Maternal education	4.06	1.41	.16	2.88	.004	1.29	6.82
Income	2.76	1.25	.12	2.21	.028	.30	5.22
Race	.16	1.70	.005	.096	.92	− 3.18	3.51
ASD risk status	− 6.41	1.69	− .27	− 3.80	< .001	− 9.73	− 3.094
Social affect	− 1.71	.45	− .37	− 3.80	< .001	− 2.59	− .82
Social affect by risk status	1.68	.53	.35	3.16	.002	.63	2.72

These parameters involve the fourth and final step in a hierarchical linear regression model in which either Mullen receptive or expressive language score is the dependent variable. Step 1 consists of the covariates sex, maternal education, and income; step 2 introduces autism spectrum disorder (ASD) risk status; and step 3 introduces social affect score on the Autism Diagnostic Observation Schedule, which measures autistic social deficits. The fourth step introduces an interaction term for ASD risk status and social affect. Bolded variables demonstrate a significant relationship with language. Social affect score is a significant contributor to variation in both receptive and expressive language. The interaction is significant for expressive language and shows a trend for receptive language. std. standard, sig. significance

ISSN: 1866-1955