Table 4.
Logistic regression predicting referral to speech‐language therapy services from demographic and child variables. Female sex, severity of language disorder and increased number of speech sound errors significantly increase likelihood of referral to specialist speech‐language therapy services
| Odds ratio | SE | p | 95% CI | |
|---|---|---|---|---|
| Age (months) | 1.04 | 0.04 | .33 | 0.96, 1.14 |
| Male sex | 0.40 | 0.16 | .02 | 0.19, 0.86 |
| Income Deprivation Affecting Children Index (rank score) | 1.00 | 0.00 | .59 | 1.00, 1.00 |
| Language composite (z‐score) | 0.38 | 0.10 | .00 | 0.23, 0.64 |
| Speech (per cent consonant correct) | 0.80 | 0.03 | .00 | 0.74, 0.87 |
| NVIQ composite (z‐score) | 1.10 | 0.22 | .65 | 0.74, 1.63 |
| Strengths & Difficulties: total difficulties (raw score) | 1.02 | 0.03 | .52 | 0.96, 1.08 |
| Early Years Foundation Stage Profile (raw score) | 0.99 | 0.03 | .78 | 0.94, 1.05 |
| Constant | 1.37 | 5.73 | .00 | 3,656.69, 5.10e+10 |
The overall model is significant, Wald χ 2 (8) = 59.04, p < .001, and explained a significant, though modest, amount of variance (McFadden's pseudo R square = .28).