Skip to main content
. 2015 Jun 1;10(6):e0127650. doi: 10.1371/journal.pone.0127650

Table 4. Regression Results: Outcomes Regressed on S/L Impairment and Cumulative Risk at Each Age.

Multiple Regression Results
Age 5/6 Age 12/13 Age 19/20
Beta SE t R 2 Beta SE t R 2 Beta SE t R 2
Depression
S/L impairment 2.08 1.56 1.80 .08 1.82 1.22 1.49 .07 1.85 1.20 1.55 .06
Cumulative risk 1.62 .43 3.77* 1.27 .47 2.67 + 1.36 0.44 3.09 +
FSIQ
S/L impairment -10.72 1.76 -3.27 + .22 -10.44 1.81 -5.78 + .22 -10.43 1.81 -5.76 + .24
Cumulative risk -3.17 0.69 -4.62 + -2.86 .69 -4.13 + -2.82 .77 -3.68 +
Logistic Regression Results
Age 5/6 Age 12/13 Age 19/20
Beta SE Wald OR Beta SE Wald OR Beta SE Wald OR
School Dropout
S/L impairment .85 .50 4.58 2.35 .52 .43 1.94 1.69 .69 .43 3.40 2.00
Cumulative risk .75 .15 29.74 + 2.13 .65 .17 26.64 + 1.91 .55 .15 18.46 + 1.73
History of Arrest
Sex 1.61 .55 9.99 + 5.02 1.86 .57 13.33 + 6.44 1.79 .58 11.92 + 5.99
S/L impairment .082 .45 0.40 1.09 -.12 .46 .42 .89 -.31 .44 0.83 .73
Cumulative risk .59 .14 20.75 + 1.81 .69 .17 27.71 + 2.00 .67 .17 24.14 + 1.96
Physical Illness
S/L impairment .38 .29 2.44 1.47 .25 3.11 1.19 1.28 .35 .29 1.94 1.42
Cumulative risk .28 .11 8.61* 1.32 .29 .11 10.57 + 1.33 .30 .11 10.37 + 1.36
Poisson Regression Results
Age 5/6 Age 12/13 Age 19/20
Beta SE Wald pR 2 Beta SE Wald pR 2 Beta SE Wald pR 2
Cigarettes per day
S/L impairment .02 .09 .00 .02 .02 .08 .03 .15 -.05 .08 .32 .02
Cumulative risk .11 .03 13.29* .13 .03 21.02 # .07 .03 5.53* .11
Cumulative outcome
S/L impairment .23 .13 2.38 .11 .29 .15 4.00* .09 .26 .15 3.07 .07
Cumulative risk .20 .05 21.97 # .20 .05 18.53 # .18 .05 13.02 #

Note: R 2 refers to variance explained by complete model; Wald = Wald χ2 statistic; OR = odds ratio; S/L impairment = speech/language impairment as assessed at age 5/6 years; pR 2 = pseudo-R 2 as calculated with Coxe, West, & Aiken’s (2009) formula 9. The pR 2 represents how much closer the model is accounting for all the variance as variables are added. (It is not comparable to the R 2 derived from ordinary least squares regression, which represents proportion of variance explained.) We derived pseudo-R 2 by comparing the complete model to a model where no predictor variables were entered.

* p < .05;

+ p < .01;

# p < .005