Skip to main content
. 2012 Oct;50(12):2773–2788. doi: 10.1016/j.neuropsychologia.2012.07.031

Table 9.

Regression coefficients of all predictor variables in the length-based model with the SES residual and the SES-based model with the WL residual in predicting PA patients' reaction times.

Word
Nonword
Length-based
SES-based
Length-based
SES-based
Beta P Beta P Beta P Beta P
WL .169 <.001⁎⁎ .232 <.001⁎⁎
OrthN −.050 .036 −.049 .038 .013 .703 .012 .732
SES .172 <.001⁎⁎ .237 <.001⁎⁎
Group .617 <.001⁎⁎ .617 <.001⁎⁎ .642 <.001⁎⁎ .643 <.001⁎⁎
WL_Group .185 <.001⁎⁎ .201 <.001⁎⁎ -
SES_Group .191 <.001⁎⁎ .202 <.001⁎⁎
LnCE_Freq −.131 <.001⁎⁎ −.131 <.001⁎⁎
SES residual .033 .033 .056 .015
WL residual .013 .390 .017 .455



R2 (%) 47.0 47.2 48.9 48.9

Note: ⁎⁎Correlation is significant at the .01 level; Correlation is significant at the .05 level; Beta coefficients presented were based on the standardized scores. WL: word length; OrthN: orthographic neighbourhood size; SES: summed error score; Group: severity group; SES_Group: the product of summed error score and group; WL_Group: the product of word length and group; Ln_CE_Freq: log CELEX frequency; SES Residual: residual of regressing SES by WL; WL Residual: residual of regressing WL by SES.