The Frequency and Degree of “intrusive /r/” in New Zealand English 47
The model effects are plotted in Figures 2.3–2.5 where the y-axes show the pre-
dicted likelihood that an intrusive /r/ will be present. The predicted probability
of insertion is calculated by deriving the log odds of insertion from the coef-
¿ cients in Table 2.3, and then converting the log odds to a probability using the
Table 2.2 Wald Statistics for Model Predicting Likelihood of /r/-insertion
Factor Chi-Square d.f. P
Gender 3.76 1 0.0525
Social class 70.64 6 <.0001
Base 79.93 10 <.0001
Af¿ x 28.09 7 0.0002
Social class x base 18.3 5 0.0026
TOTAL 131.33 19 <.0001
Table 2.3 Coef¿ cients for Model Predicting Likelihood of /r/-insertion
Intercept -4.834077
Gender=Male 0.408924
Class 0.542852
base=claw 2.206757
base=ma -0.830365
base=oprah 0.667502
base=plough -0.887871
base=sofa 3.421577
af¿ x=ify 0.27231
af¿ x=ism 0.028013
af¿ x=ing 1.000898
af¿ x=ish 0.607396
af¿ x=word -1.055286
af¿ x=y 0.570314
af¿ x=ize 0.633545
Social class x base=claw -0.264542
Social class x base=ma -0.002005
Social class x base=oprah -0.424215
Social class x base=plough -0.067133
Social class x base=sofa -0.446194