ciated with no residency training is
expð 0 : 3041 Þ¼ 1 : 36Those without previous training is 36% more likely to receive the same number
of complaints as those who were trained in the specialty.
Note:An SAS program would include these instructions:
DATA EMERGENCY;
INPUT VISITS CASES RESIDENCY;
LN = LOG(VISITS);
CARDS;
(Data);
PROC GENMOD DATA = EMERGENCY;
CLASS RESIDENCY;
MODEL CASES = RESIDENCY/DIST = POISSON LINK = LOG
OFFSET = LN;
where EMERGENCY is the name assigned to the data set, VISITS is the
number of visits, CASES is the number of complaints, and RESIDENCY is the
binary covariate, indicating whether the physician received residency training
in the specialty. The option CLASS is used to declare that the covariate is
categorical.
10.3.2 Multiple Regression Analysis
The e¤ect of some factor on a dependent or response variable may be influ-
enced by the presence of other factors through e¤ect modifications (i.e., inter-
actions). Therefore, to provide a more comprehensive analysis, it is very desir-
able to consider a large number of factors and sort out which are most closely
related to the dependent variable. This method, which is multiple Poisson
regression analysis, involves a linear combination of the explanatory or inde-
pendent variables; the variables must be quantitative with particular numerical
values for each observation unit. A covariate or independent variable may be
dichotomous, polytomous, or continuous; categorical factors will be repre-
sented by dummy variables. In many cases, data transformations of continuous
measurements (e.g., taking the logarithm) may be desirable so as to satisfy the
linearity assumption.
TABLE 10.3
Variable Coe‰cient Standard Error zStatistic pValue
Intercept 6.7566 0.1387 48.714 <0.0001
No residency 0.3041 0.1725 1.763 0.0779
360 METHODS FOR COUNT DATA