The dependent variable in the models is the
type of payment scheme. In the first model, the
dependent variable is binary, taking on a value of
1 if an hourly wage or salary is used and 0 if a
piece rate is employed. For dichotomous dependent variables, both probit and logit regression
methods are generally used. The probit model
assumes a normal cumulative density function
.(see Maddala, 1983; Greene, 1993) and is used in
this study. In the second model, the dependent
variable is polychotomous but ordinal in the sense
that the piece rate provides the least opportunity
for shirking with respect to quantity and performance speed, while the opposite is true for the
salary scheme; the hourly wage lies somewhere in
between these. Thus, an ordered probit regression is used to examine the effects of the explanatory variables on choice of payment scheme.
The dependent variable in the models is thetype of payment scheme. In the first model, thedependent variable is binary, taking on a value of1 if an hourly wage or salary is used and 0 if apiece rate is employed. For dichotomous dependent variables, both probit and logit regressionmethods are generally used. The probit modelassumes a normal cumulative density function.(see Maddala, 1983; Greene, 1993) and is used inthis study. In the second model, the dependentvariable is polychotomous but ordinal in the sensethat the piece rate provides the least opportunityfor shirking with respect to quantity and performance speed, while the opposite is true for thesalary scheme; the hourly wage lies somewhere inbetween these. Thus, an ordered probit regression is used to examine the effects of the explanatory variables on choice of payment scheme.
การแปล กรุณารอสักครู่..
