erer (version 3.1)

maTrend: Trend of probability values in a binary choice model

Description

This function computes the change of probability for a continuous variable based on a binary choice model (either probit or logit), and furthermore, stratifies the probability through a binary independant variable.

Usage

maTrend(q, n = 300, nam.c, nam.d, simu.c = TRUE)

Arguments

q

a object of class of "maBina" estimated from maBina().

n

number of points for calculating probability; the large the number, the smoother the curve.

nam.c

a name of a continuous indepedent variable; this must be given for the function to work.

nam.d

an optional name of a binary independent variable; this is used to stratify the probability.

simu.c

Whether simulation data for the continuous variable of nam.c should be used (default is TRUE); if not, the original data for this variable is used.

Value

Return a list object of class "maTrend" with the following components:

q

a list object of class "maBina"

nam.c

the name of a continous variable

mm

matrix of independant variables for all

trend

if nam.d is not specified, this is a data frame of the continous variable, its probability values, standard errors, t values, and p values; if nam.d is specified, the data frame contains the continuous variable, the probability values for all, and two additional probability series stratified by the dummy variable

nam.d

if nam.d is specified, the name of a binary variable .

m1

if nam.d is specified, the matrix of mm with the column value for nam.d replaced by 1

m0

if nam.d is specified, the matrix of mm with the column value for nam.d replaced by 0

trend1

if nam.d is specified, this is a data frame reporting the probability value, standard error, t value, and p value associated with the dummy variable being 1

trend0

if nam.d is specified, this is a data frame reporting the probability value, standard error, t value, and p value associated with the dummy variable being 0

Methods

Two methods are defined as follows:

print:

print the probability output.

plot:

Plot the probability values for a continuous variable. If a strata is specified through nam.d in maTrend(), then the stratified values also are shown.

Details

For a continous variable, its probability values can be computed on the basis of a binary probit or logit model. The trend can be stratified by a binary independent variable. In addition, the standard errors of each probability series are also computed using delta method.

References

Greene, W.H. 2003. Econometric Analysis (5th ed.). Prentice Hall, New York. 1026 P.

See Also

maBina; print.maTrend; plot.maTrend.

Examples

# NOT RUN {
data(daPe)
ma <- glm(grade ~ gpa + tuce + psi, x = TRUE, 
    data = daPe, family = binomial(link = "probit"))
summary(ma)

(ea <- maBina(w = ma, x.mean = TRUE, rev.dum = TRUE))
(ta <- maTrend(q = ea, nam.c = "gpa", simu.c = FALSE))
(tb <- maTrend(q = ea, nam.c = "gpa", nam.d= "psi", simu.c = TRUE, n=100))
plot(ta)
plot(tb)
# }