You can read the entire text in HTML format below or read it
in
Adobe Acrobat.
CONTACT: Mark
Schirmer (502-564-2851)
Sample details
The data for this analysis come from the Current Population Survey
(CPS). The CPS is a monthly survey of about 50,000 households conducted by the Bureau
of the Census.(1) The results reported in Policy Note no. 10 are based on an analysis
of 2001 CPS data. Our sample is comprised of individuals from across the United States,
not just Kentucky. Also, the analysis was limited to individuals 25 years old and
older. The question posed in the CPS survey is:
This year, have you used the Internet to take a course on-line? Approximately
2.2 percent of those in the sample answered “yes.”
Model details
The dependent variable, whether one has taken a course online, is
dichotomous (yes or no). Consequently, we use a multivariate probit model to
estimate the effect of the predictor variables (e.g., income, education, race, gender,
location, age, and having a computer in the home with Internet access). Based on the
model results, we estimate the probability of “the average person” taking a course
online at about 0.5 percent if there is not a computer with Internet access in the
home. On the other hand, if there is a computer with Internet access in the home,
then the probability of taking a course online is 4.3 times higher at 2.3 percent. The
model variables are listed below:
-
Income. The CPS does not report the precise
household family income. Rather, it is reported in 14 broad categories. We
divided these 14 categories into four groups for the analysis: group one
($0-$19,999), group two ($20,000-$34,999), group three ($35,000-$59,999),
and group four ($60,000 and over). In our analysis the first quartile is
omitted from the model and is the reference group.
We also included a variable in the model, MISINC, if the income data
are missing for a household.
-
Education. Educational attainment is collected for
individuals 15 years and older. There are 16 education categories in the CPS, which
range from “less than 1st grade” to doctorate. However, we collapsed these 16
categories into five groups: less than high school, high school, some college,
bachelors, and graduate or professional degree. Each is modeled as a “dummy”
variable, and the high school variable is omitted from the model as the reference group.
-
Race and Ethnicity. We use five variables to
test the effect of race and ethnicity. They are a series of dichotomous variables:
non-Hispanic Whites, non-Hispanic Blacks, Hispanics, Asians, and Native Americans.
The variable “white” is left out of the model and is therefore the reference group.
-
Age. This variable is also modeled as dichotomous
variables: “25 to 49” and “Over 49.” The reference group in the model is “Over 49.”
-
Gender. This variable is equal to 1 for males and
0 for females.
-
Location of Residence. This is a dichotomous
variable indicating whether the residence is in a metropolitan area. The variable
(URBAN) is set to 1 if metropolitan and 0 if nonmetropolitan.
-
Computer with Internet in the Home. This is a
dichotomous variable indicating whether there is a computer with Internet access
in the home. The variable (HCI) is set to 1 if “yes” and 0 if “no.”
The probit coefficients and standard errors for the model are
presented in Table 1. If we hold all variables at their mean and calculate a
predicted probability of taking a course online for the “average” Kentuckian,
the model predicts a probability of .121, which is much lower than the sample
mean of .022.
Table 1: Parameter
Estimates
Footnotes
1. Refer to
http://www.bls.census.gov/cps/cpsmain.htm
for detailed information. Return to text.
2. The analyses were done using individual weights that
approximately equal the inverse of the probability of being in the sample and normalized to
add up to the sample size. This produces the correct standard errors. Return to text.
|
