AT&T Telemarketing Data
SUMMARY:
The
market.survey
data frame has information
about 1000 households (rows).
The 10 variables (columns) include demographic information
about the household and information specific to the households
telephone service.
ARGUMENTS:
- pick
-
factor indicating whether the household picked AT&T as their long
distance phone company.
The level
ATT
indicates they did,
while the level
OCC
indicates
they picked another company.
- income
-
ordered factor indicating the income level of the household.
Levels are:
`<7.5 < 7.5-15 < 15-25 < 25-35 < 35-45 < 45-75 < >75'.
- moves
-
ordered factor indicating the number of times the household moved
in the preceding 10 years.
Levels are:
`0 < 1 < 2 < 3 < 4 < 5 < 6 < 7 < >10'.
- age
-
ordered factor indicating the age level of the respondent.
Levels are:
`18-24 < 25-34 < 35-44 < 45-54 < 55-64 < 65+'.
- education
-
factor indicating the highest education level achieved by
respondent.
Levels are:
HS,
Voc
,
Coll
,
BA
,
and >BA.
- employment
-
factor indicating the type of employment of the respondent.
Levels are:
F
,
P
,
R
,
S
,
H
,
U
,
and
D
.
- usage
-
numeric vector giving the average monthly telephone usage of the household.
- nonpub
-
factor indicating whether the household had an unlisted telephone number.
Levels are:
Y
,
N
,
and
NA
.
- reach.out
-
factor indicating whether the household had participated in a special
AT&T plan (before the forced choice of long-distance carrier).
- card
-
factor indicating whether the household had an AT&T calling card
service (before the forced choice of long-distance carrier).
Levels are:
Y
,
N
,
and
NA
.
SOURCE:
James W. Watson (1986) quoted in
John M. Chambers, John M. and Hastie, Trevor J. (1992).
Statistical Models in S.
Wadsworth and Brooks, Pacific Grove, CA, pg. 49.
The data were compiled from three sources: a telephone interview
survey of 1000 selected households; billing and service records for
those households, and demographic data taken from a separate marketing
database.
EXAMPLES:
survey.fit <- glm(pick ~ log(usage + 2), family=binomial,
data=market.survey)