Abstract
Ruth B. McKay (1994) "Touch-Tone Data
Entry For Household Surveys: Research Findings And Possible
Applications," Proceeding of the Section on Survey Method
Research, American Statistical Association, 509-511.
In the mid-1980's, the Bureau of Labor Statistics (BLS)
pioneered the application of touch-tone data entry (TDE) to
business establishment surveys. The success of TDE to improve
response rates and timely reporting for the Current
Employment Statistics (CES) has been well-documented (Clayton
et al, 1989). More recent work has documented increased
respondent satisfaction and reduced survey costs for TDE
reporters in the CES (Rosen et al, 1993).
The success of TDE for establishment surveys, not only at BLS
but the Bureau of the Census (Bond et al, 1993), led us to
explore the possible application of this mode of data
reporting to household surveys. In late 1991, work was begun
on a pilot TDE instrument for household survey reporters. The
BLS survey utilized for this research was a short form of the
redesigned Current Population Survey (CPS). The wording of
the questions in the short form of CPS is identical to that
in regular CPS.
The short form of the CPS questionnaire asks the minimum
number of questions necessary to assign household members to
one of three labor force categories: Employed; Unemployed;
Not In Labor Force. The shortest possible series, for those
who did any paid work last week, is 2 questions long.
Approximately 58% of respondents will be classifiable as to
labor force status after answering the 2-question series.
The CPS Touch-tone Data Entry program allows the respondent
to answer open-ended questions as well as those be answered
by "yes" and "no." Cognitive research on
the pilot version indicated that respondents of
widely-differing educational status followed the same rules
as CPS interviewers in coding their open-ended responses.
After the respondent has answered the minimum number of
questions necessary to be classified as to labor force
status, the program assigns labor force status to that
respondent.
Thus, if a respondent indicates that he/she could not have
returned to work last week, the program will request that the
respondent state the reason for this inability*. After
speaking the reason into the telephone, the respondent will
be asked to self-code his/her open-ended answer as to why
he/she could not have returned to work. If the respondent
codes the open-ended answer as meaning that he/she was ill,
and presses "1," the program classifies the
respondent as "Unemployed." If the respondent codes
the answer as "2," a reason other than illness, the
program classifies the respondent as "Not In Labor
Force."
The first phase of cognitive laboratory research on the
CPS-TDE instrument was conducted in 1993. Nineteen
respondents were interviewed using the short form of the CPS
questionnaire, following which the respondents were asked to
complete the same survey by keying their responses to a
telephone version of the interview. Two respondents
unwittingly pressed incorrect keys during their data entry,
found themselves on the wrong paths, and were not able to
complete their reports. In addition, respondents had a great
deal of difficulty in recognizing the "line number"
when using the "TDE Reporter's Guide" to report
their household roster. The TDE instrument and "TDE
Reporter's Guide" were revised to correct these
problems.
Last Modified Date: July 19, 2008
|