Methodologies
Data
Quality Assurances
First,
data validation during the interview is handled by the CATI
system, Interviewer. The system prompts interviewers for a
valid response to every question in the survey. For numeric
questions, legitimate ranges of responses are entered into
the computer so that the computer can detect out-of-range
values. When these are detected during the interview, the
computer warns the interviewer that the entered value is out
of range and prompts the interviewers for a legitimate response.
Second,
data validation at the data management level consists primarily
of accounting for all cases in the survey, and ensuring that
a data record exists for every completed interview in the
sample. Data records are passed through a SAS program
to ensure that all data fields are readable, and that all
fields are read in the format specified for that variable.
A separate data cleaning step is also undertaken for the
open-ended verbatim responses. These responses are reviewed
and spell-checked for readability. Moreover, any identifying
information contained in these responses is removed to ensure
the confidentiality and anonymity of respondents.
The final
validation step consists of checking the consistency of respondents
answers to objectives and verifiable survey questions. For
many surveys, this kind of consistency checking is kept to
a minimum, because of the additional analysis costs this
adds. All survey data is backed up and stored on microcomputer
diskettes and mainframe computer tapes for immediate access
and correction, should data corrections be needed.
Top |
 |
