Agriculture Reference
In-Depth Information
assisted (CATI, CAPI). In other cases, this operation can be replaced by optically
reading the questionnaires. Then, the operator supervises the operations performed
by a machine. Errors in the registration phase are defined as any difference between
what was recorded and what was reported on the questionnaire. These errors can be:
1. Amounts wrongly divided or multiplied by fixed factors (usually 100 or 1,000).
2. The exchange of codes (15 instead of 51) or mistyped keys contiguous on the
keyboard (A instead of S).
3. Misalignments, that is, errors caused by omissions, which typically imply that all
the subsequent data are recorded in the wrong position creating a sequence of
registration errors.
Statistically, these errors become more serious as the importance of the variables
involved increases. In this context, the most important errors are probably those that
affect the unique identification of the statistical units, or their placement in the
correct stratum or domain (Fowler
2009
).
The registration process must be designed by providing accurate definitions of
operational procedures, personnel training, and the quality control at each step. The
errors introduced in this particular phase may depend, at least in part, on the design
and implementation of the previous steps.
The design of the record must consider the true range of each variable and the
record layout must be accompanied by a list of eligible codes. Additionally, there
should be a code that unambiguously indicates a nonresponse.
Wherever possible, a controlled recording should be used so that serious errors
can be immediately identified and corrected. However, it is important to stress that
an error identified by data entry software must only be corrected if it was introduced
by the recording operator. Otherwise, the error must be recorded to avoid introduc-
ing a correction that the operator cannot perform. For this reason, data entry
software must report the errors, but should not prevent their entry.
Operator training is one of the most important aspects. This preparation should
be accompanied by examples and exercises to test the understanding of instructions
and survey procedures. Operators should also be made aware of the quantity and
quality of their work while it is being performed.
Training has a large impact on the quality of results obtained from a survey. It is
important to effectively train interviewers, because they have many duties and are
the main connection between the respondents and the researcher. By comprehen-
sively training office staff, we should enable them to process the survey question-
naires accurately and quickly. This training can be provided with the aid of manuals
containing instructions for surveyors, formal training courses, and on-the-job train-
ing. General training for interviewers may include information on the purpose and
scope of the survey, the sampling approach, the questionnaire format, the recording
of responses, interviewing techniques, ways for avoiding or reducing nonresponses
and maintaining respondent co-operation, field practice, quality control, editing,
planning the workload, and administrative issues.
Search WWH ::
Custom Search