Agriculture Reference
In-Depth Information
total nonresponses, and we should emphasize the importance of returning complete
questionnaires. Adequate attention should also be placed on the proper attitude for
helping respondents during the interview (memory aids, field measures, or a proper
use of administrative documents) without influencing the answers. In the case of
areal surveys, it is important to establish details of the route taken and to identify the
units to enumerate.
Practical exercises and discussion groups should be prepared to standardize the
behavior of enumerators. A set of checks should be defined to assess the complete-
ness of the questionnaire and to identify apparent inconsistencies.
To deal with the cases in which respondents are not immediately found or
identified, it is often necessary to prepare a plan for subsequent contacts (this
recommendation is particularly valid in the case of telephone surveys). One should
also ensure that the contacts (by telephone or in person) take place at different times
and days of the week (even weekends), and that supervisors have been appointed to
monitor the enumerators and procedures. A plan must be prepared for ensuring that
the interviews or observations have been made by returning to the field. Regular
meetings with the enumerators should be used to highlight and solve any issues not
covered in the planning phase. Moreover, if the budget and the organization allow
it, computer-assisted data collection (CATI, CAPI) methods should be used.
After the survey, it is a good practice to send a letter to the farmers and land
owners to thank them for their participation and to remind them to send back any
remaining completed questionnaires. An appropriate reminder plan can reduce
nonresponses, and it is typically reliant on the availability of a telephone number
for each non-respondent.
Finally, there are some techniques that can assess the quality of the data. We can
collect and analyze the number of contacts needed to get an accurate answer or the
reasons for nonresponses, calculate the nonresponse rate on the total sample and on
specific subsets, record the nonresponse rate and calculate the trends over time,
acquire and analyze information on non-respondents, or use administrative data and
external sources to more efficiently collect the data.
During the registration process, the collected information is converted into a
digital format. The data is transferred into the computer by typing exactly what
appears on the paper questionnaire. This conversion is normally carried out by
unskilled personnel, as it does not require any survey knowledge. For this reason,
the registration phase must be considered a significant source of potential errors.
The operations can be performed according to different organizational arrange-
ments, characterized by the degree of standardization and control of the personnel
involved. These activities are frequently distributed throughout the territory and
delegated to the bodies that were responsible for local data collection. Even the
technical methods may vary. In the simplest case, the operator enters the data into a
form that does not check for errors, or, in controlled recording, the operator is
notified if there are entry errors such as leaving a required field blank.
It is clear that some errors may still remain even after accurate controls (for
example, the year 1978 may be recorded instead of 1987). In some cases, the
registration process may be absent—for example, when the survey is computer-
Search WWH ::
Custom Search