Previous PageTable Of ContentsNext Page


Field data collection


Field data collection

Field enumeration is an interactive process between an enumerator and a respondent. It is prone to many sorts of non-sampling errors. Some of these errors may be related to the sex of the interviewer, errors occurring owing to the presence of outsiders at the interview and to other factors. The "contamination effect" occurs when neighbours are present and are influenced by the survey topics and the responses given during the interview. Later, if they are interviewed, they will naturally be affected by the previous experience and their responses may be biased.

Another important error concerning third persons is called the "presence of others effect". The presence of other persons during the interview will affect the responses given by the respondent, compared with interviews carried out in the absence of others.

The "response distortion" will vary according to different survey topics. For example, in some countries it has been found that in the presence of others livestock numbers are generally overreported while the agricultural production is underreported.

In order to avoid such errors occurring during the interviews, it is recommended that interviewers be trained to insist that the interviews should not be carried out in the presence of others.

Depending on the topic of the survey and the specific questions, a gender bias may occur for both sexes, relating to the gender of all parties (respondent, interviewer and other persons present) involved during the interview.

Pilot surveys

Pilot surveys should be designed, in particular, to detect possible gender biases corresponding to the questionnaire items and the variability of survey data by sex, as well as for estimating the cost, timing, coverage, non-response, and enumeration workload of the survey operation. Possible gender biases should be taken into consideration to test these operations.

In the light of the information collected during the pilot surveys, gender-related biases concerning those aspects of the data collection operation should be recognized and controlled.

Previous PageTop Of PageTable Of ContentsNext Page