Interviewers need to routinely be monitored in the field, both for the progress they are making and for their performance.
Most surveys will have a set period for fieldwork and it is good practice to establish milestones for interviewer progress. Where these are not being met there will be a need for intervention. Slow progress may be for work-related reasons –- e.g. working in a difficult area, completing work on other surveys (it is common for interviewers to work on more than one survey at any one time.) -- or due to interviewers’ own personal circumstances (e.g. illness, unexpected events). In some circumstances it may be necessary to transfer work from one interviewer to another. Depending on the extent of the management information (any information relating to the survey and personnel working on it) available, it may be possible to direct interviewers to work on specific cases as the survey period comes to an end. These may be hard-to-interview-at households – flats, households with access controls – which would otherwise be inclined to be under-represented in the collected data.
As well as monitoring progress, it is important to track interviewer performance.
The quality of a survey can be seriously undermined by nonresponse. Nonresponse error creates difficulties, for we often know little about the people or households who do not participate.
However, it is now more common for interviewers to collect some information about non-responding households. This will usually be by observation, for example, the type of property being sampled (flat, detached, semi-detached, terraced etc.) Typically, interviewer performance will be evaluated by the following measures of survey nonresponse which should be calculated according to standard AAPOR definitions.