Quality Control visits are repeat interview done by either a supervisor or a specially designated quality control fieldworker, as a way of cross- checking the accuracy and completeness of data. It is also a way of checking if the fieldworker actually interviewed the correct respondents. Quality control visits may be done on 5% of the households in each round of data collection. It is important to do quality control visits to as close as possible to the actual fieldworker data collection visit.

As field workers operate with a high degree of independence and flexibility, random and unannounced spot check visits may be used to check on time management and rate of data collection activities. Spot check visits also provides an opportunity for field workers to interact with their supervisors over environmental, field safety issues and the opportunity for receiving more field supplies.

In addition to instilling into field workers the value of checking their own work for completeness and accuracy, all data forms have to be re- checked by a supervisor before they are submitted for data entry. Forms with omissions and obvious errors are returned to the fieldworker for correction or revisits. At the point of data entry a further checking of the forms is performed and forms that have errors or inconsistencies are returned to fieldworker via his/her supervisor.

If a query is detected in the field office and cannot be resolved in the field office, a query form should be issued to the fieldworker. To confirm that the fieldworker goes back to the household; include other questions which may not change overtime to the query. These additional questions are referred to “red hearings” and they are means of ensuring that field workers do not answer the queries without visiting the household. Also faulty data can be generated periodically by the field office and used to test field capacities to detect problems and errors. This will help supervisors to intensify their supervision of poorly performing field workers.

To ensure quality data at each round, 5% of all visits by field workers should be visited by supervisors to conduct blind re- interviews.  Where an event is reported for example birth event, the accompanying birth form should be administered. The answers obtained from this interview should be compared to the one previously administered by the fieldworker.  This should be from the same respondent as the original interview if possible and should be as close in time to the original interview as possible.  This is a tool for monitoring the performance of individual field workers on how they are asking specific questions or to monitor if a visit was actually made to a household. The fieldworker should not know beforehand which household will be included for the re-  interview. The supervisor should contact the fieldworker following the re- interviews to review the results.  Errors should be discussed and appropriate concepts and procedures should be reviewed.  If the extent and type of errors warrant an additional observation, retraining of field workers should be done since it is possible other field workers may be making similar or the same mistakes.

In order for all expected QC procedures to be properly followed the field team must first crosscheck that the required blank forms, pre- printed forms, printed labels for forms and samples and any other required tools for data and sample collection have been prepared for the field trip. There is a need to check that all summary sheets needed and error detection sheets are packed.

The team leader must first cross check with the pre- printed household information to confirm that the members being visited are the actual ones that were previously enumerated in the household. The use of photographs taken during previous study rounds is essential. Any updates of the census form should be a priority before interviews begin. It’s the duty of the team leader to do this. Thereafter the team leader would assign respondents to members of his/her team using same sex arrangement. The list with the respondents should be updated for any new ones in households or new households. Each assignment of a respondent to an interviewer should be recorded.

The interviewer must first confirm with respondent the information on the census form and any edits should be done at this juncture before initializing the interview.

The interviewer must follow the flow of the questionnaire from the first question to the last, following skip patterns and instructions (like prompting/no prompting). When interview is completed the interviewer must check for completeness of the questionnaire, clarity of codes and open-ended answers. Clarification of any of these should be made before respondent is signed off. Once the checks have been completed then the interviewer must sign the forms and hand them over to the field editor.

The interviewers have been trained as phlebotomists and should collect the samples as soon as the interview is through from those who agree to this. There is a summary sheet that the interviewer should use to record the type of forms administered and the samples collected.

The field editor should crosscheck that a complete set of forms/questionnaires has been handed over to him/her with the interviewer’s signature. Then editor should next check completeness of forms, skip patterns have been followed and the coding is eligible. In case there is a clarification needed this should be followed up immediately with the interviewer. Its only when all the forms have been ascertained to be completed that the editor signs them. They are then handed over to the team leader

All errors that are identified by the field editor must be recorded on an error detection sheet with information on the type of form with the error, what the error was, who made the error, possible reasons for the error, whether the error was resolved and if not resolved must be discussed with team leader who may forward this to study coordinator and data management, the editor’s signature and date. These error detection sheets help identify areas in the questionnaire where most interviewers find problems. They also help identify what each interviewer mostly finds a problem. The error information is used during re- training.

Since each team leader would have assigned an interviewer to a respondent and recorded this, once the set of forms is received from the editor then this needs to be recorded.  By so doing the team leader will know at the end of the day how many were assigned to respondents and how many form sets have not been received. This should be cross checked straight away before leaving the fields.

The QC person on the field team should purposively select at least 5- 10% of those interviewed and re-interview them using a shorter version of the questionnaire which has questions that require definite answers that should not change on that day. This should be done as soon as the editor is through with an interviewer. The QC person then checks with the team leader for the questionnaires corresponding to what s/he has administered and cross checks the answers. Any inconsistencies need to be resolved by the QC, Editor, Interviewer and the respondent.

The QC and the team leader need to decide on resolutions concerning any issue that is not resolved in the above step.  The QC and team leader also have to tally the number of questionnaires and samples received at the end of the day with the interviewer assignment lists and the summary lists. Any unresolved issue needs to be discussed before the teams leave the field.

This should be the first activity as soon as the teams return from the field. Cross checking summary sheets, counting forms and samples and recording all this on log forms. Once all tally then they are handed over to data management and laboratory. All those involved in logging and receiving must sign the logbook.