Data Collection

Data collection takes place at the property with a knowledgeable adult who lives there. The survey should take approximately 30 minutes to complete. In some cases, fieldworkers will be requested to return to the property where data appears to be unreliable, or where data is missing.

General Advice

Where possible, fieldworkers should move from area to area in small groups, each managed by a fieldwork supervisor on the ground – this ensures safety in numbers as well as efficiency in covering the study area

There may be instances where residents refuse to participate in the study or are not available during the time fieldwork is conducted:

• Fieldworkers must make at least three attempts to revisit absent households during the course of the study at different times of the day and on different days of the week.

• If after three attempts the household could still not be reached, the fieldworker should leave a letter for the household asking them to come to the project office or to call the enumerator to schedule an appointment

• Surveys that are incomplete or that appear to be unreliable (where inconsistent or unrealistic responses are provided) will not be accepted and fieldworkers will be required to return to respondents to complete and verify data

• Where images are of poor quality or data captured by the enumerator does not match up with data on the images, these surveys must be returned to the enumerator for correction

• Data monitors must work closely with the fieldwork manager and fieldwork supervisors to ensure that fieldworkers perform their jobs meticulously

• It is not realistic to expect a 100% survey coverage and completeness rate, a recommended target rate is 85% or above

Data Collection Checks

In order to ensure that the data is of high quality it must be checked by data monitors. Some common things to check include:

• GPS co-ordinates Do GPS co-ordinates make sense? (i.e. was the survey conducted where it should have been conducted)

• Duration of survey Does the duration of the survey make sense? (i.e. not too short or too long)

• Spelling of names and surnames Check against the submitted ID documents, marriage documents etc. (if available)

• Correct identity numbers Check against the submitted ID documents, marriage documents etc. (if available)

• Where there are no photographs to review, check for the validity of the ID number using a formula

• Beneficiary details Check that the beneficiary profiles make sense in the context of the survey (relationship to the current occupant)

• Quality and correctness of photo uploads

  • Check if the photo is clear and the whole document is in frame

  • No corners or sections of the document are cut of

  • Photo was taken directly above the document, not from an angle

  • Photo is not blurry or dark (i.e. document details should be easy to see)

  • Check if the photographed document is accurate (i.e. if a marriage certificate is selected, the photo is of a marriage certificate)

  • Check that all pages of the document are captured (i.e. not just the front page)

  • Check if the details in the document match the details submitted in the survey (i.e. name, property details – erf number, etc.)

  • Check for missing photos (i.e. fieldworker ticked that a document was available, but no photo was taken)

  • Photo of the property is of the whole dwelling (i.e. not just the front door) and is clear (i.e. not blurry, good lighting)

• Where sale agreements are signed, these need to be carefully checked:

  • Correct ERF number

  • Correct names (spelling) and ID numbers of purchasers

  • Each page initialed

  • Signature. Where the owner is married, two signatures on the sale agreement

  • Two witness signatures