1. You have been provided with a questionnaire used in the 2016-2017 Nigeria vaccination coverage survey. If you were to use this questionnaire as a model for one that you are developing for your own survey, what factors would you consider when reviewing and customizing the questionnaire to ensure that it is aligned to objectives relevant in your own country context? (If your country context is Nigeria, consider objectives that could be considered for future surveys.)

The first consideration is whether each question will be either useful and necessary for analysis or whether it is necessary for identification purposes. Particularly for individual information on HH members it must be asked if the need for information collected justifies the time it will take. Large HHs with numerous individuals can be extremely complicated and time consuming to collect all the information in the example. Date of birth can be problematic especially for older individuals. Furthermore, is the information on whether natural parents are alive and where they live really required to address the objectives of the survey?

2. From the Nigeria questionnaire provided, develop a sample code book (data dictionary) ONLY for section 1: household information of the questionnaire. (You can attach an Excel or Word file in your Update and/or insert a screenshot of your file.)

data dictionary

3. Why is it important to have a unique identifier for each individual in a survey? Summarize key considerations for the design of a numbering system to ensure this.

The unique identifier is the mechanism that links the individual to the questionnaire information collected. This is typically through the use of an ID that may use numbers or text with each being attached to a respondent. There should be no possibility of using the same ID for more than 1 respondent and therefore, the system of allocating IDs must be clearly understood by all surveyors and their supervisors. Some identifiers may indicate the particular state, region or district follow by a series number. The allocation of identifiers would need to be taught during the training so that is understood by all data collection staff.

4. If paper data collection and pictures are to be taken, how will you ensure that the pictures can be linked to the correct child?

If pictures are to be used, the same identifier would need to be applied to a respondent or HH so that the child can be correctly linked to his individual data. A number system to allocate identifiers for picture would also need to be part of the training to ensure that the same identifiers are used for the child. With some tablet questionnaire programs it may be possible to take a photo with the HH questionnaire so that the child is linked to the HH record and his/her responses to questions.

5. Data collection is increasingly being conducted using CAPI (computer assisted personal interviewing) with tools like CSPro, ODK, etc. What features would you recommend are added to the data collection software to improve overall data quality? What considerations for support and in case of malfunctions would you put in place (contingency planning, support and troubleshooting, etc.)?

CAPI tools should have menus for choosing answers to questions and requirements that questions be filled before the questionnaire can be closed and saved. Skip patterns can also be built in so that questions that are not applicable can be skipped. These tools can help to reduce data entry errors and systematize so that multiple variations of spelling do not complicated data cleaning and analysis. Teams should also have hard copies of the questionnaire as back up in case the CAPI tool malfunctions, is lost, stolen, etc. Data should also be transmitted to a secure server daily or as soon as possible for safekeeping and to facilitate preliminary analysis.