multiple indicator cluster surveys regional training workshop i – survey design general...
TRANSCRIPT
Multiple Indicator Cluster SurveysRegional Training WorkshopI – Survey Design
General Characteristics of MICS3 Questionnaires
Some Features of Questionnaires and Modules• Standardized format as much as possible• Makes training, data processing easier
• All questions = CAPITAL LETTERS• All response categories = lowercase• All instructions to interviewers = lowercase italics.
• Words in (italics & in parentheses) = to be replaced by the appropriate word during interview
• Words in (italic-bold-in parentheses) = to be replaced before questionnaires are printed
Some Features of Questionnaires and Modules
• Question numbers begin with “1” in each module, preceded by two letters indicating the module
• e.g. HL1 or MN4.
Some Features of Questionnaires and Modules
• Any questions or filters with light grey background = will not be asked to the respondent, but completed by the interviewer, based on previous responses or observations.
Recording Responses
• All interviewers to use pens with blue ink to complete all questionnaires.
• Any editing in the field (by supervisors or field editors) to be done by using pens with red ink.
• Office editors to do all their work using pens with green or black ink – By this way, it becomes possible at every stage to
understand the changes that have been made – if necessary to go back
Type of Questions
• Questions which have precoded responses,
• Questions which do not have precoded responses,
• Filters
Questions with Precoded Responses
• For some questions, we can predict the types of responses a respondent will give.
• The responses to these questions are listed in the questionnaire, usually on the right, or below the question
• Never to be read to the respondent unless otherwise indicated (included in the question itself)
Questions with Precoded Responses
• Two types of response categories– Numeric: Only one response category can
be circled– Letter: Multiple responses can be circled
• Circle the number or letter which corresponds to the reply.
• Make sure that each circle surrounds only a single number or letter.
Questions with Precoded Responses
• In some cases, precoded responses will include an "other" category.
• Use “other” when certain that the response does not fit in any precoded category
• Always write the respondent's answer in the space provided (Specify: ______).
• Use margins, last page etc to provide more detail if necessary
Questions with Precoded Responses
• In numeric response codes– 8, 98, 998, 9998 = Don’t know– 6, 96, 996, 9996 = Other
• In alphabetical (multiple response) codes– X = Other– Y = None– Z = Don’t know
Questions with No Precoded Responses
• In entering the response for these questions the respondent's answer must be written in the space provided.
• Usually a number or a date will be recorded in the spaces provided.
• If the response has fewer digits than the number of spaces provided, leading zeroes have to be used.
Skip Instructions
• It is very important not to ask a respondent questions which are not relevant to her situation
• For example, a childless woman should not be asked about the birth weight of her last child.
• In cases where a particular response makes subsequent questions irrelevant, an instruction (usually an arrow) is shown after the particular response directing us to skip to the next appropriate question.
• It is important that skip instructions are carefully followed
Filters
• Filters require interviewer to look back to the answer to previous responses, then mark an X in the box and follow skip instructions
• They are used to prevent a respondent from being asked irrelevant questions.
• Follow the instructions and do not rely on memory• Check back and mark an "X" in the appropriate box in the
filter; then, follow the instructions.
Recording Responses
• It is very important to record answers neatly
• For precoded responses, be sure that the correct code is circled
• For open-ended responses, the reply should be written legibly
Correcting Mistakes
• If there was a mistake in recording the answer or the respondent changes her/his reply, the incorrect response is crossed out and the right answer entered
• Never erase an answer
• Two diagonal lines are put through the incorrect response, and correct response is recorded to the left or right
Adapting / Customizing Questionnaires
Customization
• No single model questionnaire can represent all human experience around the globe
• Customization will have to be made
• However, lesson learned: Adapting too freely may cause production of low quality, incomparable data
Customize,
• But retain international comparability in data you collect
• Only to reflect country-specific categories, phenomena
Adapting / Customizing Questionnaires
• Decide on your indicators
• Identify modules which collect necessary data on numerators and denominators of your indicators
• Make sure you collect information on all indicators appropriately and that a decision to take out one module does not affect collection of data on another indicator
Adapting / Customizing Questionnaires
• Retain all core modules if possible
• Assess whether you are affected by issues in additional modules and decide to include/exclude them
• Optional modules – Assess if it is an issue, if there is interest
Adapting / Customizing Questionnaires
• Assess the length of questionnaires – do not overdo it
• Consider the flow, the rhythm of the interview, do not produce a “form”
• Remember that the inclusion/exclusion of one module might influence another module, or the rhythm of the interview
Some Rules
• Retain question numbers
• Do not assign new numbers to questions already in model questionnaire
• If new questions are included, assign them new numbers – if necessary, use A, B, C etc, such as HA15A, HA15B to be inserted between HA15 and HA16
• Retain Latin characters
Some Rules
• Understand what response categories are attempting to capture
• Country-specific response categories may be already there, under a different name
• A new category should not take up the code (number or letter) of those in the model questionnaires
Some Rules
• Be cautious about changing the ordering of modules
• If not absolutely necessary, do not change ordering of modules or questions
• Do not change the definition of the eligible group, e.g. age group
• Change can only be to capture a wider group, e.g. asking questions about age group 2-14 rather than 2-9
Suggestion
• Share your questionnaires and plans with UNICEF Regional Office, UNICEF NY
• Process to do this will be discussed during the workshop