Blog

“The scientist is not a person who gives the right answers, he’s one who asks the right questions.” – Claude Lévi-Strauss

You are probably familiar with the computer science concept GIGO. GIGO stands for Garbage In Garbage Out, and it simply means that bad input produces bad output.

In order to make sound, data-driven decisions, it is important to make sure that the data you analyze is itself in good order. Any conclusions drawn from poor quality data are, by definition, poor quality conclusions.

Seeing as questionnaires are a common method of data collection, here are 5 tips to help you develop more effective questionnaires and therefore collect better quality data.

1. Be Clear

Make sure that every question is clear, and leaves no room for ambiguity. For example, if asking about how long a respondent has been pregnant, specify whether the answer should be given in weeks or months. Also, avoid offering ambiguous choices to a question. For instance: How often do you eat meat? Never, Often, Sometimes, Always. Respondents do not necessarily have the same perception of what these words mean. So it is better to provide clearer options such as; Less than twice per week, 2 to 3 times per week, 4 to 6 times per week, More than 6 times per week.

2. Account for Missing Responses

Decide before hand how you are going to handle missing data. This is especially important when creating an electronic questionnaire, which may require a response to each question before proceeding to the next one. Usually, it is better to not allow “blank” responses, as sloppy enumerators may inadvertently skip an important question. If there is a likelihood that a question may not be answered, consider providing a “No response” option. Many questions of a personal nature, such as one’s salary range, should include a “No response” or “Refused” option.

3. Be Logical

Your questionnaire should follow a logical sequence by grouping related information together. When interviewing respondents, it is typical to begin by asking for basic information, such as age and sex. You may then move on to individual thematic areas such as education, occupation, housing and so on. It is generally a good idea to save any sensitive questions for last, when the respondent has relaxed and settled into the interview.

4. Do not Tax the Respondent’s Memory

Asking about events from 1 or 2 years ago is not likely to yield good data. People typically forget details of events from such a long time ago. Instead of asking “How many times did you go to the gym last year?”, consider asking: “How many times did you go to the gym last month”. Alternatively, you may ask: “On average, how many times do you go to the gym per month”. Framing your questions this way increases the chances of getting accurate responses.

5.Carefully Implement Skip Patterns

If your questionnaire has questions that depend on other questions, make sure you describe the flow or skip patterns correctly. Again, this is especially important when developing digital questionnaires, as whatever patterns you describe will be strictly enforced. This is easier to do if you follow Tip 3. Relationships between questions are a lot harder to miss when similar questions are grouped together.

I hope you will find these tips useful when developing your next questionnaire.

Have a great week!