Getting started with survey design

Survey design 


Very useful for certain types of questions. They are useful for things that you can get information for.

We use surveys to measure attributes and attitudes/opinions about something. Also experiences, knowledge and behaviour.

Participants may not be completely honest due to a variety of reasons like shame or fear of consequences 

Think about how to administer the survey, maybe online you get more honest answers than over the phone

Especially in linked scales a 6 might not always be a 6 since in an hour the subject might give a different number.

Myers Briggs type indicator

Test retest and liability. This is a way to avoid different answers to the same questions.

Designing a survey

Define the questions to be answered 

Think about what it is that you want to find out.

Questionnaire must be concise and clear.

Consider does a suitable questionnaire already exists?

Think of a questionnaire like a conversation. The questions are ordered to build up confidence and rapport.

Start with broad questions e.g. How satisfied you are with your current health?

Then you ask more specific questions.

For a live questionnaire you ask for past behaviour then for current behaviour then for attitudes.

The final questions should be broad and general, maybe demographics. Demographics at the end because you might get biased answers based on gender for example.

Sometimes responses to one question may bias responses to another question. You can avoid that by counterbalance the order of questions or decide if it is more important to you that responses to one of the questions remained unbiased and place them first.

If it looks organised people might do it more.

Consider what will you do with the responses.

Make sure open questions ask only one thing.

In closed questions the answers should be mutually exclusive.

You have to pre test your questionnaire 

Amazons mechanical Turk you can use that to get a lot of responses 

Comments are closed.