Planning a Survey
Surveys may only be part of the background research and not the whole project. Before starting, the following questions need to be asked – relevant Expo information is added in brackets:
- What are the objectives of the survey (is it to find out opinions as part of background research or is it to obtain scientific data which would be part of the results of the investigation?)
- Are there other sources of data that could be consulted before carrying out a survey (literature search)?
- How can it be ensured that those who have a stake in the outcome of the survey support it (well-written permission letter)?
- How can a list of people/organisations to be surveyed be developed and how reliable will the contact information be (important part of research plan)?
- How can a sample (group of people taking the survey) be designed to minimize cost and maximize the accuracy and flexibility of the results?
What to know when designing a questionnaire:
- What information is required to meet the needs of the project (part of aim and hypothesis)?
What is the best way to word questions so as to get an unbiased response (procedures/method)?
- How to design the survey questionnaire to ensure the questions are clearly understood and answered properly (procedures/method)?
What is the most reliable and cost-effective method of delivering the survey (procedures/method)?
- How to pre-test the survey questionnaire (in the pilot study)?
When to use the results of the pre-test?
Answer these questions when dealing with the respondents:
- How will the confidentiality of the responses be protected (all questionnaires are filled in anonymously)?
Who will respondents contact when they have questions (the investigator)?
- How long do respondents have to respond (shorter deadlines work better)?
What to do if the respondents don’t respond (accept this as it’s the respondent’s choice to complete the questionnaire)?
- What is an acceptable response rate? What to do if the response rate is unacceptably low (for an initial study for Expo a minimum of 100 completed survey questionnaires for primary schools and 200 for high schools is needed. Should there are too few then send out more survey questionnaires)?
What to know before analysing and presenting the information:
- Whether or not the responses are biased and how data will be corrected for bias if it exists (double blind questions)?
The best way to present the data so that the audience can grasp the importance of the findings (tables and graphs of all data)
- The best way to demonstrate that the results are statistically valid, accurate and reliable (all fixed variables must be controlled and there needs to be a sufficiently large sample size for the study to be reliable)
Which techniques will be used to analyse the data to give accurate, fully results (statistical analysis)?
- How to link analysis and the discussion?
Questions sourced from: http://www.bcstats.gov.bc.ca/Home.aspx