Information Services

Using online surveys

You may find these general points helpful when using online surveys.

Bias

If you are surveying a fixed group of people (e.g. students on your course) then it's possible to direct all of them to a web page or Learn survey to complete a form. However for surveys intended for a wider audience, or aimed at a balanced sample, you should consider how using an online survey may affect the response group by excluding certain groups.

Selecting tools

By their nature, online surveys exclude people who do not use computers. The type of survey tool selected may also affect who can complete the form. Some types of survey - and indeed some types of question - cannot be completed by people using certain types of assistive technologies. The BOS tool has been found to be more accessible than software such as SurveyMonkey which uses Javascript, unreadable to many assistive technologies. There may be technical issues with certain platforms and browsers, as few tools can support every possible combination. This may create a significant bias for certain user groups.

If you wish your survey to be used by a representative sample of the general public, you should consider making your survey available in alternative formats (e.g. paper-based copies) as well as online.

Design

There is some evidence that online surveys should be structured slightly differently from paper-based surveys to help ensure best possible completion rates and reduce the erroneous completion of filter sections. However you should also check the limitations of the tools - not all will allow users to go back as well as forward, and some do not allow return visits (i.e. completing the survey in more than one session).

Before you start

Most people have more difficulty reading on screen than on paper. Keep text to a minimum and in short blocks. Include a welcome note, any general instructions for survey completion, a data protection policy statement if appropriate, contact details, any disclaimer necessary, and a thank-you note at the end. Consider what question types are available to you online, and match each question to a question type. Instructions for question completion should always be as simple and concise as possible. Use plain English, avoid technical terms, and keep sentences short and simply structured. Check the "logic flow" of the questions, so that filter questions are appropriately associated. In general it is best to keep question structure as simple as possible, and use follow-on questions rather than merge questions of different types.

Design

Wherever possible, online surveys should:

  • carry information on the first page about the estimated time it will take to complete;
  • have questions grouped together into a series of fairly short web pages, rather than one long page;
  • use filter questions where possible, which lead the user to the correct question or section as a follow-on;
  • state clearly on each page where the user is (e.g. "Page 3 of 5").

Before release

Go through the survey carefully to ensure all questions function as they should and collect data correctly. Check the survey on different computers (especially Windows and Mac operating systems) and a variety of browsers to ensure it can function in the most common ones.

Get several users to try out your survey and give you feedback. You should ask them to pay attention to ease of use and comprehension. Check that all images have a helpful "alt" tag.