1.	What is the purpose of the survey?	Surveys are conducted for many reasons. They may be required to gauge independent thought but just as easily could be used to try and gain support for a particular opinion. By phrasing the questions and structuring the answers surveys can be used in a multitude of ways and for a variety of reasons. When compiling a survey don't lose sight of its purpose.
2.	Title the survey	The importance of a survey title is often overlooked and can lead to a wasted opportunity to instantly summarise the survey's objective and grab the attention of invited respondents. Respondents are going to invest time in completing the survey and they need to feel the investment is in some way worthwhile.
3.	Do not make the survey any longer than it needs to be	Every question that is asked should be asked for a reason. There is no benefit asking a respondent's age and gender if that information is going to be irrelevant when the surveys data is analysed. Conversely surveys should not be too short; if a question is asked that relates to gender it is important to be able to analyse the responses as a whole and then to be able to further analyse them by gender of the respondents. Consider the overall time that will be required to complete the survey and appreciate that the longer the survey the greater degree of full and partial non-responses. Focus on 'need to know' questions and minimise 'nice to know' information.
4.	Use plain English, avoid jargon and acronyms, maintain consistency and don't ask questions that may result in ambiguous answers	Care must be taken in wording a question. If a question is not precise there is every chance that respondents may interpret the question differently making any analysis of the data meaningless or at the very least misleading. Be consistent with the possible answers, avoid mixing True/False and Agree/Disagree and never mix Agree/False!
5.	Avoid long questions	Try to use short sentences wherever possible. Long questions tend to cause respondents discomfort and can lead to a higher level of incidents where respondents abandon a survey.
6.	Ask one question at a time	Avoid confusing the respondent with a question like 'Do you like football and tennis?'
7.	Avoid influencing the answer	Even in a survey that is conducted with the intention to gain support for a particular view if the results of the survey are going to be useful it is important not to load the question. The results of a question phrased 'Do you support the proposal to stop irresponsible people from driving too fast outside schools' is unlikely to have any value.
8.	Ensure that the answer format used allows the respondent to answer the question being asked	If the survey doesn't allow the respondent to answer how they really feel the results will not truly reflect the respondents views and they may be less inclined to complete the survey. As a last resort consider the benefit of including a 'Don't know', 'Can't say' or similar response option.
9.	At the same time that you compile the survey consider, when the survey is complete, how the compiled data is going be analysed	If a question is asked that allows a free text or open ended response appreciate that such information is likely to be difficult to score and/or summarised. In some case free text responses may be necessary but before using them give serious consideration to see if they can be avoided. For example if the respondent is asked how long they have worked for a particular company, and they are given a free text format in which to answer, then each respondent is likely to answer differently some in days, some in weeks others in months and/or years. It would be better to consider the possible lengths of times and group them into several groups, for example 'less than 1 year', 'between 1 and 3 years' and 'more than 3'.
10.	Ensure that the questionnaire flows	When asking questions group the questions into clear categories as this makes the task of completing the survey easier for the participants.
11.	Target your respondents	In a survey that for example is considering student attitudes it may be obvious that the target audience needs to be aimed at students and it follows school and universities are where students congregate. However sometimes the target audience for a survey may be more difficult to determine. For example in a survey relating to gardening it may be difficult to target those people who have a specific interest in gardening. Consider if in a survey about gardening would the views of someone who has no interest in gardening be relevant and if they were included would their responses compromise the survey's findings?
12.	Allow the respondent to expand or make comments	Allowing the respondent to make additional comments will increase each of the respondent's level of satisfaction and will also give valuable feedback on the specific questions and/or the survey as a whole. Remember though for a large sample collection it may be difficult to analyse free text comments.
13.	If the survey you are conducting is to be confidential ensure that your pledge is upheld	If you have assured the respondents that the survey is confidential ensure that the individual data is not to be shared with anyone and the information is not going to be used for any other purpose. Confidentiality must be maintained at all times and any identifying information destroyed after the survey is complete.
14.	Weigh up the benefits of allowing respondents to be anonymous or identifiable	If your respondents are to be anonymous then appreciate that you will be unable to follow up or match 'pre' or 'post' surveys. However in some cases allowing people to remain anonymous will allow people to respond without possible peer pressure.
15.	Give careful consideration to the best response format	Where possible it is good practice to maintain a consistency in the format used for responses. Keep in mind that when analysing the data radio buttons are easier to analyse than tick boxes that offer the respondent multiple responses. Do not use a tick box if a radio response would do, on the other hand do not use a radio button in place of a tick box if it is going to be detrimental to gathering useful information.
16.	Consider the benefit of using a progress indicator to show the respondents how far through the survey they are	Respondent drop out can occur if the survey appears to be a stream of never ending questions. Consider the benefit of indicating at the top of each page how far the respondent is through the survey. The longer the survey the more important this consideration will be as if a respondent can see the end is in sight then they are more likely to complete the survey. It is good practice to give an indication as to how long the survey is likely to take so the respondents can choose the best time to complete the survey.
17.	Inform the respondents of the survey end date	In most cases it is preferable to strike while the iron is hot and encourage respondents to complete surveys as soon as possible but in some cases this will not be practical. It is therefore important to advise respondents as to the survey's cut-off date so that they have the opportunity to schedule the necessary time.
18.	Pilot the survey	Before publishing a live survey publish a small pilot survey to check for questions that are ambiguous or confusing and to ensure that the survey is aesthetically pleasing. Repeat this process until you are entirely satisfied that the survey will encourage people to fully complete the survey.
19.	Before publishing the survey proof read the survey several times	Check and check again that the survey is grammatically correct and makes sense. If possible get someone else to proof read the survey before you publish, if no one else is available then take a break before checking again.
20.	Remember to say thank you!	To complete a survey respondents need to invest their time and should be thanked either in a covering letter, at the end of completing the survey or in a follow up letter. You may even want to consider incentives such as a prize draw or reward