25 Tips For A Perfect Online Survey

On the Internet, often we are all poorly designed questionnaires. Sometimes there are too many irrelevant questions. These problems make that data is difficult to analyze, or even the wrong conclusions are drawn. Three components apply to every online survey.

1. General survey design

2. Design questions

3. Assessment Scales

General Survey Design

1. Start with the end in mind. Think about what you want to do with the answers and how to analyze it.

2. Provide an optimal routing. Do not ask questions which they have to answer not just lead over them why. Most survey tools offer this functionality. By simply skipping the question to advanced routings.

3. Keep demographic questions until the end. Respondents have a certain goodwill about the survey and the fact you want to hold as long as possible. So first set the interesting question and only at the end the simple questions such as age and gender.

4. Space for comments. For some contributors, it is important to place comments that are relevant to the investigation. So make the space for it at the end of the survey.

5. Test the online survey. Before the survey goes live, it is advisable to withdraw such by some people. So get you out the last errors.

Drafting Questions

6. Questions should be neutral to avoid getting respondents 'forces' to a particular answer. For example, "What is your opinion about ..?" Or "Can you indicate to what extent you are satisfied or dissatisfied ..?".

7. Simple questions. Make the questions as simple as possible so that everyone understands this. Ask the question in spoken language is usually the best and avoid double negatives. Twice 'not' is a question difficult for many to understand.

8. Avoid too many open questions. Closed questions are easier to analyze than open questions. Use any open questions if you have little time to explain it properly. Remember that the majority of respondents prefer not too much text in tap unless they have a strong opinion on a particular subject.

9. First general and then specific questions. For example, first set a general question about your website (such as overall satisfaction). These questions first before you ask specific questions regarding the design or content of your site. The reason is that if the respondent has a strong opinion on the latest specific issue, it affects the answer to the general question, and thus a sort of 'contamination' occurs because it is still in the positive or negative bias.

10. Use radio buttons and check boxes correctly. When only one option to choose, use radio buttons. Use check boxes if one may give multiple answers.

11. Be careful with follow-up questions with "why". Sometimes you want to know why a respondent gives a specific answer. This often provides valuable information. However, it can also generate plausible answers.

Rating Scales

12. Use scales with an equal number of positive and negative options. 4-point scales have two positive and two negative answer options like "very satisfied", "satisfied", "dissatisfied" and "very dissatisfied". 5-point scales also have two positive and two negative answer options sandwiching a reply option "neutral."

13. Use scales that end with the most positive and negative option. The scales used thus ending of "very satisfied" to "very dissatisfied". Not of "very satisfied" to "dissatisfied."

14. Use scales to measure one aspect at a time. A five-point scale for example "very satisfied" to "very dissatisfied" with between not satisfied / not dissatisfied. So don’t "average" or "satisfactory" as its centerpiece. These have their scale as "above average".

15. Take into account respondents avoids choosing extremes. Respondents tend to prefer not extreme. This tendency increases when strong extremes are used such as "very" or "always". Some researchers prefer to use an even number of possible answers which respondents are forced to choose between two extremes ("very good" vs. "very poor").

16. Think ahead. If you design the questionnaire, and you want the results of two or more questions to compare with each other, keep in mind that you use the same response scales. So then compare scales of "very satisfied / very dissatisfied" with each other and not with "satisfied / dissatisfied."

17. Points on the scale must be different. Avoid having two possible answers which are close together and too little different for the respondent.

18. Use of "not applicable" if needed. Otherwise, there is a high probability that other possible answers are clicked wrongly.

19. Be thrifty with "do not know" option. In some cases, respondents did not know to answer and can use the "do not know" option. Some respondents use this option to avoid having to think about the question. The solution is to replace this option with words that seem stronger allowing the respondent this option will let go sooner. An example is "no idea at all." and "do not know" option does not use it in any case in the middle of a scale but only at the end.

20. Use exclusive response categories for age. For example, "16-24" should be followed by "25-30" and not by "24-30."

Increase The Response To Your Online Survey

21. Write a promotional subject line. This is what recipients can see the first invitation.

22. Write a short email. This email should briefly indicate where the research is about and what the respondent has it. Feedback portion of the research motivates people to get involved. Also, the provision of an attractive incentive helps to increase the response. Tell how long the investigation will take and hold you.

23. Does the study as short as possible. Look closely at what questions you want to ask and it can do something with it. Also, if there is no overlap present in the questions. Nothing is as frustrating as respondents to think "which was just now required." An investigation should take no longer than 10 minutes.

24. Provide an incentive. Providing an incentive increases responsiveness but is often not the main reason to join. Drawing a contest with an interesting gift (car navigation, iPod) has a positive effect on the response. Also providing more vouchers so the chance that you also win increases, is also welcomed.

25. Send a reminder when it appears that the response remains. Normally, the target sample size must be received within 3 days.