In a typical survey, the success of the survey, is not really dependent on the responses to the survey, but largely depends on the entire survey design. The design of the survey needs to be given proper attention, so that the analysis and interpretation is correct. One of the challenges is to collect the relevant information in a precise manner.

Any typical survey design will involve

What to ask ?              ->    Information needs
How to ask?                ->    Questionnaire design
Who to ask?                ->    Target group definitions
How many to ask?     ->    Sample size
How to find them?     ->     Sampling design

Out of the 5 steps stated above, usually Questionnaire design is felt to be the easiest and hence given low importance. Most of us feel that we are good in conversing and hence can design survey questionnaires very easily. The point is, a survey questionnaire is not like a usual conversation between two people. The person designing the questionnaire has designed having the entire perspective, whereas the person answering will interpret the questions in his/ her own way and reply. The danger is, the data used in the analysis is devoid of the respondent’s interpretation. This aspect is becoming more and more critical, with the advent of various modes of survey like internet surveys, email surveys and so on which are self filled questionnaires.

One of the best practices is after designing a survey, take on the role of a respondent and answer the following questions

• Do you understand the survey?
• Will you be able to answer the questions?

• Example Q. “On an average, what percentage of your leisure time at weekends would you say you spend internet surfing?”

  
  

• Need to rephrase it with proper options

• Q.  “On a typical weekend, roughly what proportion of your leisure time do you think you spend internet surfing?” (Choose from the following options)
• All, Almost all, Almost three quarters, About half, About a quarter, Less than a quarter, None

• Will you be willing to answer the questions?

• Simple questions like “What is your monthly income”?
  
  • Need to rephrase it with proper options
    

• Q. In which of the following income band, does your monthly income fall?
• <Rs.20000, Rs.20000-Rs.40000, Rs.40000-Rs.80000, Rs.80000+

At all times check that

• The language is simple

• Avoid using unfamiliar and ambiguous words
  
  • Example – Use of words like Convenience foods, Horsepower

• Consistency in asking

• Decide on the sequence of asking, for example, demographics at the beginning,  followed by product usage, followed by lifestyle

• It is precise and to the point

• Length of the questionnaire will impact the quality of the responses
• As the length of the questionnaire increases, respondent may get bored and  answer the survey mechanically

• It is not boring

• Try adding pictures, interesting patterns
  
  • Instead of a simple rating scale 5- Very Good; 4-Good; 3-Average; 2-Poor; 1-Very Poor, use smileys

Always pretest the survey with a small sample & be sure what needs to be measured is in fact what is measured.

Revisiting the 5 crucial steps of a survey design stated above, if the information needs are crystal clear, it will result in a proper definition of the target group and ease the task of questionnaire design. Further a sample design and sample size decided with statistical support will ensure a meaningful survey.