In my initial blog on this topic, I touched on four areas. Specifically:
1. The differences that exist when writing a questionnaire that
respondents will fill out themselves as opposed to one in which a professional interviewer administers the questionnaire to the respondent.
2. Knowing which questions should be asked early on in the questionnaire, in the middle or toward the end.
3. Understanding how to phrase questions.
4. Being sensitive to questionnaire length.
(http://www.customerthink.com/blog/guidelines_writing_effective_questionnaire)
But the perfect questionnaire will achieve nothing if you fail to get cooperation from a representative sample of respondents. To secure respondent cooperation for an interview, the following steps will be helpful:
Immediately address skepticism:
Unless they are lonely souls just waiting for the phone to ring or for a questionnaire to appear in their mailbox or on their computer screen most people are skeptical when asked to complete a research survey. The knee-jerk reaction is that someone is trying to sell them something.
In fact, one in three people who are contacted on the phone will immediately hang up the second they realize they are being asked to complete an interview. The percent that dismiss a mail or internet questionnaire is even higher. This means that at the outset fewer people are even available to be interviewed.
As a result, if you don’t gain cooperation from people who will give you their time, your ending sample of respondents could be nothing but strange people with strange attitudes. To gain cooperation from a representative sample of respondents, follow these guidelines:
A. At the outset, assure potential respondents that no attempt will be made to sell anything as a result of their cooperation.
B. Inform respondents how long it should take to complete the interview (e.g., “It should take you about 20 minutes to complete this telephone interview.”). Being honest up front is particularly important for longer telephone questionnaires (more than 20 minutes), as respondents are more likely to terminate an interview if they feel they are being led on interminably.
C. If you don’t feel it will bias the interview, tell respondents upfront who is sponsoring the study. While this is not always appropriate, cooperation rates will significantly increase if respondents know that a legitimate company is seeking their opinion.
Offer an Incentive:
Potential respondents will be far more likely to give you their time if you offer them an incentive. While this isn’t usually necessary for shorter telephone interviews (under 15-20 minutes) it can be essential for longer ones. And it is almost always necessary to offer an incentive when conducting a mail or Internet study.
Here are a few ideas for incentive offers:
A. Tell respondents that if they complete the interview their name will be included in a drawing for a substantial prize (TV, computer, trip to Cancun, etc).
B. In addition to the prize drawing, give respondent a list of charities to choose from and ensure a donation of X number of dollars (say $100 up to $500) will be made to that charity if a minimum number of interviews are completed. Of course, you should follow through with your promise.
C. Offer to mail money for completion. Usually $5 is adequate for phone interviews of 20 to 30 minutes and $10 for 30 minute plus interviews. For mail questionnaires, send $1 when you mail the questionnaire with the promise of an additional $5 or $10 back when the questionnaire is returned. The dollar amount back should be higher the longer the questionnaire.
D. For Internet studies, offer to send the incentive upon return of the questionnaire. Again $5 and $10 incentives are appropriate given the time it takes to fill out the questionnaire.
The types of incentives that you could offer are limited only by your imagination and budget. Often, though, a pre-test is prudent to determine if you are getting adequate co-operation given the incentive that you are offering.
Remember that many companies are content to blast an email questionnaire to 10,000 customers and analyze data from the 5% or 500 that responded…and to assume 500 is an adequate sample size. In such cases, it may be better to forgo the research altogether than to assume the attitudes from 5% of your customers are representative of the other 95%.
To ensure a representative sample, your response rate for mail and Internet questionnaires should be a minimum of 35% with goal of 50%. That is, for every 100 questionnaires you send to your target, you should seek to achieve 35 or more returns.
To reiterate, more important than getting a large sample is getting one that is representative of your target. That only way to know if the opinions you are getting hold true for everyone—not just those who responded and might have a particularly positive or negative point-of- view.
And, achieving a representative random sample means doing what is necessary to gain the participation of respondents who would not otherwise give you their opinions. By not doing so, you risk making decisions on the basis of flawed data. Better to spend your money where it will give you a reliable return!