Phone Surveys

Technology from Database Systems Corp. lets you develop IVR survey applications using our interactive voice response IVR solutions. Surveys can be initiated by outbound phone calls or can be a response to callers. Using our PACER and WIZARD phone systems with the Smart Message Dialer and survey software, we can call your survey prospects and play a highly focused and custom greeting. We then can give your survey audience the option to take your survey or even talk with a representative, leave a voice message, hear additional information, or simply decline to participate in the survey. The survey can accept touchphone response or can record each question response for later analysis. To view more information regarding our automated phone applications, please visit our Automatic Phone Survey solution web page.

The following is an article relating to call survey techniques and products and services in our business.

Minimizing Bias

Remember, you need a random selection procedure to make a smaller sample an accurate representation of a larger population. The Madison example (lowering the drinking age) shows how self-selection can create nonrandom conditions that undermine the usefulness of the sample. This is just one instance of a more general phenomenon. A nonresponse bias can be introduced when part of the original random sample fails to respond.

If only 25% of sample members respond, then 75% of the population is not represented in the results. If the individuals in the sample are identical, this isn't a problem. But if they have any salient characteristics that cause or even correlate with nonresponse, the survey results can become inaccurate to a significant degree.

Unfortunately, research shows that nonrespondents often differ from respondents in significant ways, although the extent varies with the population of interest and the topic being surveyed. Some of the problems can be solved by adjusting the survey results, but it's much better to minimize bias by getting a highly representative rate of response from a properly selected sample. How high a response rate do you need? Unfortunately, that's hard to tell, because nonresponse bias is not as quantifiable as the margin of error. Certainly, the researcher must receive a high enough response to lower the error rate attributable to nonresponse bias to levels comparable with sampling error and other sources of error.

Sometimes, statistical weighting can mitigate biases somewhat. Say that the sample results differ from reliable data (e.g., circulation distribution by state) already known about the population as a whole. The results can be weighted in tabulation to enhance representation of the lowerresponding segments and reduce representation from the higher-responding segments, so that the overall totals are in proportion. This technique must be used if a stratified sampling plan is used, so oversampled groups are restored to proportionality in the final dataset. However, this approach doesn't solve all problems of fair representation. We can't conclude in advance that nonrespondents from New York City are identical to nonrespondents from Nome, Alaska with regard to the variables of interest to the survey.

Scientific Sampling Summary

This material introduces you to the paradigm of scientific sampling. These techniques are best suited for quantitative mail surveys that use known and well-enumerated populations. Magazine circulation lists are a prime example.

If you have other objectives or must operate subject to constraints on time, money, and prior information, you may need to use other approaches. Other forms of probability samples, or some nonprobability forms, may work better for you. But the best way to get bias-free, statistically precise survey results is to get a high response rate from a randomly selected probability sample of sufficient size.

Designing the Questionnaire

It's bad enough that a poorly-designed questionnaire often lowers the response rate: but whatever results are received are often unusable. A successful survey starts with a well-designed questionnaire that asks the right questions, in the right way.

Your survey should probably answer at least one of these questions:

• What do we know/want to know about our customers?
• How do our customers feel about our company or our product and its future?
• How effective has our performance been in the marketplace?
• What needs are not being met? What can be done to satisfy them?