|
Information Obtained From and About Surveys
In the September-October 2003 issue of Arthritis Today the results of a poll were printed. Readers were asked to respond to the question, "Where do you find strength?" The following is a summary of the reader responses. 35% indicated from within; 33% responded church/synagogue/prayer; 20% responded family and friends; 7% responded creative outlets, such as writing, painting, etc; and 5% responded volunteering/helping others. Conducting a poll is a method of gathering data that has not been discussed yet. This is a special case of survey research which will be labeled an opinion poll. There are several issues which need to be presented before legitimate conclusions can be drawn from the poll...
- What is the population which the sample represents?
- What is the sample size?
- Are the responses independent?
- What information is provided that is relevant for me?
- Is the question well defined?
The following contains a discussion of potential answers to these five questions...
1. The information presented in the article did not describe any characteristics of the responders. Usually, the potential population is identified and the sample to be polled is selected from this population. In addition to the primary question, demographic data such as gender, age, disease status etc. is obtained from the responders so that the sample can be described.
2. The number of responders was not given in the article. In a planned opinion poll, some criteria is employed to determine an appropriate sample size such as insuring that the length of a confidence interval is at most some number or in the case of trying to determine if certain subgroups of the population differ how large a sample is required so that there is adequate power to detect differences if they exist?
3. Nothing was stated in the article as to whether the responses were independent. It is possible that individuals participating in a support group could all respond and because they are part of the same support group could all give the same response. These responses would not be considered as independent. Likewise the members of a family where religion is an important component in the family life could also give the same response and these responses would not be considered as independent. In an opinion poll the sampling plan would usually be designed such that only one member from a given family would be included in the sample.
4. If I look at the percentages of the various responses that people indicate work for them, I could possibly see an approach that I could try that might be helpful to me.
5. The question could be more specific in that it could have been stated 'find strength to do what, such as how to cope with my disease.'
It is possible that an individual could be asked to respond to such a question over the phone. If that occurs, it is usually the case that some demographic information is also asked of the respondent. Often a specific population is targeted when these phone opinions are conducted. The person being interviewed should always inquire who is doing the polling and why and how the information will be used before giving answers to questions.
|
|
