Poll surveys are prevalent in determining the pulse of people's opinions especially on critical issues like birth control and presidential choice. Are they always reliable?
No they are not always reliable. Statistics is not about absolute truth. Statistics is about probability, and there is always a chance (though it can be incredibly unlikely). Always accurate also assumes that the object being surveyed is unchanging. Public opinion may have changed by the time you have taken the poll and analyzed the results. You also have to be very careful about the questions asked. Polling 100 people randomly selected from the phone book in Cleveland may not be representative of the national or global population.
That is all assuming that I am not deliberately trying to bias the outcome. I can bias the outcome by carefully selecting my target audience and careful use of statistics. I can also simply make up fake data and statistics to say whatever I want. As long as the table and graph look professional someone will believe the result especially if presented with sufficient conviction.
Look for the 1977 book of William Cochran, Sampling Techniques, Third Edition. The estimation of sample size is on Chapter 4 of the said book.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size