This is the other topic a friend wanted me to write about....
do you remember the tv show ..... that had the host saying
"Survey shows ...."
do you know why the producers of that show think you should be able to guess at the answers
yes, certainly. it is because the survey is supposed to represent "your" population
This implies that the sample covered by the survey is the "same" as that of the population, and stat-people, as opposed to cat-people, hypothesize by saying that the population follows the normal bell curve and taking a random sample will have the same characteristics, and hedge their bets by specifying a degree of error, expressed in standard deviation, such a plus-or-minus 3% margin of error.
Did you know that when one designs a survey, they have a hypothesis?
what is the hypothesis, you say. you mean you dont know, and there you are, making a survey
Surveys are supposed to be impartial unbiased samples of a population. Otherwise, if the survey is biased, then no significant conclusions can be made. So the survey design is important because this allows conclusions based on the sample be applicable to the larger group.
No comments:
Post a Comment