2
Samples, Good and Bad
21
CASE STUDY As discussed in Chapter 1, in February 2014 the Michigan online news site MLive ran the story, “Take our online poll: Should Michigan legalize marijuana.” Of 9684 respondents, 7906 (81.64%) said Yes, 1190 (12.29%) said No, and 588 (6.07%) said Decriminalize but not legalize. These results would seem to indicate overwhelming support for legalizing marijuana in Michigan. However, the Pew Research Center conducted a poll on March 25–29, 2015, in which they asked, “Do you think the use of marijuana should be made legal or not?” The Pew Research Center reported that the poll consisted of telephone interviews with 1500 randomly selected adult Americans and that 53% of those surveyed favor the legal use of marijuana. This is a majority of those surveyed but not the overwhelming majority that MLive found. There is a large discrepancy in the findings of these two polls. This may be because the polls were conducted at different times, the populations sampled were different (Michigan versus all adult Americans), the MLive poll had a much larger sample than the Pew poll, the questions asked were not identical, or perhaps the data from one or both polls are simply bad. By the end of this chapter, you will have learned how to assess this last issue, namely, whether the data from these polls are good or bad.