The issue: Herbert Asher talks about polling and sampling, the different ways to do both and the various ways people can influence and/or create bias in responses to polls and samples.
Major strengths: Asher spends plenty of time defining and describing polling, sampling and the different kinds of both. He particularly does a good job in the sampling chapter, going into detail about nonprobability sampling, sampling designs, simple random and systematic sampling, stratified sampling, cluster and multistage sampling and the various types of telephone sampling. His examples of how the order of questions an the use of multiple items and indexes on questionnaires were also on point.
Major weaknesses: Perhaps it is just me, but I have a hard time believing that spacing between answer choices on a questionnaire could influence a certain answer. I understand order, wording and even vertical vs. horizontal answer choices, but not shaping. I also feel that the telephone sampling section needs to be updated. More and more households are using mobile phones in place of land lines and will most likely not answer a call from a number they do not recognize. It feels weird to type this, but it seems like telephone sampling could be an antiquated form of polling and sampling.
Underlying assumption: Asher seems to assume that sampling and polling has not changed much in recent times, focusing a lot on the various types of phone sampling techniques and not so much on Internet sampling techniques. The fact that this book was copyrighted in 2012 worries me and shows that (in my opinion) he is lacking in Internet polling and sampling techniques and information. You can also add this to the “major weaknesses” section.
Provocative questions: Asher mentions how the 2000 census says that 97.6 percent of American households have telephone service and that those without are “more likely to be uneducated, poor, in a minority group, of low occupational status and living in a single-adult household.” What would those stats look like today? It seems to be that fewer houses today have home telephones, so I am interested to see what the number would like now (unless it already included cell phones).
Additionally, what percentage of people/households today have unlisted phone numbers and how are pollsters combating this issue?
Have any polling companies considered polling via text message/social media?