Reading Notes: Sharon Lohr's Sampling Book.

A Sample Controversy

"84% of women are not satisfied emotionally with their relationships." Reported in the book Women and Love: A Cultural Revolution in Progress (1987).
It probably sounds like that's a fact which applies to all women, whether they participated in the survey or not. The book widely raised millions American eyebrows and was met with a wave of criticism.

People bashed the book from various angles, including that of statistical view. Essentially, the following characteristics of the survey make it unsuitable for generalizing the results to all women:

  • Since the recipients of questionnaires decides if they would be included in the sample or not, the sample was self-selected. (Hite mailed 100,000 questionnaires, of which 4.5% were returned.)
  • The recipients were inproportionally women that have joined an "all-women" group, and their viewpoints may be hard to qualify as voices that speak for all women.
  • The survey has 127 essay questions, even with several parts under each question. Who will tend to return such a survey?
  • The wording of many of the questions are not vigoriously created, using the words such as "love". As different interpretions of love are there, it is impossible to attach a single interpretation to any statistic result to say how many women are "in love".
  • Many of the questions are leading. They indicate the "right" response to make.

Unfortunately but firmly, Hite's results can only speak for the cases of women who would have responded to this survey.

Requirements of a Good Sample

Selection Bias