Reference no: EM133661896
Assignment:
The controversial psychologist Satoshi Kanazawa wrote a book titled Why Beautiful People Have More Daughters. Apparently there are more beautiful women in the world than there are handsome men (or at least men tend to perceive/rate more women as attractive compared to the fewer proportion of men that women perceive as attractive (Rudder, 2014; see graph below). Some evolutionary psychologists might explain this by hypothesizing that it is adaptive for many men to have lower standards for perceiving someone as physically attractive when it comes to temporary or short-term mates. Similarly, you might say that it's adaptive for women to have high standards because the costs of potentially sexually reproducing are much greater to women in terms of parental investment (e.g., 9 months of gestation, years of lactation).
Controversially, Kanazawa explains the finding that "there are more beautiful women than handsome men in the world" in terms of good-looking parents being more likely to have a baby daughter as their first child than a baby son. He suggests that, from an evolutionary point of view, beauty is a more valuable trait for women than for men (Kanazawa, 2007). In a playful and very informative paper, Andrew Gelman and David Weakliem discuss various statistical errors and misunderstandings, some of which have implications for Kanazawa's claims. The 'playful' part of the paper is that to illustrate their point they collected data on the 50 most beautiful celebrities (as listed in People magazine) of 1995-2000. They counted how many male and female children they had as of 2007. If Kanazawa is correct, these beautiful people would have produced more girls than boys.
A t-test to find out whether they did. The data are in Gelman_& Weakliem (2009). Even though you have a directional prediction from Kanazawa that there will more daughters than sons, I want you to keep it as a two-tailed test (i.e., select Measure 1 does not equal Measure 2). Two-tailed tests are often the "default" in statistical programs including JASP and I'm in favor of using them in most cases even if you have a directional prediction. I know the book emphasizes recognizing one-tailed vs. two-tailed tests and it's useful to be able to do so (and it is), but in practice many researchers often stick with two-tailed tests.
Be sure to test your assumption of normality, calculate an effect size, get descriptive statistics, and a descriptive plot with a 95% confidence interval.