Replies

Dr. Bearman(the sociologist from Columbia U. in NYC) said: "Our analyses showed that pledgers are less likely to get tested for S.T.D.'s, be diagnosed as having an S.T.D. and to see a doctor because they are worried about having an S.T.D. Most S.T.D. infections are asymptomatic, and therefore, people don't know that they have an S.T.D. unless they get tested.

By including Chlamydia the sociologist may be correct about the prevalence of all STD infections, but even Chlamydia has an eventual sign or symptom in females having no children or complaining about infertility, i.e, a sign or a symptom, respectively. While they may initially be asymptomatic, almost all STDs have other signs or symptoms that get worse with time.

He said in an interview that it was "a glaring error" to use the result of a statistical test at a 0.10 level of significance when journals generally use a lower and more rigorous level of 0.05.

That means that the statistical association, i.e. probability or P, was likely to happen by chance less or equal to than 1 out 10 times using a probability of less than or equal to 0.10, versus less than or equal to 1 out of 20 times with a P less than or equal to 0.05.

Dr. Bearman is being disingenuous to say the least if he argued that using a p <.10 was inappropriate. A quick scan of the Heritage report shows that the authors reported their results from a multiple regression analysis in entirely the correct way. True they showed where some results were significant at the p < .10, but most importantly, the majority of the results they reported were at the p <.05 or p <.01 level. It is perfectly reasonable to show a weaker relationship if what you are looking at is the overall pattern of results.

His argument about STD testing seems flawed - unless he carefully controlled for the number of partners and the perceived number of partners of a partner, i.e., two individuals who were previously virgins do not have to get tested for STDs. Still since I haven't seen his study I will refrain from further comment.

Correction!!

Apologies to Dr. Bearman. On reading the actual NY Times article, it was David Landry, a senior research associate at the Alan Guttmacher Institute in New York, who made the statement about the p < .10 being inappropriate.

For those of you unfamiliar with the Institute:
"The Institute's mission is to protect the reproductive choices of all women and men in the United States and throughout the world. It is to support their ability to obtain the information and services needed to achieve their full human rights, safeguard their health and exercise their individual responsibilities in regard to sexual behavior and relationships, reproduction and family formation." http://www.agi-usa.org/about/

Now that's what I call going to a disinterested 3rd party for an objective assessment of statistical methodology!! Another example of great journalism from the Times.

He said in an interview that it was "a glaring error" to use the result of a statistical test at a 0.10 level of significance when journals generally use a lower and more rigorous level of 0.05.

10%-significance levels are commonly reported in social-science journals.