Booze and Turkey

Here's a forum post I wrote for my economics class. It is about statistics, drinking while pregnant, and myths by the National Turkey Federation. Enjoy!

Because I haven't used statistics much in my professional life, I found a website called STATS that is based out of George Mason University. It is devoted to checking out numbers thrown around in the media and using statistics and science to explain current events or trends. There are many articles on this site, and there's a great tab called "STATS Simplified" that explains how examples of tools like percentages and margin of error are sometimes misrepresented; here are two examples of articles which use statistical analysis:

In Light to Moderate Drinking During Pregnancy, Rebecca Goldin explains why the US Surgeon General recommends abstaining from all alcohol during pregnancy while the Department of Health in the UK and many OB-GYNs say that light to moderate drinking (three to six small glasses of beer or wine a week is considered moderate) is okay during pregnancy. While it is known that drinking heavily during pregnancy causes problems such as Fetal Alcohol Syndrome, Goldin references studies that show light drinking has no ill effect on babies. So why the stringent recommendation? It comes down to statistics and reporting!

One reason is that "light" or "moderate" has not been defined the same way by all researchers or by the women. "If a study is meant to examine the effects of consuming two drinks a day, but it includes some women who actually consumed five drinks per day, the statistics will be skewed" (Goldin); conversely, if women who have only one drink a week were lumped into the study, then the statistics would be slanted the other way.

Additionally, because the women in the studies are told about the health risks of drinking heavily during pregnancy, they might underreport what they are really drinking or convince themselves that they are sticking to a moderate level. Interestingly, at the end of the article, Goldin references a study by Willford, Leech, and Day (Alcoholism: Clinical and Experimental Research, 2006) which compares IQ levels of 10 year-olds whose mothers drank moderately verses those who abstained from all alcohol in the first and second trimesters. There was a small shift in African-American children whose mothers drank moderately, but no "statistically significant" change in Caucasian children; for mothers who drank lightly, there were no changes in anyone.

Goldin uses standards of statistical reporting to caution against making sweeping conclusions about genetic differences from this study. She references social and cultural effects that were not mentioned in the study; for example the interviewers were Caucasian, which might have affected the African-American mothers’ reporting. Additionally, she suggested that the researchers should have reported estimates for the Caucasian children's IQ results, even if not significant, because could hint at a trend.

Finally, Goldin mentions, "And because they were doing a “racially stratified” analysis, they should have reported the number of black vs. the number of whites in the study; statistical significance depends on these numbers, and the significance for blacks and not whites could be an indicator that there simply weren't enough whites to make the same conclusions as they did for blacks." At the end of the article, Goldin says women shouldn't just "have a hay-day" but should just relax, whether they have a drink or not.

A lighter article referenced on the website is about how numbers can be misinterpreted:

In a Claims of Thanksgiving Excess Fueled by Feast of Fuzzy Data in The Wall Street Journal, Carl Bialik uses logic to sort through the statistics that are often thrown around by the media and special interest groups about Thanksgiving.

The first set of numbers, generated by a market research firm called IBISWorld USA and the National Turkey Federation, estimated that about 46-47 million turkeys are eaten on Thanksgiving Day and another 22 million are usually eaten on Christmas Day.

The "USDA says 23% of the 270 million turkeys processed in the U.S. last year" (Bialik, A20) were slaughtered from October to December. If the research is right, that would mean almost all of the 68.3 million turkeys eaten in that quarter would be eaten on those two days alone; if this were true, people's "turkey consumption on other days in October and November would have been about half the daily average the rest of the year” (p.A20). He states that these numbers are exaggerated because of an assumption that all turkeys bought in November must be for Thanksgiving and all those in December must be for Christmas.

Bialik also debunks the myth that Thanksgiving is one of the busiest days to travel in the year. He writes, "In fact, no day last November figured in the top 220 out of the 366 days in 2008, based on the number of flights reported by airlines to the Department of Transportation" (p.A20). However, it does have one claim to fame: "busiest plumbing period. The day after Thanksgiving typically brings 6,200 jobs for Roto-Rooter plumbers. That is more than any other day, and about 50% higher than the typical Friday volume, according to spokesman Paul Abrams" (A20).


Resources:

STATS at George Mason University. http://stats.org/index.htm



Bialik, C. (2009, November 20). Claims of Thanksgiving Excess Fueled by Feast of Fuzzy Data. The Wall Street Journal, A20. Retrieved from: http://online.wsj.com/article/SB125910575363163091.html



Goldin, R. (December 12, 2007). Light to Moderate Drinking During Pregnancy. STATS at George Mason University. Retrieved from: http://alcoholnews.org/ALCO_pregnancy.html