Statistics Of Drinking, Emailing, Pupil Size

Today, three studies from loyal readers.

Drinking Creatively

First up, the New York Post’s Kyle Smith with an anecdote:

Father O’Brien was driving home after lunch when a policeman pulled him over. “What have you been drinking?” asked the cop. “Only water,” replied the priest. “Then what’s that next to you?” said the policeman, pointing to the half-empty bottle of pinot noir in the passenger seat. “Good Lord!” said Father O’Brien. “He’s done it again!”

This broaches the worthy topic of civilized lunches. Seems that researchers set up a statistical study which got volunteers drunk and asked them to play

a game in which they were given a group of words, such as peach, arm and tar, and asked to come up with another word that could be used in combination with any of the above, such as pit.

Tipplers delivered more correct answers and delivered them more quickly. Drinkers solved nine problems on average, versus six for the sober group, and came up with answers in an average of 11.5 seconds as against 15.2 for the teetotalers.

Well, this is good enough for me. Statistics proves drinking at lunch is good for you. There must be a p-value joke in there somewhere, but it’s early and my glass of breakfast Chablis did not appear.

Control Your Pupils!

You might not be surprised to learn that ivory-tower researchers, who are not known to imbibe at their midday meals, find different ways to stimulate themselves. One of these is in measuring the “profound sex and sexual orientation differences in sexual response…based on measures of genital arousal.” Profound. This is a strategy which, interestingly enough, has “potential limitations” such as—are you ready?—“volunteer bias” (pause and reflect, dear reader, pause and reflect) plus the unacceptable circumstance that there are “differential measures for the sexes.”

Academics have now solved this discrepancy! Instead of wiring up the naughty regions of volunteers, researchers looked them right in the eye. Yes, a pair of Cornellians “assessed the pupil dilation of 325 men and women of various sexual orientations to male and female erotic stimuli.” Peer-reviewed result? “[S]elf-reported sexual orientation corresponded with pupil dilation.”

In research funded by the—drum roll—American Institutes of Bisexuality, academics discovered “Among men, substantial dilation to both sexes was most common in bisexual-identified men.” Oh, the work was also funded by you, in the form of the United States Department of Agriculture’s Federal Formula Funds. Which “Formula” we are never told. And don’t you think it’s just the USA: these tricks were first tried in the Great White North in the “1950s and 1970s.”

Another hard-core finding: “sexual attraction patterns of women are less affected by a partner’s sex and more by cultural, social, and situational variables.” Yes, truly, size matters: wallet size, that is.

How did they get their volunteers? The old-fashioned way: “from a web forum where men sought both men and women for sexual reasons.” What could go wrong? Once at the “facility”, volunteers “were seated in a dimly lit room facing a monitor.” Dimly.

Then came the data manipulation and statistical modeling, which produced one or two, but only one or two, p-values of the acceptable range. Must be a joke lingering there, too.

Email Me

More peer-reviewed research (pdf) tells us that “Depression is a serious mental health problem affecting a large population of college students.” To prove it, researchers asked some college kids how depressed they were and then measured their “average packets per flow, peer-to-peer (octets, packets and duration), chat octets, mail (packets and duration), ftp duration, and remote file octets.”

Lo! There was a positive statistical correlation, with publishable p-value, between some of these measures and the depression score.

Sure, were the Pearson r values between 0.06 and 0.28? They were, they were. And did “remote file octets” (perhaps accompanied with pupil dilation) have the best correlation with depression? It did, it did. Though the correlation between gloom and “average packets per flow” was nearly nil. That’s why they switched to the Spearman’s rho for this measure, where a successful p-value was finally found. There’s more than one path to a publishable p-value!

One conclusion: “Frequent email checking may relate with high levels of anxiety, which in-turn correlates with depressive symptoms.” Which in turn leads to the dismay of statisticians who read studies like these. Depression for everybody!

————————————————————————————–

Thanks to Al Perrella for the first two tipples, and to Nate West for the last.