Tag Archives: statistics

A Quantitative Analysis of Mike Duran’s Opinions

prude-advisory(Warning:  This post contains language.  Cover your virgin eyes.)

If you haven’t had the misfortune of witnessing Mike Duran’s epic tone-trolling masterpiece, well, here it is.  Summary, Mike is mad that egalitarian women are using mean words like “enabling the patriarchy” and “glorifying abuse,” all in response to a totally polite article that only cited a woman staying with her abuser as a positive example, so he’s going to take his balls and go home to the complementarian camp.  That’ll show them!

340x-1One of the first things you’ll probably notice is the sheer mildness of the comments he’s objecting to.  No name-calling, death threats, swearing, or ad homines, just things like “shame on you” (haven’t stodgy old guys been using that one for centuries?) and “downright irresponsible” (dear God, what are you supposed to say if something is downright irresponsible?).  Obviously Mike isn’t just looking for people to avoid abusive language, but also to coddle his whiny-ass white-guy sensibilities, as demonstrated in the downright farcical comment section, where women try gentler and gentler ways of attempting to allow facts to penetrate his skull, only to be rebuffed every time with “That’s the kind of comment I’m talking about!”

I’ve, therefore, taken the opposite approach; since everything that isn’t completely deferential to him hurts his precious fee-fees, I’m just making an all-out assault on them.  Why?  Because it doesn’t change the validity of my argument in the least.  Mike, if you’re reading (and, come on, we know you are), let the record state that I don’t give a shit about the opinions of people who care more about who’s polite than they do about facts.  You’re welcome to post this on your blog as an example of shrill, hysterical womenfolk who just can’t see reason, but you’re also invited to kiss my ass.

There have been many wonderful responses to Mike’s post, so I’m not going to explore the rhetorical problems with his argument, but rather look at another question: Does Mike actually have a problem with tone, or does he just have an enormous rage-boner for ladies who don’t submit to his mighty cock by agreeing with him?  And yes, I am going to approach this scientifically.  If Mike has a real, dispassionate problem with tone, we should see that expressed about equally to people who agree and disagree with him, and about equally to men and women.  In particular, he should crack down zealously on inappropriate tone from his supporters, because he sure as hell wouldn’t want them undermining his side of the argument*.

straw-feminismMethodology: I went through the thread as of 11:30 July 13 and noted each poster (aside from Mike and me), their gender if known, and whether they generally agreed with Mike.  Then I noted Mike’s response: Did he not respond, did he respond positively, neutrally, or negatively, or did he make a crackdown, condemning the comment as the very sort of thing he was complaining about?  Finally, I made a vague attempt at quantifying how many uncivil things were said.   Since Mike is touchy beyond all belief and since words like “horrible” and “irresponsible” apparently count, I’ve used the vaguely objective metric of “negative adjectives and swearwords**.”  This puts Mike into the negative from the get-go (oops, he called the feminists “acerbic!”), but whatever.

Ready for the exciting results?  Here they are!  Raw data available as an XLS file.

  • Out of 49 total posters (not counting Mike and me), 32 were female and 13 were male (the rest were unknown).  Women left 70% of the comments, men 29%.  Top poster was the formidable Katherine Coble by a landslide (64 comments), followed by Mich Pendergrass (43 comments) and top male poster Alan Molineaux (38 comments).
  • A victory for human decency: 63% of the posters generally disagreed with Mike!  33% agreed and a few left neutral or unclear comments.  Women disagreed by a landslide (23 to 7), of course, while men agreed by a small margin (7 to 5).
  • Mike made a total of 10 comments that I classified as crackdowns: Three against Sara, one against Jill, two against Alan, two against Alise, and two against Maya3.  Thus, 80% were against women and 20% against men.  Normalized for the number of male and female posters, we find 0.25 crackdowns per female poster and 0.15 crackdowns per male poster.
  • Women received more crackdowns relative to the number of posts and negative adjectives they had posted.  Women received 0.17 crackdowns per comment and 0.25 per negative adjective, to men’s 0.05 and 0.13, respectively.
  • I’m sure you’ll be stunned to discover that 100% of the crackdowns went to people who disagreed with Mike, even though 29% of the negative adjectives came from people who agreed, versus 63% from those who disagreed (about the same as the ratio of comments, 33% to 63%).

Miscellaneous observations:

  • By far the most negative tone in the whole discussion came from…drumroll please…Mike himself!  This makes perfect sense; since his original post was complaining about others’ tone, everyone was on their best behavior so that he wouldn’t accuse them of being part of the problem, except Mike, who didn’t have to worry about such an accusation.
  • Negative adjectives often come in strings.
  • Almost everyone on his blog used a real (sounding) name, often full names.  I’ve noticed that Christian blogging circles have a higher rate of real names than secular ones, even though they rarely have official policies about real names; I don’t know why this is.
  • Violet, Sweet Pea, and Ralphie being catty

    Violet, Sweet Pea, and Ralphie being catty

    Surprise, surprise: When women tried to speak mildly and inoffensively so that they wouldn’t get labeled “shrill” and “acerbic,” they got called “wishy-washy” and “catty.”  You can’t win.

  • The phrase “shame on X” was used quite a few times in the thread.  Although Mike repeated the “shame on you” citation from the OP several times in a row as an example of hateful, inflammatory language, not once did he crack down on its use in his comments…towards people he disagreed with.
  • While Mike called Sara a troll for pointing out that he had benefited from feminism, Sally literally calls for people to take abuse in the name of Christ and Mike merely asks her to “clarify.”
  • Mike is as dense as a neutron star.  Go ahead, read that thread and despair as people try to explain simpler and simpler concepts to him and none of it makes a dent.  Abuse survivors may have a better understanding of abuse?  Saying “women submit to their husbands as men submit to Jesus” is treating men like gods?  Positively citing an abusive situation is glorifying abuse?  You wouldn’t call a man “shrill?”  Women face disadvantages in society?  LALALA I CAN’T HEAR YOU!
  • He also has trouble spelling women’s names.
  • What the hell is the manosphere doing in this thread?  (The answer is, of course, wanking about white feathers, but where did they come from?)  Same poster is also a grade-A example of mansplaining: Not female, married, or an abuse survivor, but you must totally listen to his opinions on women, marriage, and abuse!

1347584048956_3393551In conclusion, Mike is a whiny little bitch with a congenital inability to pull his head out of his anus.  He wouldn’t know a strong argument if it bit him in the ass, his massive ego routinely pulls satellites out of orbit, and as a result, his reptile brain has no more advanced view of social justice issues than “I’m not trying to oppress people and anyway I’m colorblind.”

This post won’t change him.  Being nice wouldn’t change him, either.  I’m not trying to change him, because he’s incapable of change.  For the record, I don’t hate him.  I’m not even angry at him.  I am, in fact, amused, and I’m poking him with a stick for my own amusement and yours.

*Feminists do this all the time.  If they think body-shaming is wrong, you’re not getting away with saying “I wonder if the Republicans in Congress are compensating for something?”  If it’s wrong, it’s wrong, no matter who it’s directed at!

**Negative nouns and verbs, sadly, fall outside the scope of this particular post.  I didn’t count words in quotes and exercised some discretion about skipping words that were paraphrasing something that someone else had said.  There’s an inevitable degree of subjectivity even in quantitative studies like this.

Prude advisory found here.  Privilege denying dude found here.  Straw feminists, of course, by Kate Beaton.  Whiny-ass titty baby found here.



Filed under Uncategorized


So.  The election is over.

Image from Wikipedia

For inveterate poll-watchers like myself, it’s fascinating–if a bit anticlimactic–to see a years’ worth of wild speculation finally replaced with one single, incontrovertible set of data–the final electoral map.  No more speculating on whether polls oversample certain demographics or how voter-ID laws will affect the outcome; the outcome is now known.  This abrupt switch from entirely abstract guesswork to entirely concrete fact is seen practically nowhere except elections and sports; as such, it provides an interesting glimpse into the difference between reality and peoples’ perception of it.

By which I mean, we now have proof that conservatives are completely disconnected from reality.

I’m not talking here about ideology and values; what you believe to be right or wrong can’t be proven or disproven (although, if you believe something to be wrong because of its consequences, then whether those consequences really happen can be subject to proof).  So, for instance, the election doesn’t prove that conservatives are wrong about gay marriage.  It does, however, prove that conservatives are wrong about the general population’s opinions towards gay marriage, since common conservative wisdom stated that gay marriage would never be legalized by popular vote; Family Research Council’s Tony Perkins still claims that “contrary to what the Left will say, the narrow margin for victory in these four states offers plenty of evidence that a solid majority of Americans still opposes same-sex ‘marriage.'”  Losing four out of four proves that you have popular support?  It sounds like any result would convince Tony Perkins that people agreed with him.

But let’s return to the presidential election.  Sabermetrician-turned-psephologist (there are two words you don’t get to use very often) Nate Silver, whose blog, FiveThirtyEight, rose to fame through his accurate prediction of the 2008 election, faced some harsh criticism near the end of this cycle from conservatives who balked more and more as his estimate of Obama’s win chance rose from around 60% after Denver to a final value of 90%.  At the end of October, when Obama had picked up the momentum that would carry him through to Election Day, MSNBC’s Joe Scarborough railed that Silver’s current estimate of 73% was ridiculous:

And anybody that thinks that this race is anything but a tossup right now is such an ideologue, they should be kept away from typewriters, computers, laptops and microphones for the next 10 days, because they’re jokes.

And then there was Dean Chambers.  Like many other conservatives, he believed that Silver’s model introduced bias by weighting different polls to produce the result he wanted.  And how do we know that Silver wanted to influence the election for Obama?  Quoth Chambers*:

Nate Silver is a man of very small stature, a thin and effeminate man with a soft-sounding voice that sounds almost exactly like the “Mr. New Castrati” voice used by Rush Limbaugh on his program.  In fact, Silver could easily be the poster child for the New Castrati in both image and sound.

You just can’t trust those skinny guys.  Luckily, Chambers runs his own, bias-free election model, Unskewed Polls.  Let’s see how his model fares against Silver’s.  Their final November 5 predictions are shown at left.  I’ve labeled the graphs, although I doubt I really need to.

As poll watchers already know, election night was a triumph for Silver and statistics nerds everywhere, as he correctly predicted 50 out of 50 states (his electoral college prediction, being an aggregate, was actually 20 points short).  Meanwhile, Chambers was off by a humbling 69 electoral votes.

But there’s more to this than “Silver was right, Chambers was wrong.”  Silver made a careful prediction, using a logical algorithm to weigh the different factors and laying out his reasoning every day on his blog, noting any changes from the previous day’s prediction and explaining why it happened.  Meanwhile, Chambers only gives a vague description of a variety of factors without any indication to where or how much each factor came into play.  Chambers’ map, in fact, doesn’t look like a prediction at all; it looks like a calculated effort to create a plausible-looking win for Romney while giving him the fewest necessary swing states.

The best evidence for this is Unskewed Polls’ previous prediction, from October 25, when Chambers produced the truly remarkable map shown here.

Nobody could possibly mistake that map for an actual prediction, right?  It has Oregon going for Romney.  Oregon, which hasn’t been won by a Republican since 1984, and which Obama won by 16 points in 2008.  Romney wasn’t even campaigning there!  This is obviously just partisan cheerleading.  Then, the day before the election, Chambers published the map shown above to maintain some credibility while still promoting Romney.  Theoretically he could pass the second map off as a revision based on recent polling data, similar to Silver’s procedure, but curiously, he doesn’t even try to.  The methodology instead only notes that there are 11 disputed swing states (still including some long shots like Michigan) and 39 states where the results are agreed upon…including Oregon, New Mexico, or Minnesota.  I couldn’t find any explanation for his choice to award those states to Romney in the first prediction or to subsequently change his prediction.  He seems to have just realized that his first prediction made him look ridiculous and attempted to sweep it under the rug.

Chambers had to know that his prediction was ludicrously wrong; likewise, at least some the many conservative pundits who predicted a Romney landslide had to realize what a long shot it was.  But some–most infamously, Karl Rove–really seemed to believe it, right up until it became mathematically impossible on election night.  Why?  What motivates people to cling to a prediction they know to be wrong?

I believe there are two primary reasons.  First, there’s the obvious effort to influence the election via the bandwagon effect.  Second, I think we’re seeing the effects of the ever-more-stringent Republican demands for party loyalty.  Suggesting that Romney might lose, after all, is tantamount to an attack on him as a candidate, since it suggests that he lacks the chops to actually win the vote.  There’s a degree of pride in statements that the results would be a surprise, or in predicting that Romney would win a state that most projections gave to Obama.  For instance, Dick Morris was clearly trying to set himself above other predictors when he said about his prediction of a 325-213 Romney victory:

It will be the biggest surprise in recent American political history.  It will rekindle the whole question on why the media played this race as a nailbiter where in fact Romney’s going to win by quite a bit.

Anti-intellectualism plays into this, showing up both in the attacks on Nate Silver and in the tendency to base predictions on a gut feeling, rather than poll data, such as Rush Limbaugh’s statement that “common sense tells me this election isn’t gonna be close” (Limbaugh also cited the odd “Redskins rule,” which bases the winner on the outcome of a Washington Redskins game).  So can the right’s ties to evangelicalism, as in Glenn Beck’s assertion that God put Romney behind in the polls so that Romney’s victory would clearly be a miracle.  Either way, though, the core principle remains: Find a reason why Romney will win, and if there isn’t one, look harder.

In other words, party loyalty encourages Republicans to refuse to believe obvious reality.  And, as they have just discovered, this is never worth it.  Because reality intrudes.  It stubbornly insists on happening just as it was always going to regardless of how many equivocations or justifications you make.  The only difference is that you’re unprepared.

XKCD by Randall Munroe

Although I don’t expect it will, we can always hope that the victory of the nerds will cause conservatives to rethink their out-of-hand dismissals of statisticians, scientists, economists, and everyone else who uses data to suggest that common conservative wisdom is wrong.  Maybe global warming is happening.  Maybe fracking does have a negative impact on the land.  Maybe cutting taxes on the highest earners doesn’t grow the economy.  And maybe acknowledging these facts and crafting a platform around them is wiser than clinging to a position contradicted by plain evidence.

We can only hope.


*This links to a Gawker article, rather than Chambers’ original (which is linked at the beginning of the paragraph), because Chambers later deleted that paragraph and apologized.  After the election.  Almost as if, had Silver been wrong, Chambers would have redoubled the “you can’t trust femmy guys” line of attack.

**I’ve linked this HuffPo article, rather than FiveThirtyEight (linked in the body text), because FiveThirtyEight’s actual prediction maps are shown in a sidebar, rather than in an article, and therefore will probably become unavailable eventually.

Leave a comment

Filed under Uncategorized

Gender in Film: A Brief Quantitative Analysis

Attempts to point out the gender imbalance in the movie industry often meet with allegations of anecdotal evidence; pointing out that a film doesn’t pass the Bechdel Test is likely to be greeted with one of three responses: Either that it’s only one film and not representative, or that this particular film had a good reason not to pass and ought to be the exception, or that the Bechdel Test is arbitrary and meaningless.  Thus, I’m here presenting a quantitative study of the relative number of film roles for men and women.

Since men and women each comprise about half the population, you’d expect each to have an approximately equal number of roles in movies* (unless you think that women’s lives are inherently less important or interesting, in which case you may feel free to smack yourself).  This doesn’t mean, as is often assumed, that each individual movie should have a cast that’s half male and half female.  Rather, it means that films should fall into a rough bell curve; 50 randomly-selected films might look like this:

The majority have roughly equal numbers of men and women, some have more men, some have more women, and a small minority are all or almost all male or female.  If you prefer, the curve could be flatter, with more films at the ends and relatively fewer in the middle, but either way the overall male/female ratio should be about even.

I looked at Wikipedia’s 50 highest-grossing films of the 2000s, using IMDB cast lists to count the male and female characters.  When the cast was listed in order of importance, I used only the first billed cast members; when listed alphabetically (The Lord of the Rings) or in order of appearance (Harry Potter) I used the entire cast.  In either case, I usually skipped unnamed characters and characters without IMDB character pages, a strong indicator that it’s a tiny bit part, but exceptions had to be made on a film-by-film basis because neither naming nor IMDB pages are necessarily consistent (for instance, I wouldn’t want to omit Tigress from Kung Fu Panda for not having a real name).  Characters without gender were usually counted as the gender of the actor or actress, because there is significance to the sorting hat from Harry Potter and the computer from WALL-E being voiced by male actors.  The results are below.

Drastically different.  It’s closer to an S-curve than a bell curve: Until you pass the 90% male mark, each bracket contains more films than the one below it.  Thus, not only is a film extremely likely to have more men than women, but it’s more likely to be two-thirds male than half male, more likely to be three-quarters male than two-thirds male, and more likely to be four-fifths male than three-quarters.  Only Mamma Mia! prevents all 50 films from containing more men than women.

The obvious objection is that high-grossing movies are usually summer action films and children’s movies, which are not representative of movies in general–aside from Mamma Mia!, chick flicks are absent, for instance.  The 2000s might also be non-representative because they’re skewed by franchises like The Lord of the Rings and Transformers.  The first objection raises the question why action movies and children’s films are not expected to have a strong representation of female characters, and the second is effectively canceled out by the more gender-equitable Harry Potter films, but fair enough.  Let’s have a look at the 20 most-acclaimed films of the decade, as analyzed by Metacritic.

Better; the 90%-100% male category is now empty, and, even though there are fewer total films, the 40%-50% category contains multiple films.  If you take out the three Lord of the Rings films, only one remains in the 80%-90% category (The Dark Knight).  On the other hand, the foreign films (Amelie, 4 months, 3 weeks, and 2 days) fall towards the lower end of the graph, so maybe it’s just American cinema that’s the problem here.

But the graph is still centered at 60%-70%.  That is, the average acclaimed film has twice as many men as women, or to put the same thing another way, men’s voices are considered twice as important to express in film.  And there still isn’t a single film with more than 60% women: Everything is either mostly men or about even.  Women’s stories, with a large majority of women in the cast, are simply not present.

And now for the second round of objections.  Yes, there are many films that are neither popular nor critically acclaimed, but at that point you’re grouping movies starring women with Gigli.  I don’t have time to get into the thorny mess of rebuttals–nobody wants to see movies about women; all movies about women are bad–except to say that a few really good female-dominated films could start changing that landscape.

Men are not twice as important as women.  We can do better.


You can find my original data here (.xls).

*People who don’t identify as either male or female are, of course, egregiously underrepresented in film as well, but that’s a much more complex issue to address.  As far as I know, none of the films studied included human characters who identified as neither male nor female.


Filed under Uncategorized

A Statistical Analysis of Women’s Pastoral Roles

Scot McKnight of Jesus Creed posed this question a few days ago:

How many of you, in your church today  had a woman preach, a woman teach, or a woman lead worship? Second, how many of you have a woman who has the title of “pastor”?

He has received nearly 200 responses, which I have compiled into some interesting, if statistically meaningless, data, which I present here hopefully free of my usual veneer of sarcastic commentary.

Question 1: How many of you, in your church today,  had a woman preach, a woman teach, or a woman lead worship?

Yes: 104 total

  • 38 had a woman preach
  • 17 had a woman teach
  • 42 had a woman lead worship
  • 30 did not specify

No: 70 total

  • 28 have had women participate in these capacities in the past
  • 8 haven’t in the past (or didn’t specify), but would allow it
  • 12 would not allow it
  • 22 did not specify


  • 120 would allow women to serve in at least one of the mentioned capacities
  • 34 would not allow it or did not specify

Question 2: How many of you have a woman who has the title of “pastor”?

Yes: 87 total

No: 83 total

  • 5 have had female pastors in the past
  • 15 haven’t in the past (or didn’t specify), but would allow it
  • 20 would not allow it
  • 5 have no “pastor” or equivalent title
  • 37 did not specify


  • 107 would allow women to serve as pastors
  • 62 would not allow it, did not specify, or have no “pastor” role

I reiterate here that these statistics are not actually indicative of anything because of the large selection bias inherent in Scot McKnight’s readership: as a politely but vocally egalitarian blog, Jesus Creed tends to attract egalitarians, or at least people who are open to the viewpoint.

Here are a few other interesting trends that I noticed but didn’t enumerate:

  • Many “yes” responses were husband-wife pastoral teams.
  • UMC churches were the most widely represented among “yes” responses.
  • “No” responses were the most likely to quote a Bible verse (only 1 Tim. 2:12).

More statistics, methods, and caveats after the cut.

Continue reading

Leave a comment

Filed under Uncategorized