Tuesday, February 27, 2018

The "title track law"

The "title track law", as proposed by me, states that every album has a title track, in the sense that even if there is no song on that album with the same name as the album, there is still going to be a song with the name of the album in its lyrics. (Of course, this only applies to genres where the music has lyrics.)
Here is a partial list of albums with "title tracks" that contain the album's title in their lyrics but do not have the same title as the album:
U2's "All That You Can't Leave Behind". Title track (TT) = "Walk On" (The only baggage you can bring/Is all that you can't leave behind)

Everclear's "Sparkle and Fade". TT = "Summerland"

(Everything we want to be
We could get lost in the Fall
Glimmer Sparkle and Fade
The Sparkle and Fade)


Modest Mouse's "Good News for People Who Love Bad News" = "Bury Me With It"

(Good news for people who love bad news.
We've lost the plot and we just can't choose.
We are hummingbirds who are just not willing to move.
And there's good news for people who love bad news.)


The Pixies' "Doolittle". TT = Mr. Grieves ("Pray for the man in the middle/The one who talks like Doolittle")


The Shins' "Chutes too Narrow". TT = Young Pilgrims ("I fell into a winter slide/And ended up the kind of kid who goes down/Chutes too narrow/Just eking out my measly pipes")


Party of One's "Caught the Blast". TT = Baghdad Boogie ("Caught the blast from 30 feet away/85 degrees in the shade")

The Russian Futurists's "The Weight's On The Wheels". TT = Hoeing Weeds Sowing Seeds ("And that great weight that you carry/From your born-day 'til you're buried/Well, now that weight's on the wheels"

Saturday, February 17, 2018

How not to do a literature review

This is a step-by-step guide to how to do a literature review as badly as possible, following the remarkably good example of a bad review provided by Rushton & Templer (2012). For additional debunking of this "review article" (if it even deserves to be called that), see this blog post by another blogger and the comment section thereof.

Step 1: Come up with a predetermined conclusion based on that of your own research, and make sure the entire review article fits this theme. This second part is important, so keep it in mind for later on. 


Here's what this looks like in our example: if you're "race realists" like Rushton & Templer, don't pay attention to the numerous criticisms of your methodology that have been published over the past several decades. Instead, make a bald-faced statement in the abstract like this: "Both within human populations (e.g., siblings), and between populations (e.g., races, nations, states), studies find that darker pigmented people average higher levels of aggression and sexual activity (and also lower IQ)." But if you wanna compete with the masters of writing bad review articles, don't stop there: make sure to shoehorn the bald-faced statement into your own pet theory to create the appearance of empirical support for the theory. 


So how do R&T do this? Right after the quote I included in the paragraph above (from the abstract), they write: "We conceptualize skin color as a multigenerational adaptation to differences in climate over the last 70,000 years as a result of “cold winters theory” and the “Out-of-Africa” model of human origins. We propose life history theory to explain the covariation found between human (and non-human) pigmentation and variables such as birth rate, infant mortality, longevity, rate of HIV/AIDS, and violent crime."

Step 2: Cite a bunch of articles published by you and your BFFs who totally agree with all your ideas--don't let any of the haters who say you're just part of a tiny clique of ideologically motivated cranks get you down! No no, what you need to do is cite articles published by you and other researchers with the exact same view on this subject. Make sure that you do not include any studies criticizing your or your buddies' work.


So when you're discussing the putative relationship between race and violent behavior, for example, you don't want to include any of the numerous papers that have criticized your methodology. No, instead, you want to do something like what Rushton & Templer did in the "Human studies" section of their paper, when they wrote,


" In the US, Taylor and Whitney (1999) analyzed the FBI Uniform Crime Statistics and National Crime Victimization Surveys from the US Department of Justice and found that since record keeping began at the turn of the century and throughout the 1960s, 1970s, 1980s, and 1990s, African Americans engaged in proportionately more acts of violence than other groups...Rushton and Whitney (2002) analyzed the 1993–1996 INTERPOL Yearbooks and found that across 100 countries, the rate of murder, rape, and serious assault is four times higher in African and Caribbean countries than elsewhere in the world. In violent crimes per 100,000 people, the rate for African countries was 149; for European, 42; and for Asian, 35. These results are similar to those carried out on other data sets from INTEROL and the United Nations. They show the Black overrepresentation in violent crime to be a worldwide phenomenon."

The key lessons to take from the above passage, students, are to only include studies by hereditarian "race realist" researchers like Rushton himself, Jared Taylor (founder of white nationalist magazine American Renaissance), and Glayde Whitney (animal behavior geneticist turned white supremacist).


Then they follow step 2 even more later in the same section, citing a paper on psychopathic personality by eugenics-advocating psychologist Richard Lynn, again without mentioning published criticisms of this research or other studies that contradict it.

Later they one-up themselves yet again by citing two studies on race and sexual behavior co-authored by Rushton (one from 1987 and one from 1988) without mentioning the numerous criticisms of both studies (e.g. here and here). 


3. Sprinkle a bit of sources by third parties in that vaguely have to do with your argument and subtly imply that they totally back up your conclusions when they do no such thing. This may be because they don't control for confounding factors (as is the case in the example discussed below). Just cite racial disparities in various outcomes that you're looking for without considering representativeness of samples, how outdated they are, or potential environmental causes of the observed disparities (this last one is especially important). Just assume everything's mainly genetic in origin.


Let's take another look at the R&T paper: 


"In Canada, a government commission found that Blacks were five times more likely to be in jail than Whites and 10 times more likely than Asians (Ontario, 1996). In Britain, the Home Office (1999) found that Blacks, who were 2% of the general population, made up 15% of the prison population."


In the quoted passage from R&T (2012) reproduced above, the authors misleadingly cited 2 government reports (one in Canada and one in Britain) that look at racial disproportionality in prison populations without considering the sources of this disparity. Then they imply that this disparity implies not only true race differences in offending, but also that this, in turn, validates their claim that individuals with darker color skin (i.e. Blacks) are more likely to engage in violence than are members of other races. Why is this misleading? Because 1) Differential treatment (i.e. discrimination) in the criminal justice system at any or all of its many stages undoubtedly plays at least a small role in the disparity in question, and 2) there is no reason to assume that all the crimes for which individuals in any country are incarcerated are those of violence (unless you wanted to make a point to fit your narrative with no regard for its accuracy, of course). Of course, this also ignores the very fundamental and important fact that what is or is not a crime is not invariant across time or cultures, but instead is a social construct determined by forces in a specific society.  


Next, you might want to add some statements about race differences in STD rates, and you should know by now to avoid even discussing environmental factors like incarceration that may contribute to such differences. Here's how the experts do it: "African descended people are over-represented in rates of sexually transmitted diseases [STDs] such as syphilis, gonorrhea, herpes, chlamydia, and HIV/AIDS (US Centers for Disease Control, 2009). Of the more than one million people living in the US with HIV/AIDS in 2007, almost half (46%) were Black. The Black–White difference in HIV/AIDS is found worldwide with high levels in sub-Saharan Africa, for example, Botswana (24.8%), South Africa (17.8%), Zambia (14.6%) and Zimbabwe (14.3%) (CIA World Factbook, 2010)."


4. Make very broad and controversial statements to support your arguments. Be careful not to cite sources for these statements, which is probably just what the haters want you to do.


There is perhaps no better example of this than in the review we are examining here: immediately after the excerpt quoted above, R&T write, "
Since victims’ surveys tell a similar story, the differences in arrest statistics cannot just be attributed to police prejudice." No citation is given for this claim, despite it being an obviously controversial one.


5. Last but not least, make sure to cite outdated sources rather than more recent ones that contradict your overall argument.


R&T follow this point while also following point #2 (i.e. don't mention critics), when they write the following: "...Blacks have the most testosterone (Ellis & Nyborg, 1992), which helps to explain their higher levels of athletic ability (Entine, 2000). Testosterone acts as a “master switch.” It goes everywhere in the body and affects many bio-behavioral systems. It affects self-concept, aggression, altruism, crime, and sexuality, not just in men, but in women too. Testosterone controls muscle mass and the deepening of the voice in the teenage years. It also explains why Black women have the most premenstrual syndrome (PMS) and East Asians the least." 


Hoo boy, there's a lot of BS there! But the most salient point relevant to point #5 is that they include only a paper from 26 years ago (Ellis & Nyborg), while neglecting to mention more recent studies like this one and this one which find no race differences in testosterone levels. 

Saturday, February 10, 2018

Abortion vs. guns: which one should (and do) we regulate more strictly?

[Note: Another old post that's been sitting in my archive since mid-2016, when I originally wrote almost all of it. I recently decided to dig it back up, add a bit of content to the end to make it seem at least somewhat complete, and post it. Enjoy! -JPS]

The gun control-abortion laws comparison has been brought up many times, typically by Democrats like Trevor Noah. The argument is that Republicans oppose gun control very staunchly, but they seem to support "abortion control" (i.e. laws that make it harder for women to obtain abortions) with comparable fervency. This, Democrats say, is logically inconsistent, since if Republicans are really "pro-life", as they claim to be all the time, they should support measures designed both to reduce abortions and gun violence.

In this post I will try to answer two questions, and discuss others' answers to them:
1. Why don't Republicans support reducing gun violence through restrictive gun laws if they support reducing abortions through restrictive abortion laws?
2. What would it mean if, as one Missouri politician proposed last December, we regulated guns like abortions?

The typical Republican response to this first question tends to be that they aren't opposed to gun control but rather, are merely "favoring the Second Amendment." Similarly, they may say that restrictive abortion laws are aimed at "saving the unborn." This seems to cover the first question, so let's move on to the second.

So let's see whether guns are more strictly regulated in this country than abortions, or vice versa. Let's start with a popular meme, which you can see at this link. The meme's first observation is that people who want to have abortions have to undergo mandatory waiting periods in many Republican-controlled states, but this is not the case for people who want to buy guns. Now it's worth noting that abortion waiting periods vary from state to state: 27 states mandate them.
2. ArizonaCurrently Enforced24 hours1
2. GeorgiaCurrently Enforced24 hours
2. IdahoCurrently Enforced24 hours
2. KansasCurrently Enforced24 hours
2. KentuckyCurrently Enforced24 hours
2. LouisianaCurrently Enforced24 hours1
2. MichiganCurrently Enforced24 hours
2. MinnesotaCurrently Enforced24 hours
2. MississippiCurrently Enforced24 hours1
2. NebraskaCurrently Enforced24 hours
2. North DakotaCurrently Enforced24 hours
2. OhioCurrently Enforced24 hours1
2. PennsylvaniaCurrently Enforced24 hours
2. South CarolinaCurrently Enforced24 hours
2. TexasCurrently Enforced24 hours7
2. VirginiaCurrently Enforced24 hours7
2. West VirginiaCurrently Enforced24 hours
2. WisconsinCurrently Enforced24 hours1

The length of mandated waiting periods varies considerably among these states: 18 hours in Indiana, 24 hours in 16 other states (shown in the chart above), 48 hours in Alabama, Arkansas and Tennessee, and 72 hours in Missouri, North Carolina, Oklahoma, South Dakota, and Utah. Now what do gun waiting periods look like on a state level? Because it's worth noting that there is no federal law requiring a waiting period for guns any more than there is one for abortions. Only 9 states out of 50 (as well as the District of Columbia) require waiting periods for some or all gun purchases. So clearly, abortions are regulated more strictly than guns in this respect. Or are they? Because the longest abortion waiting period is 72 hours, but among the 9 states and DC that have them for guns, they are mostly longer than that:
States Imposing Waiting Periods for Purchases of All Firearms
State (Waiting Period)
California (10 days)15
District of Columbia (10 days)16
Hawaii (14 days)17
Illinois (24 hours) (long guns); 72 hours (handguns)18
Rhode Island (7 days)19
States Imposing Waiting Periods for Purchases of Handguns and Assault Weapons
State (Waiting Period)
Minnesota (7 days)20
States Imposing Waiting Periods for Handguns Only
State (Waiting Period)21
Florida (3 days)22
Iowa (3 days)23
Maryland (7 days)24
New Jersey (7 days)25

As you can see, all the waiting periods above are at least as long as the longest for abortions (72 hours), except the 24-hour waiting period in Illinois for long guns. Interestingly, only one state has mandated waiting periods for both guns and abortions: Minnesota, which has a waiting period for handguns and assault weapons (but not other guns) of 7 days, 7 times longer than its abortion waiting period. So in that state, guns are regulated more strictly than abortions, but overall it's clear that more states have abortion waiting periods than gun waiting periods. Though, to be fair, the 10-day waiting period in California has been struck down in the case of experienced gun purchasers who have already passed a background check.

Let's move on to the issue of parental permission. In all but 13 states, such permission is required for women under the age of 18 who want to have abortions. Under federal law, by contrast, you must be at least 18 to buy a long gun or a handgun, with one key exception: you can legally buy a long gun no matter how young you are if you do it from an unlicensed dealer. The number of states where parental permission is not required for people under the age of 18 who want to buy a gun appears to be 0: there are 10 states where you can buy a long gun without parental permission, but in all 10 of these states, you must be 18 or older to do so legally.

One area where there's no competition is in the sheer number of abortion clinics vs. licensed gun dealers: there were a total of 788 abortion clinics in the entire U.S. (as of 2014). The total number of licensed gun dealers in the US appears to be the sum of the values in columns 1 and 2 in the ATF's monthly table (here is the one for September 2017). This shows that just in New York, there were 1,769 total gun dealers that month--over twice the # of abortion clinics in the entire country!


Sunday, February 4, 2018

On the impact factors of two pseudo-academic journals

Here I try to examine the credibility of 2 not-very-highly regarded "peer-reviewed" journals devoted largely, or entirely, to the field of psychometrics: Mankind Quarterly (MQ), the most notorious pseudoscientific "race realism" journal in the history of the entire world, and Open Differential Psychology (ODP), a new-ish open access journal started by Emil Kirkegaard, who only has a bachelor's degree in an unrelated field (linguistics). Kirkegaard also publishes two other journals on the website openpsych.net: Open Behavioral Genetics (OBG) and Open Quantitative Sociology & Political Science (OQSPS).

Consider MQ first. The most cited article on Google Scholar (GS) published in MQ in 2015 or 2016 is "Admixture in the Americas: Regional and National Differences" by Kirkegaard and Fuerst. The article in question has been cited 22 times on GS since it was published in 2016. Sounds pretty solid-but of these 22 citations, all of them were either to non-peer reviewed sources or articles published in MQ or one of Kirkegaard's "Open" journals...except for 3 (14%) (articles 1, 2, 3). When one looks at the 7 citations for "The genealogy of differences in the Americas" by Fuerst, one similarly finds that only 1 of them (14%, article 3 in the last paragraph) is in a peer-reviewed journal outside of the "race realist" echo chamber (i.e. MQ or one of the three openpsych.net journals mentioned above).

A more precise picture can be obtained by searching ProQuest for articles published in each journal. When this is done for MQ, one gets 90 results, similar to the 88 obtained using GS. But citation counts are much more discouraging here, which is to be expected since it's a more selective database. In fact, of the 90 results for MQ, only 6 of them (7%) are even cited once on ProQuest. Currently, when I add up citation counts for the 6 results I got most recently I get a total of 10 citations. 

I then try to figure out how many of those citations were in 2017. The answer is...5 (12, 3, 4, 5). This yields an IF of 5/6 = 0.833, which isn't encouraging (but not that terrible: Psychological Reports' IF is a bit lower.) 

What about the openpsych journals? For ODP there are 7 citable articles in the period of interest, the most cited of which is, no surprise, the one about OKCupid data. Of its 11 citations, 3 of them are outside the echo chamber (1, 2, 3). The other 6 articles weren't cited at all in legit sources, yielding an IF of 3/7 = 0.429.

Neither of the IFs obtained here are very impressive, certainly, which doesn't bode well for the journals' credibility. This is obviously truer for MQ (which is over 50 yrs. old) than for ODP (which is only a few years old).