Check it out!
”(…) Error Management Theory suggests that, in your inference, you can make a “Type I” error of false positive or “Type II” error of false negative, and these two types of error carry vastly different consequences and costs. The cost of a false-positive error is that you become paranoid. You are always looking around and behind your back for predators and enemies that don’t exist. The cost of a false-negative error is that you are dead, being killed by a predator or an enemy when you least expect them. Obviously, it’s better to be paranoid than dead, so evolution should have designed a mind that overinfers personal, animate, and intentional forces even when none exist.
Different theorists call this innate human tendency to commit false-positive errors rather than false-negative errors (and as a consequence be a bit paranoid) “animistic bias” or “the agency-detector mechanism.” These theorists argue that the evolutionary origins of religious beliefs in supernatural forces may have come from such an innate cognitive bias to commit false-positive errors rather than false-negative errors, and thus overinfer personal, intentional, and animate forces behind otherwise perfectly natural phenomena. (…)
In this view, religiosity (the human capacity for belief in supernatural beings) is not an evolved tendency per se; after all, religion in itself is not adaptive. It is instead a byproduct of animistic bias or the agency-detector mechanism, the tendency to be paranoid, which is adaptive because it can save your life. Humans did not evolve to be religious; they evolved to be paranoid. And humans are religious because they are paranoid. (…)”
— Satoshi Kanazawa, Why do we believe in God?, Psychology Today, March 28, 2008. (More). See also: Martie G. Haselton and David M. Buss, Error Management Theory: A New Perspective on Biases in Cross-Sex Mind Reading, University of Texas at Austin (pdf)
The invited talks at ICWSM were especially good this year. I want to highlight a few points from Duncan Watt’s talk and Jon Kleinberg’s talks.
- Social influence makes the selection for success less predictable. In other words, judged against independent measures of quality, if an audience is influenced by knowledge of community behaviour, it will select or promote with less correlation to quality than you would think. You may think ‘so much for the wisdom of the crowds’ but, of course, WOC is all about aggregating over independent judgments, not socially influenced ones – see Experimental Studies of Inequality and Unpredictability in an Artificial Cultural Market.
- We know less about our friends than we think we do. In the Friend Sense experiment, it was demonstrated that we project our opinions onto friends about whom we make assumptions regarding political beliefs. Watt’s concerns about the misrepresentation of polarization might be contrasted with the experiments reported in Nick Carr’s book The Big Switch in which a) small preferences lead to deep segregation and b) homophilly leads to extremism.
- Diffusion of information may ‘long circuit’ the small worlds of social networks. In Kleinberg’s presentation regarding the study of the largest internet chain mail (a petition) he described the role of the threshold model of diffusion in which we require multiple receipts of a stimulus (e.g. a chain mail letter) to pass it on, we are more sensitive to our immediate community – our strong links – than to small-world building weak links. This seems to have some relationship with Watt’s work on Challenging the Influentials Hypothesis and both his criticism of the disease analogy and his focus on the importance of the network structure, not some magical power of the ‘influential’.
From the blog of the “Freakonomics” authors…
Tall women are the smartest
“Once height (measured in inches) is controlled, women have significantly higher IQs than men. Net of height, women score 2.14 points higher on the PPVT. In contrast, each inch in height is worth more than half an IQ point (0.56). A comparison of standardized coefficients shows that the effect of height is more than twice as large as that of sex. Because American men on average are 5 inches taller than American women (5’10” vs. 5’5”), this translates into 2.80 IQ points, overcoming the 2.14-point advantage of wom- en and making men appear more intelligent when height is not controlled.”
They also controlled for sex, race, age, physical attractiveness and health.
As a relatively short man, this is discouraging news.
Here is the complete study.
This is what happens when smart people get some data, run a regression, and try to interpret… overgeneralized, ridiculous conclusions that don’t translate to life in any practical, useful ways. Not that I’m bitter or anything.
As psychology students past and present will be only too aware, statistics are a key part of every psychology undergrad course and they also appear in nearly every published journal article. And yet have we ever stopped to recognise the statisticians who have brought us these wonderful mathematical tools? As psychologistDaniel Wright puts it: “Statistical techniques are often taught as if they were brought down from some statistical mount only to magically appear in [the software package] SPSS.”
To help address this oversight, Wright has compiled a list of ten statisticians he thinks every psychologist should know about. The list is strict in the sense that it only includes statisticians, whilst omitting psychologists, such as Jacob Cohen andLee Cronbach, who have made significant contributions to statistical science in psychology.
Wright divides his list in three, beginning with three founding fathers of modern statistics. First up is Karl Pearson (pictured), best known to psychologists for the Pearson Correlation and Pearson’s chi-square test. He was a socialist who turned down a knighthood in 1935. His first momentous achievement was his 1932 book The Grammar of Science and he also founded the world’s first university statistics department at UCL in 1911.
Ronald Fisher was the author of Statistical Methods for Research Workers, which Wright describes as “one of the most important books of science.” Fisher was also instrumental in the development of p values in null hypothesis significance testing.
Together with Pearson’s son, Egon, Jerzy Neyman produced the framework of null and alternative hypothesis testing that dominates stats to this day. He also created the notion of confidence intervals. Neyman and Fisher were big critics of each other’s theories. After a brief spell at UCL with Fisher, Neyman moved later to Berkeley where he set up the stats department - now one of the top such departments in the world.
Wright also lists three of his statistical heroes: John Tukey of post-hoc test fame, who made major contributions in robust methods and graphing (and who coined the terms ANOVA, software and bit); Donald Rubin who has conducted influential work on effect sizes and meta-analyses; and Brad Efron who developed the computer-intensive bootstrap resampling technique.
Wright devotes the last section of his list to four statisticians who have gifted psychology particular statistical techniques: David Cox and the Box-Cox transformation; Leo Goodman and categorical data analysis; John Nelder and the Generalised Linear Model; and Robert Tibshirani and the lasso data reduction technique.
“The list is meant to introduce some of the main statistical pioneers and their important achievements in psychology,” Wright concludes. “It is hoped learning about the people behind the statistical procedures will make the procedures seem more humane than many psychologists perceive them to be.”
What do you think of Wright’s list? Is there anyone he’s overlooked?
Daniel B Wright (2009). Ten Statisticians and Their Impacts for Psychologists. Perspectives on Psychological Science, 4 (6), 587-597. [Draft pdf via author website].