Welcome to The Shores of Academia
The Shores of Academia
The Shores of Academia will present analyses and commentary on issues in social sciences and education that are seldom if ever addressed elsewhere.
The focus in the near future will be the troubling rise of adolescent suicide and related matters. Many of my posts will be critiques of pertinent research and theories.
The Youth Suicide Rise
The doubling of adolescent suicide between 2007 and 2017 (The Rise) and the continued perseverance of high suicide rates is a tragedy that might be largely preventable if we better understood this phenomenon.
In 2019 I started a series of CDC data analyses in which I showed that:
- The Rise is not explainable by shifts in demographics or suicide methods but coincides with increases in the likelihood that adolescents were affected by adult suicides and fatal overdoses during childhood ('abandonment deaths').
- Teen and tween suicide rates are, at least since late 1980s, highly predictable from 7-year aggregate adult suicide rates, pointing to childhood trauma as potentially the primary explanatory factor of trends in adolescent suicide.
- Female share of adolescent suicide started to increase in late 1990s, long before The Rise, and so it is not necessary for The Rise to be explained by factors that affect girls more than they do boys.
- Male share of teen suicides tends to increase when male teen homicide increases, suggesting a violence factor that is affecting boys much more than girls.
These results remain relevant today, especially since the youth suicide rise is still widely considered a mystery or attributed to various other causes. I plan to update and deepen the analyses and perhaps also formulate a more general theory to account for depression and anxiety disorders among adolescents.
Adolescent Mental Health Research
When Jean Twenge, later joined by Jonathan Haidt, started to argue that the main culprits behind declines in the mental health of adolescents are smartphones and social media, it induced considerable opposition and criticism. I myself contributed -- see Blinded by Gender, The Crime of Parsimony, and The Perils of Improper Terminology -- and I will continue to critique their work.
Some of the responses to Twenge and Haidt from their colleagues, however, have also been severely flawed, especially when dealing with statistics. This is unsurprising given that statistics are not only poorly understood within social sciences, but also greatly misunderstood by some of those in psychology who think they do understand statistics.
For this reason I may occasionally end up both criticizing and defending Twenge & Haidt.
In fact there are persistent misconceptions about statistics that are widely shared by both proponents and opponents of the digital tech hypothesis within psychology -- in these cases I will be criticizing both the pro-tech and anti-tech camps together.
Junk Science
A number influential of papers regarding the mental health effects of technology amount to outright junk science: everything novel within these papers is either trivial, irrelevant, fuzzy or false.
I've so far avoided detailed critiques of junk science because the considerable time and effort one must invest does not contribute to getting closer to any underlying truth as there is nothing left to improve in junk science papers once we dispose of their rotten core.
Junk science, however, is starting to dominate the field of psychology concerned with technology effects. This is why I've decided to critique, in the near future, at least a few of the worst examples of the recent wave.
A Monstrosity of Bad Science
Among the most atrocious examples of recent junk science is The association between adolescent well-being and digital technology use published in 2019 by Nature Human Behavior (NHB).
This widely celebrated monstrosity begot one of the most asinine yet beloved memes in current psychology, namely that the effect of screen time / digital technology / social media (the meme varies) on the mental health of adolescents is no greater than that of eating potatoes or wearing glasses.
The paper is so packed with ludicrously flawed science that one might suspect it to be a hoax in the tradition of the Sokal affair -- alas the parody of science here is entirely unintentional.
Much of the paper is one giant tower of babble produced by taking 'magical' medians of mindless iterations of absurd combinations of indefensible mismeasurements -- typically rank correlations with dichotomous outcomes misrepresented as validated regression by the authors. The rest is a mixture of blatant bias and genuine incomprehension of elementary statistics and rudimentary science.
The authors declare associations of digital technology with adolescent mental health to be 'so minimal that they hold little practical value' -- even though, in all their three data sets, heavy users of digital tech consistently report up to double the risk of depressive symptoms and suicidal conduct (facts entirely withheld from the readers).
The methodology of the authors could be easily adapted to conclude that concerns about smoking are unwarranted because smoking cigarettes 'explains' at most a tiny portion of health and because it has a similar impact on well-being as eating potatoes or wearing glasses.
To comprehend the utter lack of quality controls at Nature: Human Behavior, note that even though the authors attenuate correlations with mental health by controlling for every inappropriate mediator available, they fail to control for age (or grade) and sex (or gender).
The list of serious problems with this paper is so long that just enumerating them amounts to a major task, especially since much of the methodological mayhem is revealed only once you look under the hood to see the strings of folly implementing this carnival of madness.
The Monstrosity Project
The authors of the NHB paper ignored the criticism I raised with them after its publication.
Even a severe programming mistake that I pointed out to the authors in 2019 -- the correlations code meant to select students who answered yes to at least one of the risk questions in a given subset is actually selecting only students who answered yes to all these questions -- remains uncorrected as of 2023.
I refrained from writing a public review of the NHB paper because to do so comprehensively would be a giant undertaking, especially if the criticism is to be written so as to be easily comprehended by psychologists, which is necessary if it is to have much of an impact -- at least when such critique is published by someone unknown.
Given the continued influence of the NHB paper, however, I am no longer willing to wait until some figure of sufficient authority -- ideally a renowned statistician -- finally addresses this leviathan of bad science.
That is why I will soon launch The Monstrosity Project on Substack -- an attempt not only to identify, in a series of articles, the most egregious errors and flaws and misconceptions of the NHB paper but also to explain these failings in a manner that ordinary psychologists can fully understand without having to constantly consult a statistics textbook.
Pseudoscience with a Postmodern Twist
Some of the recent junk science crosses into pseudoscience where the mimicry of science without understanding becomes a central characteristic, key conclusions are outright absurd, inconvenient evidence is meticulously censored, and deceptive methods of communication as well as insinuation are common.
The authors of pseudoscience usually immunize themselves to scientific discourse by simply ignoring or evading criticism and yet the most prominent purveyors of such pseudoscience promote themselves as advocates and model practitioners of scientific rigor and objectivity.
This strategy of positioning oneself as a defender of science while simultaneously undermining it is the kind of double coding irony utilized by postmodern demagogues and can be extremely detrimental to social sciences when propagated under the veil of (bogus) statistics expertise and protected by a Teflon scholarship bubble inflated by an informal alliance of similarly minded academicians.
I hope to find the time to expose some of this recent pseudoscience and critique not only its authors but also the journal editors who publish it.
The Integrity of Scientific Institutions
When the scientific integrity of entire fields -- such as the study of digital technology impacts on mental health -- is collapsing due to the disintegration of quality controls at prestigious journals, the proliferation of statistics misuse, and the creeping intrusion of ideologies, it is crucial for key scientific institutions to preserve high standards.
One such prominent institution is the Centers for Disease Control and Prevention (CDC), the administrators of the Youth Risk Behavior Surveillance System (YRBS), a biannual survey of U.S. high school students that supplies a wealth of data on adolescent behaviors and feelings.
Unfortunately, the recent misrepresentation of 2021 YRBS results by the CDC, accompanied by abhorrent misuse of statistics and lack of transparency, impinged efforts to improve adolescent mental health and damaged the reputation of the CDC.
The CDC also continues to imperil the usability of the indispensable YRBS by its senseless refusal to release vital information, such as what portion of YRBS questionnaires omitted items related to suicide and sexual violence.
I will continue to pursue these institutional matters, at least as long as they directly affect adolescent mental health research and policies.
Summary
If all goes well, The Shores of Academia will present analyses and criticism that will be valuable and rare, addressing important topics in social sciences (and later education).
The Shores of Academia, however, will not fit easily into the current world of ideological divides, be they in science or education or politics, and is likely to irritate rather than please its various camps.
As it should.