By night we play music with townies. But by day we anesthetize, philosophize, and preach in attempt to numb the pain of surgical knives (Nate), market failure (John), anxiety (Adam, albeit loosely), and existential malaise (me, even more loosely). My anesthetic is academic research. In this post, I discuss the articles that have most transformed the way I think, write, and see the world.
Time is scarce, so I limited myself to one article per decade. If you haven’t read these articles, I encourage you to do so. If you have different picks, add them in the comments.
1970s
Year: 1974
Authors: Amos Tversky and Daniel Kahneman
Title: “Judgment under Uncertainty: Heuristics and Biases”
Source: Science, 185 (4157), 1124-1131.
Comparable album: Sgt Pepper’s Lonely Hearts Club Band (The Beatles)
To call Tversky and Kahneman the Beatles of judgment and decision-making research would undersell their contribution. “Judgment Under Uncertainty” was Sgt. Pepper if it had come out in 1947. T&K cite no previous research in their introduction—not because they were poorly read, because probability-Presley, bias-Berry, and heuristic-Lee Lewis had yet to shake, rattle and reveal the ways that biases distort our judgments and decisions.
T&K discuss three psychological ticks that tack our judgments into squalls.
Tick 1 is the “representativeness heuristic,” wherein we guess how likely an event is by judging how similar it is to a prototypical outcome. Imagine hearing an unfamiliar country song about trains and trucks and pickup trucks and mama and gettin’ drunk. You might think the singer is from Texas because the song sounds Texan. The representativeness heuristic often works. The problem is that we overlook other useful information. In this case, we might ignore the fact that most country singers are from somewhere other than Texas. In this case, the singer is actually from Akron, Ohio.
Tick 2 is the “availability heuristic,” wherein we judge how likely an event is by assessing how easily it comes to mind. You would [correctly] guess that there are fewer top 40 songs that use piccolo than use guitar because it is easier to think of songs with guitar. The problem is that sometimes it is easier to imagine less common events (dying in an airplane crash) than more common events (dying in a car crash). People fear flying but feel at ease in a Ford—or even a Harley—because we hear more about people who die in planes (Buddy Holly, Patsy Cline, Lynyrd Skynyrd) than on the road (Bessie Smith, Duane Allman, Marc Bolan). The media exacerbate this bias, as one of my favorite scholars explains: “The press is an availability machine. It serves up anecdotes which feed our impression of what is common in a way that is guaranteed to mislead.”
Tick 3 is called “anchoring,” wherein we make judgments by adjusting our estimate from an initial starting point. The problem is that this starting point can be arbitrary, and we usually don’t adjust enough. To judge the length of “Freebird,” you might start by assuming that most songs are about 3 minutes long and adjust your estimate upwards because you know that Freebird is epic. For instance, you might guess that Freebird is 6 or 7 minutes rather than the 9 [glorious] minutes for which it actually shreds. On the other hand, if you were to guess the length of Freebird after learning that Pink Floyd’s “Shine on You Crazy Diamond” is 26 minutes long, you would instead adjust downwards from 26 and likely arrive at a longer estimate, say 12 or 15 minutes.
The moral of “Judgments under Uncertainty” is that we aren’t nearly as rationale as we, or economists, like to think. T&K remind us to, in the words of Professor Kendrick, “be humble” (Lamar 2017).
1980s
Year: 1986
Author: Grant McCracken
Title: “Culture and consumption: A theoretical account of the structure and movement of the cultural meaning of consumer goods”
Source: Journal of Consumer Research, 13(1), 71-84.
Comparable album: Workingman’s Dead (Grateful Dead)
Psychedelic rock and the field that is now known as consumer culture theory (CCT) have much in common. Both are ambitious, break boundaries, have a cult following, and are misunderstood by those that haven’t tasted the Kool-Aid.
Workingman’s Dead and “Culture and Consumption” occupy a similar place within their respective spheres. Workingman’s Dead blazed less trail than Surrealistic Pillow or the acid test concerts, just as “Culture and Consumption” cut less edge than “The Experiential Aspects of Consumption” or the Consumer Behavior Odyssey. Yet, few psychedelic rock albums are as palatable to mainstream listeners as Workingman’s Dead, just as few CCT articles are as palatable to mainstream marketing scholars as McCracken’s article. If most CCT papers are a 22 minute drum solo, “Culture and Consumption” is “Casey Jones”: four-and-a-half minutes of catchy countercultural country-folk-pop-rock perfection. Importantly, both the albums and the article opened my ears, eyes, and mind to a richer palette of music and research possibilities.
“Culture and Consumption” starts with the premise that products have meaning. No one buys a 24-carat emerald-cut diamond or a leopard skin pillbox hat for their functionality. By the 80s, the idea that products have meaning had reached legal drinking age (Levy 1959), but continued to be a wallflower trampled in a field that thought of consumers as “information processors” and “utility maximizers.”
McCracken argued that cultural meaning moves in a semiotic samba orchestrated by advertisers, journalists, product designers, influencers, and consumer subcultures.
We assign meaning to phenomena such as time (second, day, year, now, later), space (here, there, up, down), nature (trees, sheep, wheat, rock), and people (old, young, boy, girl) to make sense of the world. Products acquire meaning from the people and objects paired with the product in ads and on packaging, runways, sidewalks, boardrooms, and grocery aisles.
For example, the Velvet Underground became “avant-garde” in the 60s by using the pop art of Andy Warhol, featuring discordant string solos, and singing about drugs, prostitution, and sadomasochism. Similarly, the Red Hot Chili Peppers became cool in the 1980s by performing at underground punk clubs and mingling with Andy Gill (Gang of Four), George Clinton (Parliament, Funkadelic), and Maceo Parker.
Meaning also transfers from products to the people who use them. People wear Velvet Underground shirts to show that they are avant-garde, and they brag about seeing the Chili Peppers open for Gary and Neighbor’s Voices at the Rhythm Lounge to try to be cool.
Importantly, meaning changes. The Chili Peppers became less cool when middle managers and minivan moms started listening to them. When meaning changes for the worst, people flee. The Chicks and Lady A. changed their names when “Dixie” and “Antebellum” began to evoke white supremacy more than southern tradition, and liberals stopped listening to Kanye West when Ye wore a MAGA hat.
“Culture and Consumption” was one of the articles that made me want to become an academic researcher. It inspired my articles on coolness, ironic consumption, and why an endorsement from Jessica Simpson is more likely to make a product ditsy than sexy.
1990s
Year: 1998
Author: Thomas C. Veatch
Title: “A Theory of Humor”
Source: Humor, 11(2), 161-215
Comparable Album: Boys and Girls in America (The Hold Steady)
Neither Boys and girls in America nor “A Theory of Humor” have had as wide an influence as they deserve, but they revolutionized my musical and professional lives. Before hearing the Hold Steady, I sang about drifters and levees and cowboys. Now I sing about townies and the boys and girls with whom they party. Before reading Veatch, I studied memory and meaning and whether Jessica Simpson makes products sexy. Now I study why having sexual intercourse with a dead chicken is funny.
Before Veatch, the leading theory was that people laugh when they perceive an “incongruity.” Incongruity is a squishy concept, but it usually refers to something unexpected—like a cloud that farts pistachios—or a juxtaposition of two things that don’t normally go together—like headphones and donuts.
The problem is that many incongruities are not funny. Surprises can be awesome (winning the lottery) or awful (toenail cancer), just as juxtapositions can be tragic (getting mugged at a yacht club) or intriguing (sampling a gentle Dido chorus over a rap about a homicidal fan).
Veatch argues that funny is a function of N + V, where N indicates that people think the situation is “normal”, and V indicates that they also think there is a “violation”. Veatch notes that N and V are incongruous, but most incongruities are either not normal (cancer, yacht club muggings) or not a violation (winning lotteries, sampling Dido). But when a joke or prank or tickle seems both violating (it’s a tickle attack!) and normal (it doesn’t hurt), laughter ensues.
“A Theory of Humor” was a conceptual wallet lying on the academic pavement. My co-author, Pete, and I found the wallet and pocketed the “violation” idea—though we left “normal” on the sidewalk. Funny isn’t normal, we argue, but it is benign. We found that participants would chortle when they read about a man who snorts his dead father’s ashes or a woman who sells her virginity on eBay, but not because they think these behaviors are normal. Rather, they are amused when they think of these hypothetical transgressions as benign. For instance, more people laugh at a story about a man named Matthew who rubs his genitals against a pet kitten when “the kitten purrs and seems to enjoy the contact” than when “the kitten hisses and does not seem to enjoy the contact.” Distance helps too. It would seem less benign, and be less funny, if Matthew did this to your kitten.
2000s
Year: 2006
Authors: Matthew J. Salganik, Peter Sheridan Dodds, and Duncan J. Watts
Title: "Experimental study of inequality and unpredictability in an artificial cultural market"
Source: Science 311(5762), 854-856.
Comparable Album: Are You Experienced? (The Jimi Hendrix Experience, North American edition)
It was hard to find an album that compares to “Experimental Study of Inequality...” Like the article, the Ramones debut was a succinct and scathing critique of overconfidence and self-aggrandizing bravado. But the article was more like Dummy (Portishead 1994) in that it was technically innovative and distinct from what came both before and after it. I also thought of Duke Ellington’s big band recordings because the article used a massive experiment to test new research questions, much like the Duke et al. used a massive orchestra to create new sounds.
In the end, I went with Are You Experienced. Songs like “Purple Haze” and “Fire” bring the heat in under three minutes, just as Salganik et al. bring the data in under three pages. Jimi’s musical experiments and technical guitar work were ahead of their time, just as Salganik et al.’s market experiments and technical data analysis were ahead of theirs. Finally, listening to “Are You Experienced” made me question the spiritual world, just as reading “Experimental Study of Inequality” made me question the social world.
People think that hit songs become hits because they are better or catchier or more relatable or something else about the song itself. Salganik et al. reveal that popularity is mostly a matter of luck, alongside a smattering of social influence. They ran an online experiment in which 14,341 music fans listened to and downloaded songs. The experiment randomly assigned listeners into one of nine different worlds. The worlds started the same. Listeners chose which of 48 unfamiliar songs to download. In eight of the worlds, listeners could see how often each song had been downloaded by other listeners who had been randomly assigned to the same world. The songs started with zero downloads, but the number of downloads accumulated independently in each of these eight worlds. The ninth world did not reveal download history. The experiment lets us see whether the same songs become popular in each world, which would occur if songs become hits because they are better, or if different songs become popular in different worlds, which would show that it’s a total crapshoot.
The results show that it’s mostly a crapshoot. “Lockdown,” by Metro, for example, was the most downloaded song in one of the worlds but finished 40th (out of 48) in another. “She Said,” by Parker Theory, finished tenth in one world, but first in another. It was impossible to predict which songs would be most downloaded. Moreover, it was more difficult to predict success in the worlds where listeners could see download history because small, random differences in which songs were first downloaded would cascade into large differences over time. For example, the tenth listener in world one might see that “Lockdown” had already been downloaded twice and decide to join the crowd, whereas the tenth listener in world two would see that “Lockdown” had not been downloaded and pick something else instead.
Outside of extraordinary sociology experiments, we don’t have the luxury of comparing different, randomly assigned worlds. History deals just one hand. If we happen to be dealt the hand in which “Lockdown” becomes popular, we argue that “Lockdown” was a hit because it is catchy or has a great beat or listeners can relate to it. If dealt a different hand, we invent a story about why “She Said” was a hit. As Watts, one of the study’s authors, writes: “We deceive ourselves into thinking we can make predictions about things that are impossible, in principle… You could know everything about individuals in a given population—their likes, dislikes, experiences, attitudes, beliefs, hopes, and dreams—and still not be able to predict much about their collective behavior.”
If Tversky and Kahneman’s research is a holler for intellectual humility, Salganik and colleagues research is the Chrysler Air Raid Siren.
2010s
Year: 2011
Authors: Joseph P. Simmons, Leif D. Nelson, and Uri Simonsohn
Title: "False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant"
Source: Psychological Science, 22(11), 1359-1366.
Comparable Album: Nevermind (Nirvana)
Nevermind was a reaction to the produced, glossy, market-oriented hair metal that passed for rock and roll in the 80s. “False-positive psychology” was a reaction to the surprising, cute, unreplicable studies that passed for science in the 00s. Nirvana may not have cleansed cock-rock from MTV (see “November Rain”), and Simmons et al. may not have cleansed p-hacking from JCR (see O’Donnell et al. 2021), but both declared the dealer was selling oregano and demanded real herb.
Simmons, Nelson, and Simonsohn show how researchers alchemize junk data into statistical gold. When running a study, researchers need to decide how many respondents to recruit and which respondents, measures, and analyses to use. Because they want to publish their results, and journals are unlikely to publish a study that finds “there’s nothing up with acai berries,” researchers choose, and report, analyses that yield significant results while explaining away those that don’t.
Say that you want to test whether listening to Limp Bizkit is mind numbing. So, you create an experiment in which one dollop of undergraduate saplings listen to “Nookie” (the treatment condition) while others listen to either Radiohead (a control condition) or the sound of paint drying (another control condition). After listening, the sampled saplings solve algebra problems and write a haiku. You’re not made of Dogecoin, so you initially recruit just 30 people. But the data are difficult to interpret. The people who listened to paint drying answered more of the math questions correctly, but the difference is not significant. It’s impossible to know if the difference is real, so you recruit another 30 people to see what happens with a larger sample. Then you realize that 10 of the participants are professional mathletes, and this might contaminate your sample, so you throw their data in the e-trash. Now the study worked! People solved more algebra after listening to paint than after listening to Limp Bizkit. You report “Limp Biskit is more mind numbing than listening to paint dry (p < .05)” but don’t mention that you doubled the original sample size, dropped the mathletes, asked participants to write haikus, nor that the Radiohead listeners did the maths as poorly as Bizkit listers. Simmons and colleagues call this p-hacking and show how researchers can p-hack random data to statistical significance in over 60% of experiments.
Like Nevermind, “False-Positive Psychology” is remembered for its angst, but is also a masterpiece in storytelling. Most method papers plunge directly into mathematical proofs or statistical simulations that make paint drying a stimulating comparison. In contrast, Simmons and colleagues illustrate the problem by creating two real experiments, one of which shows that listening to “When I’m 64” by the Beatles caused participants to be younger, on average, than listening to an instrumental song called “Kalimba” (p < .05). They next bring the stats to unravel how they hacked their way to the implausible yet statistically significant results.
On the radio, Nevermind replaced Poison, Bon Jovi, and Motley Crue with Nirvana, Pearl Jam, and Soundgarden practically overnight. The pace in academic journals is slower, but “False Positive Psychology” is gradually starting to spread better science. It has encouraged researchers to preregister their studies, publish their data, and avoid the most egregious forms of p-hacking. No paper has done more to change the way I run, analyze, and report my studies.