**** PLEASE NOTE **** This is NOT a political post, and political comments will NOT be tolerated.
One of the many problems associated with doing scientific studies on the psychology of people is that you have to ask them how they are feeling and what they are thinking. The problem, of course, is aptly pointed out by fictional character Dr. Gregory House: “Everbody lies.” How can scientists determine whether or not people are actually telling the truth in these studies? There are some techniques. For example, if the study is based on a survey, the survey can ask the same question in many different ways, and a model can compare a person’s answers to those slightly-different questions to see if there is a consistent pattern. However, I recently ran across a study that did something radically different, and I found it very interesting.
The scientists noted that many people said the results of the 2016 presidential election here in the U.S. caused them a great deal of stress and anxiety. As the authors of the study put it:
One therapist, Inger Burnett-Zeigler, wrote in Time, “In the weeks since the election, many of my patients have come to therapy with anxiety, fear, and worry…It’s obvious to me that this highly contested election is already having real mental health consequences.”…A full 72% of Democrats reported that the presidential election outcome was “a significant source of stress,” as compared to 26% of Republicans. (references removed for clarity)
The researchers wanted to see if this was really the case, so they decided to do something interesting: They studied how people searched the internet after the election. After all, in a survey, you are either responding to a person or filling out a form that you know a person will read. The fact that you know someone else is going to evaluate your responses might lead you to say things that aren’t really true, so as to look better to that person or to make a point. However, your internet searches are (supposedly, not really) private, so people might be more “honest” with Google (or in this case, Bing) than they are with people. As a result, if you really want to know how people are feeling, look at their internet searches.
With that in mind, the authors looked at the searches of over a million Bing users. They found 300,000 who had also answered a political survey on MSN prior to the election. From that survey, they identified the users as Republican/Trump voters or Democrat/Clinton voters. They then looked at the users’ general Bing search habits prior to and after the election. They found that neither Republican/Trump voters nor Democrat/Clinton voters increased their searches regarding mental health issues after the election. This made the authors conclude that neither side of the political isle actually experienced increases in stress or anxiety after the election. They just told people it was happening.
The authors coined a term for this: reverse cheerleading. It is the practice of lying about how you feel so as to signal your dislike for your opponents. The authors say:
This suggests that some Democrats reported mental health declines after Trump’s election as a form of reverse cheerleading, where partisans report evaluations that are more negative than their true beliefs to reflect badly on a president of the opposing party.
Now if you are wondering whether or not the methodology of the study is valid, they did find one interesting result. If they compared the searches of people who sometimes searched the internet in Spanish, they did find an effect. Those users were more likely to search Bing for mental-health-related issues after the election than before the election. As the authors say:
This finding suggests that while Democrats’ descriptions of mental distress after the election had an element of expressive reporting, the mental consequences of Trump’s ascendance were very real for Latinos.
Since many think that the President’s policies will be bad for Latinos, it makes sense that at least some Latinos would feel genuine stress or anxiety after the election. Thus, the authors’ method seems to be sensitive to at least some changes in psychological health.
So in the end, it seems that reverse cheerleading is real. As a result, you have to be skeptical of survey results and other voluntary reports about psychological issues.
Of course, those who are worried about privacy might be concerned that a study like this can actually be done. After all, most people think that no one will ever see their internet searches, especially if they clear their browser history. This study demonstrates otherwise.
In a separate study researchers found that internet users visiting http://blog.drwile.com sought out psychological resources directly after visiting that site.
: )
Indeed!
I have to sometimes. 😛
This is interesting; thanks for posting it, Dr. Wile.
Ignoring all the problems that arise with this kind of research (though I’m glad the sample size is large), I’m not at all surprised at the notion that the people claiming increased anxiety and stress probably weren’t truly affected by either. I always thought that particular news story — that Trump is damaging people’s mental health — existed primarily because (i) people were truly upset by Trump’s victory, and (ii) some of those people were journalists with actual mental health issues. Since our culture teaches us to conceptualize ourselves in terms of our political beliefs and encourages us to argue and worry about them even if we don’t actually know much, losing an election does bother people unreasonably. I had very wise and solid friends who considered disavowing me after the election; they reconsidered because they knew better, but most people don’t. And these are exactly the people who feed on the media narrative, so they’d connect their worry with the anxiety/stress story.
But what occurs to me now is that this story is really an abuse of the notion of serious anxiety and stress that does involve mental health. It’s common for people who don’t have mental health issues to abuse the language: they’ll say they or others are OCD, bipolar, or autistic in things that have little to no relation to what the actual disorders are like. I remember my fellow first-year grad students joking that I was autistic because I calculated the Clebsch-Gordan coefficients for spin-3/2 plus spin-1 and wrote down the 12 by 12 matrix representing the transformation between bases. This was not me being autistic; it was them not yet having a concept of how difficult and laborious physics is (and of course, Clebsch-Gordan coefficients are easy; I’m pretty sure the guy who said it actually went on to work with 3nj symbols, and those are intense). Since people are already flippant with mental health terms, it’s easy for them to virtue signal about their supposed anxiety.
That’s an excellent point, Jake. We do that with other things as well. I remember the first time I heard someone say that vaccinating a kid was “child abuse.” As the adoptive father of a girl who actually experienced child abuse, my hackles went up immediately.
I just realized you taught me a new word: hackles. And I meant to ask, have you ever done anything with 3j, 6j, or higher symbols? Or do heavy ions just have too many nucleons to bother with adding up angular momenta of all of them?
I also meant to clarify my reasoning: I think the primary reason the Trump-anxiety story exists is that some journalists from influential publications had friends (or were themselves) going to counseling whose feelings about the election upset them enough to make the connection to their mental health. And whether or not that truly added to their anxiety, their journalist friends saw it was a good story.
Wigner symbols aren’t useful for the kind of physics I do. Part of the issue is the fact that there are so many nucleons. As the number of nucleons increases, the nucleus has a lot of collective degrees of freedom that take away from the intrinsic character of the nucleons. Thus, the larger a nucleus gets, the less you can treat it as an ensemble of individuals.
Thanks; that’s what I figured.
I wouldn’t assume that if people search for mental health topics on the Internet, that it means that they are experiencing such difficulties. Another possibility for the Latinos was that they were unfamiliar with the mental health references after the election and they were searching it to learn about it. This could be either to determine if they fit the bill or simply to understand what claims people who they perceived to be on their side were actually making. I often search for things like this, even if I’m familiar with something, to see if I understand it correctly. I especially search for things I may disagree with in an effort to understand opposing positions.
That’s a good point.
I’m surprised that most people think that no one will ever see their internet searches. The fact that Google and Bing (and Facebook, twitter, instagram, etc) save and collect all the information put into it, should be common knowledge by now. That’s why I usually use the search engine Duck Duck Go.
Also, I’m surprised when I meet people that don’t know what they put in hot dogs. That should also be common knowledge by this point (hint: it’s not pieces of pork tenderloin)
To pick at the nits:
The photo you’ve cited isn’t of someone finding out who won the election…it’s a well-known photo of a person (protesting) who is reacting adversely to someone she disagrees with vocally being sworn into office as the president.
There are many photos of folks screaming at the election of the president from election night. Most of those are news personalities. The majority of folks were either quietly resigned, or openly despairing…but not screaming.
Good point. The caption was incorrect, so I corrected it.