13

Evie Magazine reports on a 2020 Pew Research study, in an article titled Over 50% Of Liberal, White Women Under 30 Have A Mental Health Issue. Are We Worried Yet?

Conservatives label younger liberal generations as snowflakes or as having Trump derangement syndrome if they didn't like the past president; liberals and progressives label right-leaning individuals or conservatives as racists, bigots, misogynists, etc. Resorting to this type of lowbrow behavior might once have been seen as an excuse not to address the actual issues or beliefs at hand, but now ad hominem attacks are more common than not.

But what if what was once a cheap shot or a personal insult has actually been found to bear scientific correlation between the individuals who hold progressive ideologies and an increased risk of mental illness? That’s exactly what Pew Research has found — and all politics aside, the shocking diagnosis of over 50% of liberal women with some form of mental health medical diagnosis is a public health concern that no one seems to be discussing, let alone taking seriously.

Is there a correlation with being a young, liberal woman and mental illness diagnoses?

Oddthinking
  • 140,378
  • 46
  • 548
  • 638
Lambert macuse
  • 503
  • 3
  • 11
  • 3
    Stop commenting on this question unless you are trying to improve it. [We don't care](https://skeptics.meta.stackexchange.com/q/3858/11643), and you can jabber all you want in [chat](https://chat.stackexchange.com/?tab=site&host=skeptics.stackexchange.com). –  Aug 17 '21 at 05:49

1 Answers1

25

Not based on this evidence.

Your first source claims "A 2020 Pew Research study reveals that over half of white, liberal women have been diagnosed with a mental health condition at some point." The author bases this not a Pew document itself, but on a Twitter thread by political science graduate student Zach Goldberg, who describes himself as a "Wokeness Studies scholar".

The source in question is a Pew Research Center American Trends Panel (ATP) survey conducted March 19 to March 24, 2020, in the early weeks of the covid-19 pandemic "to understand how Americans are continuing to respond to the coronavirus outbreak." This isn't a random telephone survey, but a Pew effort to track changes among a generally static, broadly representative group of participants. Each year the participants are sent survey questions; this year 11,537 people from the pool responded via web surveys or the wireless tablets they were provided.

Several questions asked directly about the developing pandemic, such as how closely the respondent had been following the news, whether they had prayed about the outbreak, and what they thought the government should do in response.

The last question was "Has a doctor or other healthcare provider EVER told you that you have a mental health condition?" According to Pew, 18% of respondents indicated "yes."

From the associated data you can slice and dice the results to identify how various population groups answered the question: Men vs. women, rich vs. poor, and, indeed, conservative vs. liberal. In his number-crunching, Zach Goldberg mostly focuses on white liberals vs. white conservatives, though he points to a large gap between liberal women 18-29 and liberal men 18-29. (Though overall numbers are lower, there are similar gaps between white conservative women and white conservative men.) graph of results by age, gender, and political leaning

However, what this doesn't do is establish "mental illness" among the respondents. It establishes whether they're willing to say they were told by a healthcare provider they had a mental health condition" -- however the respondent interprets "healthcare provider" and "mental health condition." It doesn't say "diagnosed." It can be affected by whether the respondent has access to health care, is willing or unwilling to seek counseling, or does or doesn't have someone willing to talk to them about depression, an eating disorder, a substance abuse disorder, PTSD, or any of a variety of other issues. These are also factors that can vary by whether someone lives in a rural area versus a city.

Goldberg, whose Twitter thread sparked this discussion, acknowledges this: " "It's possible that the disparities in self-reported diagnosis are simply or partly a function of white liberals being more likely to seek mental health evaluations. I don't have the data to answer this question." He later notes "...this is a question that's underexplored and which may shed light on attitudinal differences towards various social policies."

Another element is that this survey result, even if we accept it as telling us something about the circumstances of women based on their political leanings, doesn't tell us anything about causality or currency. The question doesn't ask, and the answer doesn't tell us, whether the respondent was currently experiencing a mental health issue, just whether they had ever been told they had a "mental health condition."

We also don't know how the respondents interpreted "mental health condition" as applying to themselves; the Diagnostic and Statistical Manual of Mental Disorders (DSM–5) counts not only depression and schizophrenia as conditions, but also autism, stuttering, and attention deficit hyperactivity disorder.

jeffronicus
  • 5,412
  • 3
  • 28
  • 27
  • 6
    +1 for the Skeptics-style analysis. But I have a feeling that those who were hoping for a broad take on the topic (and who amply commented in that direction under the Q) are going to be the disappointed by your answer. – Fizz Aug 14 '21 at 01:07
  • 2
    @Fizz, true, but I don’t envision anyone doing a conclusive study of diagnosed and undiagnosed mental health and political beliefs, especially given the stigma associated with counseling in some quarters. – jeffronicus Aug 14 '21 at 01:17
  • 1
    For a catch paper title in regard to the penultimate para in your answer, ["Conservatives report, but liberals display, greater happiness"](https://pubmed.ncbi.nlm.nih.gov/25766233/), which in some sense speaks of core [measurement] issues here... – Fizz Aug 15 '21 at 02:47
  • 3
    @Fizz There is a whole rabbit hole to fall down there as researchers try to tease results out of self-reported measurements of "liberal" and "conservative" and "happy" and "unhappy." Once you see there are no objective measurements involved the results start to seem like noise to someone immersed in the physical sciences; I can measure the conductivity of seawater objectively, but how do you meaningfully reduce the complexity of human ideology across Western nations to a self-reported scale of "1=left, 10=right?" – jeffronicus Aug 15 '21 at 16:12
  • Since you've answered No after looking at the numbers, can you please try and estimate the likelihood that "young liberal women are NOT more likely to have a mental health condition" based on the statistics and the confidence intervals? – Ark-kun Aug 16 '21 at 06:28
  • @Ark-kun It's not really a question of "the numbers" in this survey, it's that this survey doesn't collect the right information to address the question. You'd need something comprehensive like the Nurses' Health Study at https://nurseshealthstudy.org/ – jeffronicus Aug 16 '21 at 14:49
  • I'll reframe my question a bit: Given that the ratio of people who answered "yes" is: Group A: 56%+/-12% (44%..68%) and Group B: 27%+/-17% (10%..44%), what is the probability that after a medical examination the ratio of positive cases in group A is not higher than in group B. – Ark-kun Aug 16 '21 at 18:55
  • 4
    @Ark-kun Again, it doesn't particularly matter what the numbers are; you can't meaningfully extrapolate anything to a population when you have a single self-reported measure that's poorly defined as opposed to, for example, knowing the results of clinical examinations for all participants. There are too many confounding factors, as noted in the answer. – jeffronicus Aug 16 '21 at 19:14
  • 1
    So young white liberal women are far more likely to be told by a doctor that they have a mental health condition, but that is disconnected from whether they actually have a mental illness? I imagine there's some validity there, but it's also a stretch to think there's no correlation. – Ask About Monica Aug 16 '21 at 21:56
  • @AskAboutMonica: This answer explains that the provided poll does not provide evidence. If you have other evidence, please post an answer. Otherwise, this comment is just your opinion. – Oddthinking Aug 17 '21 at 12:52
  • I'm just trying to summarize the essentials of the answer. To put it another way: If the question was "Are young liberal women more likely to SELF-REPORT a mental health condition?" it seems like the answer would be a clear yes. The reason it's a no is because the study doesn't provide evidence tying self-reporting to actual conditions. I think there's probably some correlation there, but the study doesn't speak to that. – Ask About Monica Aug 17 '21 at 21:13
  • I think it’s worth adding that the *entire* study is of ~11k people, of which the white 19-29 population is an even smaller subset. We also know historically people who respond to voluntary polls aren’t often representative of the entire population. So all you can really say (imo) is “in a *smaller subset* of a subset of roughly 0.0034% of the population, white liberal women are more likely to self report mental health issues”. I know studies are meant to extrapolate from evidence but… I think you’d need 11k respondents in every major city, and ~2k in numerous rural areas to make it worthwhile – TCooper Aug 18 '21 at 13:04
  • "However, what this doesn't do is establish "mental illness" among the respondents." Well, that's not what the OP asked about... – tuskiomi Aug 30 '21 at 14:42
  • @tuskiomi The question was substantially edited after this answer was posted. The details still apply, though; there is no substantial data to support the claim unless you add so many caveats that the claim becomes meaningless. – jeffronicus Aug 30 '21 at 15:12