This is an independently authored statement based on material presented by“Particles for Justice“. The authors and signatories of that statement have been notified of this response. Click here to read professor Strumia’s own response to his critics. Click here to read his (draft) paper about the data and analysis used in his talk.
“Particles for Justice“
On the 4th of October, a blog post titled "Particles for Justice" was published by a group of physicists (henceforth referred to as "the authors") condemning the views of professor Alessandro Strumia as "morally reprehensible". They claim that the presentation he gave at a CERN conference on "High Energy Theory and Gender" was incorrect in asserting, that the degree of sexism in physics is overestimated, that biological factors adequately explain the observed performance disparity between men and women in science. He further claimed that there exists a widespread, ideologically motivated push to hire people based on factors other than merit. Strumia is accused of "mining" and "misrepresenting" data to support his claims, and "belittling" the legitimacy of female scientists.
In this piece, I will review the claims and sources presented by the authors. My findings lead me to believe that their outrage is misplaced and unsupported by the data they themselves cite. Furthermore, I will attempt to analyze the philosophical underpinnings of contemporary equality movements and explain why their intellectual foundations make controversies like the present one inevitable. It is my hope that this piece will help physics remain (as it should be) a field where the nature of reality can be investigated rationally without the distracting influence of moral outrage, unjustified accusations of bias, and impetuous campaigns to fix what is not demonstrably broken.
For context, the exact reason for Strumia's suspension from CERN should be clarified. Several major news sources (BBC, The Guardian, The Independent) have reported or give the impression that his suspension was due to the supposedly sexist viewpoints presented in his talk. However, CERN's official press release on the controversy and correspondence from their press office make it clear, that Strumia's opinions were not the reason for his suspension. According to CERN, no genuinely discriminatory or sexist remarks violating their code of conduct were made. The concrete reason was that Strumia directly criticized the hiring decision made by a named individual.
The interesting thing is that, media attention and the indignation of the authors have primarily been directed at the immorality of Strumia's viewpoints - not the professional faux pas of including personal criticisms in his presentation. There are two ways of viewing this situation:
A) CERN is wrong to consider his viewpoints acceptable.
This would imply that expressing one of the following notions (in order of their apparent importance in Strumia's eyes) violates their code of conduct:
"Attempts to increase the number of women in science using non-merit based policies such as quotas is ideologically motivated, harmful and discriminatory".
"The degree of sexism and discrimination in science is generally overestimated".
"Sexual dimorphism (a well-documented phenomenon) in humans is a significant factor for explaining behavioral differences and performance disparities".
While certainly unconventional, there is nothing inherently discriminatory or malicious about holding these opinions.
B) CERN is right to consider his viewpoints acceptable.
This would imply that the decision of the authors to morally condemn Strumia for his viewpoints is unprofessional. One thing is to disagree with the conclusions a fellow researcher draws based on the available evidence - another is to paint him as malicious for these conclusions.
Review of sources
Moral condemnations aside, are Strumia's viewpoints correct/plausible or is he guilty of misrepresenting and manipulating scientific data as the authors claim? In this segment, the sources and articles cited by the authors to argue their case will be reviewed.
Like the following two sources, this article is cited as evidence that discrimination in physics is "present and well documented". However, the source in question solely contains statistics showing that a smaller percentage of physicists are women, but no evidence that this discrepancy must be due to discrimination.
Observing that the Nepalese excel at mountaineering at a higher rate than the Belgians is not evidence of discrimination. Nor does it deny the humanity of 0.15% of the world's population to suggest, that this difference may be primarily explained by evolutionary factors. Suppose that (all things being equal) the nationality of the applicant has an impact on their chance of getting hired for mountaineering expeditions. Is this an argument for compromising objective standards of competence (past demonstrated experience, physical endurance, cold resistance) and replacing them with quotas/targets to "get more Belgians people into mountaineering"? Or is it an argument for making the application process more rational and objective, i.e. by anonymizing applications?
According to a study from Harvard the latter strategy has successfully increased the number of qualified women in top US orchestras by making auditions "blind". In contrast, the former strategy arguably introduces more bias and arbitrariness into hiring - why should one aim for 20% Belgians instead of 30%? Should Dutch or French Belgians be given priority? Who gets to decide? In spite of these issues, the implementation of quotas is recommended by the source, The European Parliament Committee on Industry, Research and Energy, is practiced at several institutions (see side 18 of Strumia's talk), and was earnestly discussed at the CERN conference.
This source reviews two studies where "long, open-ended interviews" were conducted with 80 women at various US universities. Based on this sample, the paper draws the conclusion that women are indeed more or less subtly discriminated against by their peers. In the highlighted interviews, it is frequently noted how the interviewees feel inferior and alienated because of their backgrounds.
Leaving aside the question of whether subjective personal experiences constitute objective proof that others engage in discriminatory behavior, there might be a reason for these findings, which does not involve prejudice. On page 25 of his slideshow, Strumia argues that if societal discourse overestimates the degree of bias and sexism in science, it will scare away people, who would otherwise enjoy this career path. One can easily imagine, that anyone who has listened to exaggerated warnings and decides to study science in spite of them will assume that they are entering a hostile environment. This will make them more likely to mistrust peers who belong to a majority group and to interpret ordinary social misunderstandings as evidence of discrimination. In short, one could hypothesize, that media attention creates exaggerated awareness, which creates false positives in studies, which further justifies more attention and more studies.
Is this hypothesis plausible? Certainly the examples of Matt Taylor's shirt, Tim Hunt's self-deprecating joke and Strumia's own talk indicate that even spurious evidence of discrimination can quickly be broadcast to millions. Apart from particular cases, the next source presented by the authors seems to apply impractically broad definitions, which result in an overestimation of the problem.
This 300 page report, reviews available data on sexual harassment in academia. For context, Altmetric.com states that since June 2018 the report has been downloaded +16.000 times (12.000 times in the US alone), is cited by 76 mainstream news articles, and ranks #614 in fame out of 12 mio. publications tracked by the site. The influence of this work in academic circles and on the public at large is thus significant.
The bulk of the data stems from a survey conducted at 14 institutions in "The University of Texas System" regarding the rate of faculty/staff-on-student sexual harassment. Student-on-student data was excluded in the report due to ambiguity. The report argues that students making suggestive remarks to each other is less clearly unacceptable than if staff made the same remarks to students. In the report, various types of sexual harassment are categorized as:
A) "Sexist hostility" - remarks such as "what a 'pussy'", "This is a man's job!" (etc.).
B) "Crude behavior" - vulgar behavior (hand gestures etc.).
C) "Unwanted sexual attention" - repeated requests for dates/romance despite rejection.
D) "Sexual coercion" - favorable professional treatment depending on sexual favors.
Note that this categorization goes from ambiguous and less severe ( A) and B) : remarks, jokes and hand gestures) to clear and highly inappropriate ( C) and D) : attention in spite of rejection, sex for favors).
The authors cite this work as evidence that sexual harrassment discourages women from joining STEM fields. However, the data presented in the report (see for instance Figure 3-3 on page 60) shows an almost equal rate of "Sexist Hostility" in non-STEM fields compared to STEM fields.
A) and B) are most prevalent in Medicine - the field where women according to the report (p. 54) have been graduating as frequently as men for decades, and where women are overrepresented among faculty compared to other scientific fields. Note that the more severe (and more unambiguous) types of sexual harassment ( C) and D) ) occur at an approximately equal rate (less than 5%) across all fields.
Across the disciplines, women are more prone than men to experience and/or notice the two first categories ( A) and B) ) from faculty/staff. It could be that male staff is more prone to be juvenile and vulgar for reasons stemming from either nature or nurture. Whatever the case, the field of Medicine indicates that such actions are not a determining factor in women's choice of career.
The report (Appendix D, p. 281) further states that female students who had experienced sexual harassment felt "significantly" less safe and missed "significantly" more classes that those, who had not. Here, "significantly" does not refer to the absolute magnitude of the measured effect. Rather, it means that there is a statistically significant difference i.e. in class attendance between students who have/have not experienced harassment.
Interestingly, the report shows, that what it defines as sexual harassment has little absolute impact on class attendance, feeling of safety and mental health of affected students. The maximum difference reported between having experienced some or no sexual harassment is around 0.2 points on the scale used by the report. No explanation for this scale is given for figures D5-D7, but for D8-D9 it is a 5 step scale ranging from 0 (never) to 4 (frequently). Assuming they are all measured on the same scale, the impact of SH on students' well-being is at most around 0.2/5*100=4%.
Why does the data show that something as serious as sexual harassment by staff/faculty has a relatively small impact on the well-being of students? One explanation for this surprising result is that the definition of "sexual harassment" (as employed in the report) is too broad. A large number of students experienced common and harmless remarks, jokes, and hand gestures (A) and B) ), reporting no detrimental effects. A much smaller number of students experienced severe breeches of conduct (C) and D) ) and report a large, negative effect. When added together a small net negative effect is found. One could test this hypothesis by re-doing the analysis of negative effects including only students who have experienced the severe types.
Sexual harassment is a serious matter. For the sake of the victims, it is essential to not dilute the definition by including actions, which have no discernable negative impact. Overstating the scope of the problem might draw attention and resources away from where victims need them the most. The authors have cited this work as proof that sexual harassment is widespread in science. However, the report's own data does not unambiguously support this conclusion. Insofar as the present criticisms are valid, it demonstrates the existence of widely publicized studies painting a disproportionately bleak picture of the STEM fields based on broad definitions and ambiguously analyzed data.
This source is a repository of articles analyzing the degree of bias in various academic fields. The page is cited as evidence that unconscious bias occurs at an equal rate in both men and women. Since no particular study is pointed to by the authors, addressing the veracity of this exact claim is difficult. However, assuming that the observed differences between men and women documented by the repository are indeed due to unfair bias, this is (as mention in the "mountaineering example") at most an argument for enhancing the use of objective measures in hiring.
Strangely, the creators of the repository argue (see the section, "Structuring Equality") for selectively hiring underrepresented groups without considering the possible negative impact of such a policy. As previously mentioned, this is not an uncommon or controversial opinion in spite of its under-addressed potential impacts. For example, the authors dismiss Strumia's complaint over a less cited (female) scholar being hired instead of him as sour grapes - a dismissal which overlooks the forest for the trees. The crux of the matter is, that any institution which deliberately implements selective hiring or quotas, makes it reasonable to question the competence and reliability i.e. of female colleagues. If a university hired staff partially based on the credentials of the applicant, and partially on their stated shoe-size, it is natural to doubt anyone with boots larger than one's own. Such a policy could conceivably yield a result opposite of the advertised one - more prejudice and more mistrust (this exact effect is anecdotally reported by one scholar cited by the authors). In an age where collaboration in research is more important than ever, scientists need to be able to trust in each other's competence and professionalism.
For genuine, impactful improvements, it is crucial to do the right thing in the right way, and for the right reason. In line with other sources cited by the authors, the repository campaigns for quotas etc. without critically considering their drawbacks. The authors condemn Strumia for supposedly belittling female scientists with his complaint. Noticeably absent is any critical evaluation of the fundamental reason for his complaint - the very popular quota policy, whose broad implementation threatens to make objections like his commonplace and perfectly legitimate.
5) Patat F.
This source is cited by the authors as evidence of sexism in allocated telescope time among astronomers. The article itself states that a 50/50 composition of men and women in review panels would not produce a significant change in acceptance chance for female researchers:
It further states that their analysis does not provide evidence for bias in the selection process.
The biggest discernable influence on acceptance chance was career advancement (senior professors accepted more often than students). It may well be (as the paper points out) that more men tend to become senior professors. The important point is that the authors cite this article as evidence for a conclusion, which the article itself explicitly states cannot be drawn from the data. This would not be a problem if the authors offered some criticism or explanation of how the data in fact supports their viewpoints. In the absence of any such remarks, their citing of the work seems hasty and unthorough.
Summary of source review
The aforementioned sources do not provide a strong case for the arguments presented by the authors. They object to Strumia's talk for "mining" and "misrepresenting" data, but their own case suffers from comparable flaws. At best, the sources presented moderate Strumia's secondary point about the prevalence of bias in physics without addressing his main one. If one looks beneath the surface, the controversy surrounding Strumia is not about personal criticisms in presentations, whether he mines data more or less than his critics, or how skeptical one should be about citation numbers. None of this would have sparked international headlines or provoked a major reaction from anyone. Nor is it about him "demeaning female scientists" (because he did no such thing), or even the demonstrable degree of bias in review processes.
Fundamentally, the controversy is about Strumia correctly identifying the existence of ideological forces affecting the physics community. In his talk, Strumia points out a number of high profile cases, where scientists have lost their jobs and been publicly humiliated for saying or doing even fairly benign things. One of them, James Damore was fired from Google for producing a document reviewing the practical, harmful effects that Google's strategy for hiring more women in their staff might have on the company. He argues that since the majority of people in the tech industry are liberally/left-wing minded, their disposition towards compassion for those they perceive as weak can get out of hand if unchecked. This in turn creates an ideological atmosphere, where objections to specific policies or lines of reasoning justified by compassionate goals are taken as evidence of malice.
Damore's argument (insofar as this description of it is accurate) that political overrepresentation itself causes the delicate ideological atmosphere, has some merit (see the Asch Conformity Experiments). However, does it fully explain why the reaction to expressing disagreement is not simply limited to looks of incredulity or perhaps an occasional heated argument? Why does it frequently escalate to firings and public condemnations? In the next segment, I will attempt to explain the underlying logical structure of a mindset prevalent (although not exclusively so) among contemporary equality-advocates, why it is detrimental (especially to their own stated goals), and how it may explain many of the recent controversies within and outside of academia.
In his magnum opus, "The Open Society and its Enemies", philosopher of science, Karl Popper (famous for the falsification principle), identifies a pernicious philosophical strain running through the Western tradition. He attributes its popularization to the German philosopher Hegel, but whatever its true origins, its existence and prevalence stands to reason. Its basic tenet is that the way a person thinks is entirely determined by the group to which they belong. For example, the experience of growing up English fundamentally structures the customs, habits and worldview of the Englishman. Like the island of Robinson Crusoe, his nationality simultaneously keep him afloat in the world and restricts his (intellectual) mobility. In this philosophical school, it would be reasonable to say, that nationality not only determines the starting parameters of his thinking - it *is* his thinking, and one cannot be understood without the other.
Therefore, to understand why an Englishman holds a particular opinion, one must first investigate how his national background might compel him to hold it. If he expresses delight in eating toast, the interesting thing is not to study the tastiness, affordability or health benefits of this food. Rather, one should note that Great Britain is a major exporter of bread, and infer that he and his country have a national interest in him espousing this opinion (even if he is not conscious of it). To this notion, Hegel (according to Popper) added the belief that history is essentially an arena, where these "group-ideas" compete for dominance (occasionally devouring and "synthesizing" each other). Any argument for the objective merits of a given opinion, institution or practice can instantly be dismissed as an expression of group-prejudices. One could even argue that rational discussion itself is an invalid notion and purely product of groups, who have found it convenient to propagate this idea.
To see why such a mindset is detrimental, assume that the English and the French generally hold it, and that Great Britain has historically monopolized fishing spots in the English Channel. Any argument a Frenchman makes for sharing the fishing spots (i.e. because his family is hungry), or for trading fishing rights for some other commodity can instantly be dismissed as an expression of his national interest - more fish for the French. If an Englishman feels strongly attached to the fishing trade, which has defined his family's identity and secured its survival for generations, a Frenchman can always reject this sentiment as the result of Great Britain's national interest - keeping the fish for themselves. Since reasoned discussion is out of the question, only passions (and usually not constructive ones) remain for solving the conflict. To both parties a tempting resolution is to neutralize the other side.
The danger, which this line of tribalistic thinking poses, is not exclusive to cases, where the groups in question are nation-states. It could be argued that Karl Marx "turned the logic 90 degrees" and applied it to economical classes across borders. One could summarize his version of it as:
It should be stressed, that the problem runs much deeper than mere disagreements about specific opinions. For example, the policies favoured by an upper class conservative and a blue collar socialist respectively are obviously mutually exclusive. However if both sides have the attitude, that the other party can be persuaded by a strong argument or the evident success of a certain policy, disputes are likely to remain civil and even productive. The same is not true if the socialist maintains this stance, but the conservative (more or less consciously) believes, that one must have been born into the upper classes to fully appreciate why taxes should be low. In this case the disagreement is not only about specifics, but on whether the two sides are able to communicate at all.
Even if the socialist generally favours tax cuts, his arguments for cutting sales tax rather than income tax (as the "group-thinking" conservative advocates) could easily be dismissed as a result of his plain upbringing. Since he cannot hope to persuade his lowborn counterpart by argument (or even reach a suitable compromise), he might demand, that he takes intensive dinner-etiquette courses to truly understand what it means to be upper-class. In the long run, seeking to abolish institutions of working-class culture (pubs, football games and similar crude sources of entertainment) is a natural move for our conservative. Afterall, their responsibility for defining the mental landscape of the socialist is the true reason why he refuses to accept cutting income tax - not that the conservative's own arguments for this policy might be inadequately presented or elaborated.
Note that it is irrelevant whether income taxes are objectively too high in the specific case. If disagreements cannot be discussed rationally in general (because the socialist can always be instantly disqualified due to his malicious biases) good decision-making relies on the blind luck of the conservative being right from the start. His mindset obviously threatens the accomplishment of his own goals, because disagreements within the conservative ranks can only be a sign that some members are insufficiently purified of lower-class bias. Thus, personal purity (rather than objective skill) becomes the figure of merit.
Popper's analysis of tribalistic thought systems and the consequences of acting according to them allows one to explain the so-called "Horseshoe theory"; the observation that far-left and far-right political groups (in spite of their specific aims being diametrically opposed) are similar in their behaviour w.r.t. mob mentality, authoritarianism, and intolerance of criticism. These groups may advertise nominally noble goals ("help the poor", "love your country"), and their members may honestly believe in these slogans with every good intention imaginable. However, the tribalistic logic that defines both their ideologies renders positive sentiments entirely irrelevant. With the toolbox of reasoned discussion firmly closed, no progress through compromise can be made - for "progress" is synonymous with furthering the power of my group-idea at the expense of others.
Fundamentally incompatible worldviews
I contend that the pushback stemming from people, who otherwise identify as liberal and open-minded, against the various contemporary equality movements is not about veiled contempt for women or whoever else is being advocated for. Rather, it is a symptom that underlying and fundamentally incompatible worldviews are in conflict. The current conflict is between the tribalistic worldview outlined above and one, which could be labelled "rational". The hypothesis that tribalistic thinking shapes equality advocates is consistent with certain observations about their rhetoric and methods. A clear example is the frequent calls to scrutinize each other's backgrounds instead of analyzing the veracity/practicality of the presented ideas. Moreover, whole professions, hobbies or subcultures are routinely denounced as insufficiently equal and in need of restructuring. Examples include university fraternities, fantasy card games, secular activism, movies, comic books, video games, hiking and skiing, and board games. While it remains an underlying, unarticulated premise in most cases, the belief that group-background/interests entirely shapes one's thinking is sometimes directly expressed by influential activists ("meritocracy" is a myth propagated by the already advantaged - not a preferable ideal everyone should strive for).
Granted, all people occasionally engage in Bulverism (as C.S. Lewis dubbed the phenomenon). The central problem is that the use of this logical fallacy has been systematized, justified and amplified by various influential schools of thought descending from Hegel and Marx (according to the view presented by Popper). Its contemporary prevalence has been attributed to the popularity of certain thinkers (sometimes categorized as "Post-Modern" or "Neo-Marxist") like Michel Focault, Jean-Paul Satre and Jaques Derrida in Western universities. One can debate the accuracy of these labels as the mentioned schools and their thinkers disagree on a number of specific points. What they do seem to share with each other and with classical Marxism is a preoccupation with analyzing the background factors defining how people think. Examples include arguing that power controls and defines thought, rebranding Marxist thought or criticizing Western thinking for being "phallogocentric".
In short, ideologies which share a tribalistic approach to thinking are highly popular in Western intellectual circles. Their underlying mentality is transmitted to students, which shapes how they put their ideas into practice and how they tackle disagreement. Because the mindset is defined by its focus on exposing malicious hidden motives behind institutions or opposing thoughts (rather than seeking objective truth) it becomes natural to antagonize people and subcultures outside its orbit. This not only damages the specific individuals or institutions in question, but prevents the stated goals of the activists ("equality", "respect", "tolerance") from actually being achieved. Legitimate criticisms and useful but unorthodox ideas can always be dismissed as lingering group-prejudice, so either one must submit fully to the prevailing opinions or be labelled an irreparably biased outsider.
The other option
How can one characterize the opposing system of thought, with which the tribalistic one clashes? In this school, people (in spite of their prejudices, and if they exert themselves) are indeed capable of changing their mind in the face of superior evidence or a more logically coherent argument. It views hidden psychological reasons for holding an opinion as irrelevant compared to the concrete evidence for or against it. This mindset is clearly aligned with the one required for scientific progress, where knowledge is gained by putting ideas to the test through experiments - not by picking the brains of the author for hidden prejudices. It could be argued that the tremendous success of natural science is evidence, that we are not as constrained by prejudices as one might think; a large scale, international collaboration like CERN is possible because the nature of reality can be independently identified by people no matter their starting point.
It should further be stressed that tribalistic thought systems are inherently unscientific. According to the aforementioned Falsification Principle (perhaps more accurately: "Disprovability Principle") a scientific theory must make concrete predictions that allow experiments to prove it false. For example, Newtonian mechanics is scientific, because it makes highly specific claims about the behavior of light near massive objects, which upon inspection turn out to be false. The same cannot be said for astrology, whose claims are so nebulous, that proving them wrong is neigh impossible in practice. Tribalistic thought systems are arguably more unscientific than astrology. Not only are their claims about reality usually vague and difficult to test, but criticisms can always be deflected with reference to group-bias. Paradoxically, tribalistic thought systems are usually eager to boldly criticize everything under the sun, but themselves remain comfortably shaded from scrutiny.
The rational worldview is also well aligned with liberal democracy, where disputes are settled by discussion and compromise. This might explain why self-identified liberals take issue with various equality movements - one cannot hold reasonable, liberal attitudes and simultaneously support groups premised on the idea that certain people are incapable of reasoning due to their group-membership. Because the ability to engage in rational discussion and experiment to settle disagreement is essential to science as well as liberal democracy, it is natural that the presence of tribalistic mindsets should cause tension. Their prevalence and their internal structure help explain why so many recent controversies involve their adherents. Fortunately, nothing fundamentally prevents these adherents from observing the problems with this family of ideas and changing their minds accordingly.
With any luck, the lengthy analysis presented here has managed to shed light on the Strumia controversy itself as well as the bigger picture surrounding it. The central thesis is that logically fallacious, tribalistic ideologies and the actions of people following their principles are to blame for recent public and academic controversies. They clash with principles that are indispensable to good science, and the functioning of a free, just and democratic society. Moreover, critical looks should be cast on any movement based on these ideologies, as they are not likely to deliver on any of the noble goals they advertise.
In spite of the many criticisms presented in this post and the decision to defend a controversial figure, I hope that those who disagree will demonstrate good faith in their response, and resist any temptation to react with tribalistic tactics. Those who agree are welcome to sign here and join in support of a man, who correctly (if imperfectly) spoke out against unscientific ideologies affecting a community whose ideals strive to let anyone (no matter their background) contribute to the pursuit of truth.
Please direct comments and inquiries to:
@ForStrumia on twitter