Gender Controversy Comes to Physics: A Response to the Statement Against Alessandro Strumia

Introduction

Although physicists tend to fall on the liberal side of the political spectrum, at a professional level it’s generally accepted that science unites us and personal politics should be reserved for conversations over drinks. We aren’t known for social controversy. That exterior peace was broken, however, by a recent controversial talk that sent waves through the community and out into the wider world.

On September 28, 2018, Prof. Alessandro Strumia presented a talk entitled “Bibliometrics Data About Gender Issues in Fundamental Theory” at the 1st Workshop on High Energy Theory and Gender, hosted by CERN. The talk provoked an immediate outcry from some attendees and a public statement condemning it was drafted by several physicists, under the name Particles for Justice. Reportedly, thousands of other physicists have now cosigned the statement. The affair prompted coverage by the BBC, the Guardian, BuzzFeed and many other outlets. Currently, Prof. Strumia is under suspension, pending an investigation by CERN and the University of Pisa, where he holds his position.

The talk that provoked so much outrage was, as might be guessed, about the gender balance in high-energy physics. To give some background: around 20% of PhD-holding physicists are women. Although this varies by country, that number has remained roughly consistent in the US over the last twenty years. Physics is thus near the bottom among STEM fields, as ranked by proportion of women. Naturally, there has been a big push to get more women into the field, with invited talks on the status of women, efforts to promote role models to young girls, surveys, scholarships, networking and so on. The CERN workshop was one such effort and featured mostly talks by women about their physics research. However, several presentations dealt with gender issues directly. Apparently, Strumia asked if he could contribute a talk, and the organizers accepted.

I am a high-energy researcher and first heard about the fracas after it erupted over the weekend. I was directed to the Particles for Justice statement. After reading it through, and then checking Strumia’s slides, I found I could not in good conscience sign it. Indeed, there are problems with Strumia’s talk, but I was greatly disappointed by the reply, and frankly embarrassed that many respected scientists would sign a statement that seems so lacking in scientific and ethical rigor. I thought to myself, how can we endorse a polemic that is at least as bad as the talk it criticizes? I brooded on it for a while, looked into the competing claims and sources, and eventually started writing my own take to ease my mind. This evolved into an open letter addressed to the high-energy community.

Areo has been kind enough to offer me a platform to publish that letter. Unfortunately, I feel that anonymity is the wisest course. Although I am a genuinely liberal person, and although I have striven to be fair and conscientious, I fear attaching my name could harm my career and my relationships. I know there are many other physicists who were also put off by the polemical nature of the response, and who would at least be willing to discuss these things privately, but the social atmosphere is toxic right now.

I would encourage anyone following the controversy to read Strumia’s slides, the Particles for Justice statement, and then my own letter, which I’m afraid is nine pages long—the length required to address the numerous issues in the talk and statement with care. Even then, there are many other points which I might have raised, but I tried to limit myself to the core claims.

The short version: Strumia’s talk advanced the argument that there is no anti-female bias in the field and that gender differences are driven by interest and ability. These are complicated issues and he can’t claim to have settled them. There remain important caveats and confounds, but the argument is not one that has been refuted either and I think it deserves consideration. Additionally, he made a serious mistake in one comment, by bringing up a personal comparison, which was rightly condemned by the community.

The criticism of Strumia, however, has not been confined to the unprofessional comparison or to a measured evaluation of the science. Instead, there has been an effort to demonize him and cast his arguments as pseudo-science and easily refuted bunk. In doing so, his critics have repeatedly misrepresented him, invoked specious arguments, misled people as to the state of the evidence and generally behaved contrary to the spirit of good science.

An Open Letter to the Physics Community on the Strumia Affair
Anonymous

Preface

Recently, the high energy physics community was beset by controversy after a talk given by Prof. Alessandro Strumia at the 1st Workshop on High Energy Theory and Gender. The talk argued that there is little evidence of discrimination against women in high energy physics, and that differing numbers of men and women are consistent with explanations in terms of ability and interest. Strumia’s comments have provoked much criticism and, in particular, a statement of condemnation called Particles for Justice which has been signed by more than 1000 academics from the larger physics community and other fields.

I am writing this because I believe Prof. Strumia’s comments are problematic, but that the response has also been deeply flawed. I am a scientist within the high energy community and I know many of the signatories personally or professionally. I have never met Prof. Strumia although I am aware of his work. I believe the researchers who signed the Particles for Justice statement did so out of a sincere concern for women and legitimate desire to publicly confront dubious conduct on Strumia’s part. However, I find I cannot sign it myself due to multiple ethical and scientific problems in the statement. It may be that some of the signatories have not carefully read the statement or checked the references on which it bases its claims. There have been a number of other online criticisms of the talk, I choose to focus on this one due to its stature. I did not attend the talk, which does not seem to be available as a recording, but the slides can be found online. As such, I cannot comment on any verbal statements from Strumia that aren’t reflected on his slides. The statement at hand, and the other criticism I have seen, is almost exclusively focused on the slides as well. I apologize for the length, but careful statements require more time than sweeping conclusions.

I have chosen anonymity for fear of losing my career or even friendships. I believe physicists are both open-minded and skeptical, but tempers run high on such topics. I have tried very hard to be accurate and compassionate. I hope this letter will be read in that spirit.

Statement of Principles

To begin, let me say that I emphatically wish for physics, and all other disciplines, to be places where all people feel welcome, regardless of sex, ethnicity, sexual orientation, gender identity, nationality or other group memberships. I believe the ideal is that anyone interested in physics should be given the opportunity to involve themselves, and that they should advance in the field according to merit and contributions to science. I also believe scientists should conduct themselves with professionalism befitting their situation, and that their fields benefit from respectful collegial interactions.

We must also adhere to fundamental scientific standards. The above are statements of values, while in statements of fact we strive for objectivity, intellectual honesty, critical evaluation of ideas, the need for empirical data and openness to correction. Good scientists welcome constructive debate and criticism. As a field we make judgments about the validity of various ideas. Some are accepted as essentially true, at least within their domain, and some are for all practical purposes falsified. Such ideas are not the subject of ongoing scientific debate. However, many others are and though they may have partisans who are convinced on one side or the other, we value academic freedom. We support the principle that ideas must be met with evidence and arguments. Consistent with both ethical and scientific values, an idea should stand or fall independent of the person who articulates it.

Problems with Strumia’s Talk

Here I will discuss what I believe are the most serious problems with the talk. The merits of his arguments are discussed in more detail later. In my judgment Prof. Strumia acted very unprofessionally in naming a competitor for a position he ultimately lost, and in comparing his citation count to hers. In this I am in agreement with the Particles for Justice signatories. Such a comparison is needlessly personal and inappropriate for a scholarly talk. Hiring decisions within the community are often fraught with disagreements, but a talk is not the right venue to air such grievances, especially personal ones. CERN has removed the slides from their conference website and suspended Strumia pending an investigation. Insofar as the personal comparison violates their code of conduct, I support this action.

It has also been alleged that Strumia misrepresented himself to the organizers. If true, this would be a significant breach of professionalism. Based on the available information I don’t agree with this assessment. I do think Strumia knew his talk would be provocative. However, although the issues surrounding gender are often felt personally, a provocative conclusion or strong criticism of other theories is not misconduct. Arguably this was an unnecessarily confrontational approach from Strumia. At the same time, it seems the organizers would have rejected his talk regardless of the personal comment issue. He might have gone to the wider community after rejection, but it is not clear his points would have received the attention his talk garnered. Regardless, it was not his duty to ensure the organizers agreed with his talk.

Turning from ethical issues, the content of the talk itself has come under much scrutiny. The talk is written in a casual style, including cartoons and jokes, along with a lot of plots and bullet point items. This is not uncommon within high-energy physics. Links are included in his talk to various sources supporting his claims. The basis for his own analysis is a paper by Strumia and Torre on the arXiv depository. His analysis employs a large dataset of citations and job positions from the INSPIRE website which is used widely among high energy physicists to monitor papers and authors.

Details of his methods can be found in the paper. It should be noted that results shown in the talk are not included in the paper, which was not focused on gender issues. Needless to say, Strumia’s talk does not address every criticism that might be made. He does not include references to every paper claiming to find sexism in some aspect of the STEM fields, or every study that might question his own sources. This would be an impossible task and it is not asked of any researcher. His sources appear to be reasonable, in the sense of not being dismissed in the field. This leaves a number of topics that are not addressed by Strumia’s talk, and possibly evidence that may not fit the viewpoint he is promoting. His analysis does not incorporate every variable and in some places uses a crude model, which he admits. Thus, the summary phrase in his conclusions “Physics is not sexist against women” can’t be taken as a settled issue. There are aspects of sexism, such as reports of harassment, that his research did not address, and the question of discrimination can’t be solved by a single analysis. In this sense I find his presentation too dismissive.

Response to Particles for Justice Statement

Here I will address major points in the statement issued against Strumia, pointing out the flaws which I find make it unworthy as a response from the community. I will take these in the order they appear.

After a description of the situation, the statement begins: “We write here first to state, in the strongest possible terms, that the humanity of any person, regardless of ascribed identities such as race, ethnicity, gender identity, religion, disability, gender presentation, or sexual identity is not up for debate.” This is a fine sentiment, but a deeply deceptive bit of rhetoric. Nowhere in Strumia’s slides is anyone’s humanity questioned, nor does he even raise the issue of race, ethnicity, gender identity (as distinct from sex), religion, disability, or gender presentation. It is unethical to imply that he did. Strumia’s views can be summed up in the argument that the distribution of women in physics is consistent with personal choices and ability, that it does not show signs of anti-female bias, and that the theory that the numbers are driven by discrimination is inconsistent with the data. Those claims are certainly disputed, but they no more question women’s humanity than the claim that male weightlifters dominate world records due to greater male size and interest. Similarly, the authors condemn his “open discrimination and personal attacks.” While I share the concern over the personal comparison, nothing in his slides discriminates against women or calls for it.

The authors write, “It is clear to all of us that Strumia is not an expert on these topics and is misusing his physics credentials to put himself forward as one.” Strumia does not present himself as an expert in gender studies, nor does he rely on his credentials as a physicist. This is also not a standard applied to other speakers at the workshop. For example, Dr. Jessica Wade has a degree in physics and works on light emitting diodes. Although serving on several women’s councils, she lists no academic papers in gender studies. Nonetheless, the fifth slide of her talk announces, in regard to the issue of women’s participation in STEM fields, “this has NO biological origin.” Her slide links to a paper on sex differences in school grades which makes no such claim. In fact, its findings are quite compatible with Strumia’s view. I don’t wish to single out Dr. Wade, who is entitled to her argument, but to point out the hypocrisy. The criticism that Strumia is not an expert is an appeal to authority. If his arguments can be easily refuted there is no need to rely on unspecified experts who disagree.

Several alleged errors are detailed in numbered points to show Strumia’s bad science.

1. The authors argue that the greater presence of women in humanities can’t be evidence against disparity driven by bias because it doesn’t account for systemic discrimination. They then link to three putative pieces of evidence for said discrimination. The first, Hodari et al., does not describe discrimination, it only lists relative prevalence of women in various categories. The second, Johnson et al., draws on “qualitative findings” from interviews with a handful of women of color and finds that some feel a sense of isolation. The third is the NASEM report on harassment in STEM, the only example cited which actually discusses sexist behavior across STEM fields in any detail. Reported harassment is certainly a serious topic to consider, and a large survey of the field is helpful. No one should be subjected to harassment and the reported incidence should be addressed.

Strumia, however, did not claim that harassment never happens, he argued that discrimination was not the cause of the gender imbalance. According to the NASEM report, for faculty/staff-on-female-student harassment in the University of Texas system, 1% of male and female science students report sexual coercion and 4%(2%) of females (males) report unwanted sexual attention. Some 17% (13%) report sexist hostility and 8% (5%) describe crude behavior. What is notable is that these numbers are the same or lower than those in non-STEM fields, in Engineering, or in Medicine. In particular, female students in medicine reported a 45% rate of sexist hostility. This is not the full picture of harassment, but it is the only comparison across fields in the STEM chapter. Hence, we don’t see evidence of exceptional rates of harassment in science, and we see them much higher in a female-majority profession like medicine. While harassment surely happens in physics as in all fields, including it would apparently only strengthen Strumia’s assertion that it does not drive the gender differences between fields. This criticism of Strumia also elides his claims that women are better represented in fields like Business and Law, which have “real power” and better pay. It also fails to address his contention that they are better represented in administrative positions than as physicists or technicians.

2. The statement criticizes Strumia’s citation of the “Gender Equality Paradox,” which finds higher-equality countries tend to have bigger gender gaps in STEM. They write “This claim ignores cultural differences, and also the possibility that women in such countries have fewer career options outside of academia. Without controlling for such effects, any attempt to draw conclusions is meaningless.” This is a deeply unscientific response. The “paradox” is in fact a result of attempting to control for cultural differences. The Gender Equality Index is a composite measure of women’s reproductive health, access to education, participation in government and participation in the workforce. The surprising result is that cultures with higher general gender equality have more separation in STEM fields. No one advancing a discrimination-based theory would have predicted this, but it is compatible with interest- and ability-driven explanations. This includes the idea that participation in unequal societies is artificially inflated by limited career options. Of course, it remains possible that some other correlate explains this incongruity, but it is frankly irresponsible to declare this evidence “meaningless.” It is up to those advancing the discrimination theory to explain what more complicated model might account for their failure.

3. The authors note that even though Strumia finds no difference in the way men and women cite papers, the fact that men are cited at a somewhat higher rate could be due to a shared anti-woman bias. They write “Such unconscious bias is often found at similar levels in both men and women.” The link goes to a webpage collecting a number of recent articles discussing gender issues in academics. The articles are quite disparate in methods, goals and findings. In fact, they include an entire section on articles that find lack of bias or bias in favor of women in STEM. A talking point in the introduction states “Women are as likely as men to make biased judgments that favor men,” but this is not the subject or conclusion of most of the papers.

A relevant article among the collection is Moss-Racusin et al., which does find anti-female bias. A blind study of 127 faculty members in science disciplines found that both genders tended to rate women lower on a hypothetical evaluation of hireability for an undergraduate position, despite identical applications. This is certainly worth considering, but one small study of non-real-world conditions can’t settle the question of the state of high-energy physics. For example, another article found on the same webpage, by Williams and Ceci, performed a randomized study of 873 faculty and found a 2 to 1 hiring preference for women in Biology, Engineering and Psychology. One may compare the roughly 10% male advantage on a point scale in the Moss-Racusin et al study.

So, while anti-female bias in citations isn’t ruled out by Strumia’s data, we don’t seem to have compelling evidence that there is such a prejudice. If there is, it seems to be the same for both women and men. It is also disingenuous to say that “the equal citation rate at most only suggests that male and female scientists are equally capable of identifying the most cited papers in their field.” Researchers do not select citations based on “the most cited papers in their field” but on those which are deemed most relevant to their work. This will often include seminal works, or those considered the most up-to-date and thorough, which we expect will correlate with citation count.

4. The authors write “Strumia argues that since the most cited papers are disproportionately by men, this gives evidence that men are intrinsically better at physics. In between intrinsic ability and citation counting however, there is the huge and complicated process of how physicists are raised, trained, hired, and perceived.” It’s true that one’s career path involves many factors and citation metrics could be caused as well as causal. I believe a significant element of luck enters for an individual depending on who mentors them, who hires them, where the interest in the field is, etc. However, we are looking at averages over large numbers of people and papers. Why is there a male-female split in the distributions and what does it mean? In fact, Strumia does not use the words “intrinsic” or ”innate” on his slides, so this is a bit of a strawman. The question of why men produce a higher citation index throughout their careers does not change the fact that they do. One can speculate, for instance, that men are preferentially hired by more prestigious mentors, leading to more citations; it would still be the case that they produce more by this metric.

According to the statement “Even at the professorial level, discrimination can still play an important role (such as e.g. the imbalance in telescope time awarded to female researchers).” Perhaps it can, but the citation doesn’t show this with any certainty. That paper found that men submit more proposals for telescope time and have mildly higher ratings on average. Both men and women graded the men’s proposals as better. They also rated professors’ proposals as better than those of postdocs, who were better than students. Unlike the citation differences in Strumia’s study, this study finds a small difference between how men and women grade, thus changing the fraction of women graders would change the difference in male vs. female grades. But, if there is such a thing as a “true” grade this study can’t tell us which way the bias runs, and even making the graders entirely female wouldn’t eliminate the gender differences in outcome. How much any of this could affect one’s overall career is well beyond the scope of the paper.

5. The statement here focuses on Strumia’s “case study” comparison of citation count between himself and a woman who was offered a job they both applied for. As stated above, I agree this was a major breach of professionalism. As they say, use of raw citation count for a job interview is simply not tenable, and many other factors should be considered for any hiring. Aside from ethics, this was a dumb move for Strumia. I don’t know if it was meant tongue-in-cheek, or if he carries a deep resentment, but it allows his critics to assume the latter and dismiss his arguments with ad hominem. It also allows them to focus on the raw citation number he used in this one comment, rather than the more plausible metrics used throughout the paper. In particular, on the same slide Strumia shows gender comparisons of CERN fellows where men have a clear advantage in number of papers, number of citations and his preferred metric, N_icit.

The authors write of citation counts that “using them as a substitute for scientific quality is very problematic.” This is undoubtedly true, especially for individual papers, also for individual authors. Yet, no one seems to believe they are meaningless in the aggregate. For instance, people are very concerned about impact factor, which is a citation-based rating of journals and which does indeed seem to correlate with those considered most prestigious, at least within a field. The basic idea that important papers should accrue citations and forgettable ones should not is hard to ignore.

As an example of problems with raw citations, Strumia’s critics note that a huge chunk of his come from the Higgs discovery paper, which lists thousands of authors associated with the CMS collaboration. One wouldn’t assume that Strumia’s contribution is somehow equal to the total effort, but at the same time it seems entirely correct that that paper has thousands of citations. It was a Nobel Prize winning discovery and will influence the field for all time. This is why the measure N_icit averages over the number of authors on a paper. This corrects for the effect that large experimental collaborations are central to high-energy physics but individual contributions can’t be simply sorted out.

Another point raised is that a big number of citations comes from Strumia’s 750 GeV bump paper. This is a more complicated case though, and can’t be simply dismissed. The bump is described as “statistically insignificant,” but this is misleading. The bump did not cross the high threshold used to announce a “discovery,” but it appeared to be well in the range considered “evidence” which is why people, including many respected researchers, wrote papers about it. More data revealed it to be an unlikely fluctuation, but this does not mean the papers written were intrinsically bad. Because it is ruled out it will garner few citations after the active period, and will fade into obscurity. This is true of most theory and phenomenology papers. Many physicists study possible models of new physics and how to detect it, and most of those models will eventually be ruled out or lose the community’s interest. As I write men and women are working hard on manuscripts about current anomalies, all of which may turn out to be statistical flukes. Being a good author of such papers means writing those that are innovative, interesting, plausible and scientifically strong. Within the field this is how quality is judged and ideally citations should track this. There is more to say about the sociology of the 750 GeV excess, and this is not an argument about the exact merit of Strumia’s paper. However, it is important to recognize that the bump’s disappearance with more data is not an argument that citations are meaningless.

Strumia’s N_icit further averages over the number of references in a paper. The goal is to compensate for the inflating number of references per paper over time. This would also tend to normalize different sub-fields that may have different referencing habits. In the Strumia and Torre paper, the authors consider several other metrics which try to account for the state of the overall field, to eliminate the effect of self-citations and citation “cartels,” to estimate the long-term impact of a paper, etc. They find that N_icit correlates well with these other sophisticated measures, and not with naive citation counting. Strumia’s preferred metrics have not been widely adopted in the field (the paper is from March 2018). Nor will this end the debate on how to measure quality or contributions to research. However, no one can accuse him of not having thought about this, or of not making a good faith effort to establish an indicator that behaves the way we want. No proxy will be a perfect measure of quality, but other measures are not obviously better.

6. Addressing the graph which shows a growing divergence in citation index over male and female careers, the authors posit that this may be due to differences in parenting roles or departmental appointments. The former at least seems quite possible, as research in other careers suggests that family decisions play a role in the earnings gap. This would not change the fact that men produce more, as measured by various metrics, over their careers. Such an explanation is one of personal choice, perhaps reflecting larger social influences, but not of discrimination within physics. They claim that this hypothesis may be supported because “the decline he claims does not begin until after the postdoc level.” This is not true, we see a small split early on which gradually grows with time and this is mentioned explicitly on the slide. The rate of divergence seems to increase after ten years, then again around 25 years. It’s an interesting pattern without an obvious cause. One might also point out that after the postdoctoral phase a researcher is more independent. None of this addresses Strumia’s finding that women are hired to faculty earlier than men, and with lower individual citations. We wouldn’t expect this pattern if the citation metrics were random noise.

It is worth commenting on Strumia’s ability model here. The model assumes that the total ratio of men to women is set by interest, that they have equal mean ability, and that men have higher variance, leading to male predominance at the high end. He is correct that studies have found greater male variance on the 10% level for various tests of academic performance (including in the reference cited by Dr. Wade), and that the genders exhibit average differences in interests. It is also true that physicists are at the high end of the IQ range. That said, the model he uses is very crude and can’t be taken too seriously. We don’t know how well IQ or other test scores correlate with citation success in physics, we don’t know how normally the tails of the distributions behave, and it’s not clear if changing representation with time has an effect. Nonetheless, the idea that differing variance plays a role is worth considering and might also include factors besides intelligence. It’s not obvious, though, how bias can fit the pattern shown. There are additionally many questions about how environmental effects might contribute to differences in the first place, but that is a separate issue from bias within the field.

7. As the authors point out, it would be silly to say that Marie Curie faced no discrimination in her day. They then list four women who arguably should have won the Nobel Prize. This list tells us nothing, unfortunately, since they did not attempt to compare with the many men who might have deserved the prize. They write “a gap of 55 years since the last woman won the Nobel Prize in Physics does not suggest that women in our field face no external obstacles to success.” At the same time, a gap of sixty years, then another gap of fifty-five years, is also not a pattern that suggests discrimination, unless one believes that almost nothing has improved in gender attitudes since Curie’s time. The authors state that the accomplishments of the women mentioned “were not formally acknowledged,” but this is simply false. Wu, Rubin, Meitner and Bell Burnell all received awards, prestigious positions and accolades, just not one particular prize. None of this means they faced no obstacles, but the case can’t be made with bad arguments.

8. The last numbered point concerns commentary about gender roles and discrimination in other aspects of society. I think there is nothing much to say here; the topic is far too broad to condense into a coherent summary. Both men and women face different legal and social environments, the relative benefits and penalties of which can’t be simply judged.

Wrapping up, the authors write, “Ultimately, answering questions of cause and effect is subtle and requires carefully designed studies.” No one could disagree with this statement, but Strumia’s data is interesting and the arguments he made can’t be dismissed out of hand. They present questions and challenges to those who believe physics is particularly biased against women, and as seen above, strong conclusions are often drawn on misunderstood or shaky evidence from the other side. They “reiterate that Strumia’s arguments are morally reprehensible.” I cannot agree. Whether his scientific case is sound or not, his hypothesis is not reprehensible. Strumia believes physics is on average a meritocracy, which I think most scientists would agree is the ideal. If he is mistaken then he is scientifically wrong, but this does not reflect on his morals.

They continue, “Finally, we would also like to underline how grossly unethical it is to misrepresent the topic of one’s talk to workshop organizers to promote an agenda which is antithetical to the workshop itself.” The title of his talk “Bibliometrics data about gender issues in fundamental theory” is accurate. It does not indicate his conclusions, or commentary towards the end of the talk. This is commonplace in academic talks, and does not seem out of context in a session on gender participation. Indeed, no one has raised this issue regarding the other talks. The problem is apparently that the organizers assumed Strumia would not advance a conclusion they disagreed with, and that he would not criticize other viewpoints. This does not comport with a scientific attitude in my opinion.

The authors then urge future organizers to seek guidance from certain experts. The first link goes to a book by Dr. Karen Barad on her theory of “agential realism,” which is described as “Offering an account of the world as a whole rather than as composed of separate natural and social realms, agential realism is at once a new epistemology, ontology, and ethics.” The third link is to a psychoanalytic analysis of interpretations of quantum mechanics. This is not the kind of work most physicists would endorse, and certainly not what is considered rigorous science. The other citations are to opinion pieces with two exceptions. One is an uncontroversial report on the number of women in physics. The other is a study which finds women-led papers in astronomy receive fewer citations, consistent with Prof. Strumia’s findings.

Concluding Remarks

I wish I could endorse the statement issued by Particles for Justice. There are serious issues of professional conduct which need to be addressed concerning Strumia’s personal comparisons. There are also many open questions about the status of women in physics, how to measure performance, what affects performance and so forth. I would be happy to sign a community statement which addressed these issues in a well-considered way. Unfortunately, there are major problems with the statement we have, including unethical misrepresentation of Strumia’s talk, misleading citations, poor analysis, and unscientific attitudes towards questions of fact. I have tried to confront the situation in a balanced and cogent way. I have not tried to address every issue that might be raised. Surely no one can claim the final word on gender representation, bias, or rankings.

I fear that some women reading this will see it as an attack, or a chilly attitude towards their presence in physics. I sincerely do not intend this. I know many women in physics, including those who have undoubtedly made bigger contributions to the field than I, and I believe they all belong. I do not claim any certainty as to what governs the different trends we see across disciplines. It seems clear that biology, culture, and circumstance can all interact in complex ways. My wish is for the field to be welcoming to women and to people from all walks of life who are interested. It is highly competitive, but I believe the vast majority of physicists strive to be impartial and to judge people as individuals based on their work.

Undeniably, some women still report harassment. This is an issue across society, however, some evidence suggests it is less so within physics. We cannot easily prove or disprove subtle biases, they may still be present, but available evidence does not seem to be conclusive and biases may play out in different directions at different levels. I would urge people not to let fear keep them out. Good work is recognized and women have achieved success at all levels of the discipline. Whether there are biases or not, we must constantly guard against assuming that individual ideas or people fit the patterns we may perceive. Ultimately, regardless of the current state of the field, we must all strive to uphold our ethical and scientific ideals. I have tried to so this in writing this letter.

If you enjoy our articles, be a part of our growth and help us produce more writing for you:

20 comments

  1. Strumia’s slides are certainly a bit dickish in places, and mentioning somebody by name was unforgivable.
    However, it came over as a reasonable attempt to see what the reality is based on what little “objective” data we have.

    One the other hand, the Particles for Justice reply is absolutely dreadful, for all the reasons you say. Quite epic straw-manning.
    I really hope that any physicists who signed it either did not read the slides or did not read the letter. We expect better from scientists.

    Caveat: the verbal presentation might have been much worse than the slides.

  2. The author rightly describes the following sentence in the Particles for Justice statement as deeply deceptive rhetoric:

    “We write here first to state, in the strongest possible terms, that the humanity of any person, regardless of ascribed identities such as race, ethnicity, gender identity, religion, disability, gender presentation, or sexual identity is not up for debate.”

    In philosophy language which is classified as “rhetoric” is usually just dismissed, but I do not think we should let those behind the PFJ off so lightly. There is no good way to view this sentence (and the paragraph that follows it). At best it is cringe-worthy virtue signalling and at worst it is an outright lie (the implication that Strumia denied someone’s “humanity”). Either way rhetoric like this is far from harmless – both for the individual they are vilifying and also the wider debate.

    Strumia may have acted inappropriately by including information about his personal experience – but there are ways of responding to that which do not include a mob with pitch forks. Ethical considerations aside, if you do not find that part of his presentation convincing the principle of charity requires you to ignore it and evaluate the remainder of his argument.

    I am a big believer in science, but I find it extremely distressing and also frightening that such a large group of scientists who should know better either failed to read something they signed their name to, or seem to have forgotten basic principles of logic and critical thinking, not to mention standards of human decency.

  3. The problem with Strumia’s talk is its general rudeness littered with sarcasm reflecting his biased agenda. if you’re going to give a talk, then try and remain impartial and remain focused on the data. For example, these non-experts:

    Yes, women in science still have a disadvantage by physicist Sabine Hossenfelder,
    http://backreaction.blogspot.com/2018/10/yes-women-in-science-still-have.html

    Jordan Peterson: Women want workaholic men in positions of power
    https://www.youtube.com/watch?v=7STtjfoWz38

    There’s a way of expressing controversial views without coming across as a trolling jerk.

  4. The reason why Strumia was suspended and underwent disciplinary procedure was that he accused openly two female colleague of having got the job that he wanted because they are women and he bragged that he deserved the position because he had more citations (which are, of course, by far not the only issue that is considered in any academic recruitment process). His opinions on the causes of under-representation of women in STEM were not the reason for the sanction. Besides that, his talk was bellicist and provocative, clearly intended to be offensive to the female audience (he said that “physics was built by men” and similar stuff). There can be a reasoned discussion about the under-representation of women and men in certain disciplines and roles, but I doubt that Strumia has done much to promote it.

    5
    25
    1. Did you actually read and try to comprehend this article? The author acknowledges using names was wrong. However the 2 to 1 hiring preference for women cited in one if the papers and the staggering difference between the hired woman (and her boss) in citations I could imagine to be frustrating.

      And what is wrong with a provocative talk? Isn’t that good for the discussion?

      5
      1
  5. Great letter. This is one I would sign with my real name if it were to be sent to CERN as a petition. I encourage you to think about doing this.

    The only thing left out (but Strumia left it out as well) is the fact that occupational choice depends on what economists call “comparative advantage”. It’s not how good you are at (and how much you like) doing math that determines whether you will enter a STEM field. It’s how good you are at / how much you like math relative to how good you are at (and like) doing other things.

    Girls who score very well on math tests tend more than boys to also score very well on language tests. So now there are three reasons why women avoid STEM fields. On average, women are more interested in people over things and boys in things over people. Boys are more numerous than girls among the very high achievers in math. And girls who are among the very high achievers in math are (more often than boys) also very good at other things.

    The people who think it is all discrimination should be obligated to test their hypothesis. Show us that the women who pioneered the movement into medicine and law in the 1970’s were met with open arms relative to the women attempting to enter STEM today.

    Lorne Carmichael
    Professor of Economics (emeritus)
    Queen’s University

    24
  6. The thing about cultural differences is that cultural variation in gender roles isn’t one-dimensional. It is not only possible but extremely common for Society A to be stricter about one thing than Society B, but less strict about other things.

    Here’s just one example. In the United Kingdom, surveys find that half of gay men aren’t comfortable holding hands in public (https://www.bbc.com/news/uk-41179976). But in Saudi Arabia, men hold hands with each other in public all the time, and nobody bats an eye (https://www.youtube.com/watch?v=UtzpQf3O1K4). Is this because Saudi Arabia has more liberal gender roles than the United Kingdom? Of course not. This is because Saudi Arabian gender roles, while far stricter than British gender norms in many other aspects, do not include nearly as much stigma on physical affection between platonic male friends.

    Is it so hard to imagine that Western cultures could also have weird taboos about girls and technical fields that some other cultures, including ones which are more patriarchal in other ways, don’t share? Personally, I don’t find it hard to imagine at all.

    Also, people should really stop using the Global Gender Gap Index. Among other questionable metrics, it considers countries “more equal” if women live longer than men. Russia (along with many other countries) is ranked number 1 for “health equality” because a lot of Russian men drink themselves to death at a relatively young age, resulting in an especially large gender gap in favor of women in life expectancy.

    12
    4
    1. Sure, maybe ‘Western cultures could also have weird taboos about girls and technical fields’ but girls choosing to go into STEM in Western cultures won’t find those taboos enforced with eg. fines, flogging, prison time up to life, death, whipping, or deportation (which are the possible penalties for same sex sexual activity in Saudi Arabia). In western cultures, girls and women especially have been encouraged to flout taboos in recent decades. It really does seem that women choose not to do STEM because they don’t like it. Of course, their not liking it might have a cultural cause, but I think we should consider such things as the inculcated perception that every girl born deserves to be a princess, alongside the unproven idea that a male-dominated profession exhibits sexist behaviour towards its female members.

  7. The alt-Left made it clear with Google’s firing of James Damore’s and making him unemployable in the industry, that questioning is not allowed. Since it is obvious that men and women are different, the only course of action open is to falsely represent the enemy’s arguments and then attack those faulty positions. This is known as the “strawman fallacy”.

    James Damore did not say what most people think he said. He did not say women are unsuited for tech jobs. He did not say women only have tech jobs due to affirmative action. That is all bullshit piled on by those mau-mauing him, in an effort to expel someone whose questioning made them uncomfortable.

    Why it made them uncomfortable is the fascinating part of all this.

    It makes them uncomfortable because it shatters the brittle shell of their enforced narrative, which can only be kept by keeping everyone from saying anything against it. It is very much a “the king is naked” situation. If James Damore is allowed to say men and women are different, the entire project of feminism shatters.

    And yet everyone who doesn’t live in academia or in a social justice echo chamber knows that men and women are different.

    36
    3
  8. The very last comments in Strumia’s slides are so sad, and ironically so. He felt confidence in the intellect of his audience only to find his confidence was misplaced. And even sadder still, through the reaction of his audience, we must now consider the possibility that the differences we see in balance of sexes in the sciences may indeed have something to do with sexually dimorphic limits on the ability to employ intellect instead of emotion: an unintentional self-indictment. This could, in fact, make it even harder for those very bright young women I knew in my mathematics and physics classes to make their mark. Truly, this is a tragedy.

    Although I believe it is good that those of us out here in layperson land see that scientists are human, one cannot come away from this questioning some scientist’s understanding of human cognitive faculties and their ability to interpret nature. On the other hand, science has made a lot of progress against the influence of soothsayers, diviners, and fortune tellers, and it should not be to surprising that they are striking back, and they are doing it through the social sciences and grievance studies. Let’s hope that sociobiology can tame this bunch before it’s too late.

    14
    4
    1. correction: …one cannot come away from this *without* questioning some *scientists’* understanding of human cognitive faculties…

  9. As much as I would like to consider the arguments against Allesandro Strumia, The fact that the video has been scrubbed from any and all servers induces me to conclude that the various statements of censure are entirely without substance or validity. It would appear that CERN learned from Google’s mistakes in the Damore incident. It’s rather difficult to prosecute a proper witch hunt when the disputed tract has been made widely available for general review. CERN has made no such error, cannily forcing the unwashed masses to form an opinion based almost entirely on hearsay. How delightfully “scientific” of them.

    32
    1
  10. Evidently everyone involved is discussing citation numbers, or perhaps rates per person, on a base population in which men outnumber women four to one. On the number of citations, of course citations of male-written papers outnumber citations of female written papers. They should outnumber them four to one.

    You might expect rates per person to correct for this, but they won’t. With four times the number of people in the population, males will have four times the variance, and four times the representation on the right tail. But males will also have a greater rate on the undesirable left tail as well, right? Nope. The left tail ends at zero. Both women and men will stack up on the left tail. The distribution is Poisson, not normal.

    So the male four to one employment count will strongly bias citation count towards men without any cause other than the four to one employment count.

    5
    13
    1. So if you double a population, but keep other citation practices the same (number of citations in a paper, authors per paper, papers per author), you get a higher average citation count per person? After all, there are double the amount of people on the right side of the distribution, but not on the left (as it it’s zero).

      This doesn’t make any sense.

    2. You would get that if males disproportionately cite males and females disproportionately cite females, but I’m not aware of any evidence that this is the case.

    3. That is just a question of using the right sample size for a representative comparison. You could arbitrarily select one forth of the male candidates and only compare their citation numbers to the female candidates to avoid this. The higher your sample size, the less significant the difference in variance anyways, so that few outliners would not skew the statistical relevance out of the confidence interval.

Leave a Reply

Inline
Inline