search

WDscholia / scholia

Wikidata-based scholarly profiles
https://scholia.toolforge.org
Other
216 stars 78 forks source link

Investigate gender citation patterns #442

Open fnielsen opened 6 years ago

fnielsen commented 6 years ago
prototyperspective commented 2 years ago

How is this relevant to Scholia? Are you sanely suggesting that researchers decide whether or not to cite something based on the gender of the authors? This should be closed as not relevant to Scholia code.

fnielsen commented 2 years ago

There might be citation patterns that correlate with gender. The two linked articles suggest such a correlation. It does not necessarily mean that researchers cite based on gender, - directly. The correlation could come by other means. I think it would be interesting if Scholia could put a light to the patterns, although I don't think that is "just a SPARQL query" issue.

prototyperspective commented 2 years ago

Yes, but it's not appropriate to add a highly-visible panel on all pages for this. This just facilitates controversy, polarization, unfairness, inefficient work, misassessments, and so on as people call for opinion-based corrections to what they simply consider to be injustices or unfairness or something similar and institutions and journals attempt to bow to public outcries.

Just like it's usually not good for the government to act upon loud public outcries about pandemic mitigation measures, it's also not good to act upon this data and complaints based on them by trying to make science mirror global, national, gender, etc demographics. For the same reason we probably wouldn't add an extra panel for studies that suggest COVID19 not to be a substantial threat, just inviting calls to carry on with business as usual. You may still find such data and studies (e.g. misrepresentations of vaccine side-effects relative risks and so on) if you look for it, but it's not supported/amplified by Scholia. This is just to illustrate what I mean.

Another use-case of these statistics in practice could be students looking to start studying in Universities with a high share of females as this info is included in the institution's page, but I don't think that's necessarily a good factor when deciding such.

Furthermore, it can also lead to various controversy-facilitating results, that neither people interested in this probably would want to accept nor does it add much of value. Example from here:

Strumia’s major finding is that, on average, women write fewer papers than men, their papers are less cited than those written by men, and they are hired with lower bibliometric indicators based on these measures. His findings are significant and robust.

Some more infos are in the issue linked above.

I propose removing that panel from the pages and adding info about how to query such data elsewhere.

fnielsen commented 2 years ago

"The rise of citational justice: how scholars are making references faire" https://scholia.toolforge.org/work/Q111333081

prototyperspective commented 2 years ago

"The rise of citational justice: how scholars are making references faire" https://scholia.toolforge.org/work/Q111333081

I would just laugh at this if it wasn't so sad...these people seem to really believe this and are writing in Nature news. This is how you effectively damage science and the scientific method and these people should be be held accountable for such unethical attempts at distortions which could cause a lot of problems such as more preventable human suffering due to declining science including a loss of reliability, fairness and effectiveness and trust in science.

I'm pretty sure nobody checks which genders or races (usually it's more than one author these days anyway) the cited study authors have, when I reference anything I basically never even know which genders, skin-color or ethnic -group even just the main authors have.

Even worse they go the typical route of discarding objections as "racist" and "sexist", here is the relevant part:

Outright racist or sexist comments — such as “white people or men just write better papers” — are rare, but people often say that they just cite good science or that they don’t see race or gender. These arguments are problematic, Bassett says, because they indicate that people are not actively trying to address their own explicit and implicit biases — or that they are not willing to dig more deeply into the literature in their field to diversify their citations.

Note that they should have put that cited objection more accurately, something like "white people or men may often just write better or more relevant and significant papers and more papers overall".

The worst of all of this is they actually suggest people to artificially "diversify their citations" (based on race or gender).

I don't know why this Nature news article was linked here but wanted to quickly address it. Campaigns like this are doing great harm to science and Scholia definitely should not non-neutrally amplify them.

Edit: many others don't agree with this Nature piece either, see for example here and here, a quote from the latter:

Thank you for posting this article, but I will continue to cite papers that inform my work, without regard to the race, gender or any other such characteristics of the authors of those papers. That imo is what fair citation is.

I always try to cast as wide as possible net in seeking out papers that pertain to my work, if someone has conducted a study or developed a theory that should inform my work I do not want to miss that, but without regard to those characteristics. This notion of "Citational Justice" doesn't resonate with me.

fnielsen commented 2 years ago

"citational justice": When judging from citations to my own articles from others, I would say that there is at least a Matthew effect. One of my articles is (if I may say so) more cited than it should have been, while another is less cited than it should have been. To me it seems to be a "social effect". Such an effect/effects are interesting to model mathematically.

My current pet theory of why there is a citation gender imbalance is due to males writing more foundational papers, e.g., methodological papers. They might be cited more. If I look at functional neuroimaging there is fairly gender balance among authors, but males are typically the ones that write papers about the most used methods, SPM, FSL, AFNI, etc. If one look at the most cited papers in overall science it is typically some bio- or chemistry method paper, e.g., the Bradford test is highly cited.

fnielsen commented 2 years ago

Mishra et al., Self-citation is the hallmark of productive authors, of any gender https://scholia.toolforge.org/work/Q57903876

prototyperspective commented 2 years ago

These are all very unsolid hypotheses with no real foundational basis by which one could postulate these.

For example the study you linked says "As a result, papers by authors with short, disrupted, or diverse careers miss out on the initial boost in visibility gained from self-citations. Our data further suggest that this disproportionately affects women because of attrition and not because of disciplinary under-specialization". I don't think that there is even just remotely sufficient data about

  1. the usefulness of self-citations (e.g. in comparison to alternatives, across fields, and considering correlations)
  2. that this disproportionately affects women
  3. that if this would be because of attrition as the main or even only cause (or even one cause at all)
  4. and that this attrition would be somehow the result of something that's unfair rather than, for example, the reason for why it may in general be just fine or to be expected that it's more men doing science than women and that on average these are getting disproportionally cited

Issues like unaffordable article processing charges (APCs which I mentioned here elsewhere) and things like the snowball effect or Matthew effect where things like early reception (or already-existing popularity or whether or not the authors papers got drowned out by reviews which are cited instead) have an inadequately large impact on the paper's impact are things worthy of a) research in general b) attempts of appropriate(!) changes towards a fairer, more efficient system (for example by improved education about science communication or research topic selection etc) c) reasonable "amplification" by high visibility as tile in Scholia.

However, I think that the focus on gender is undue and effectively counterproductive distraction from such actual issues.

Also note that it's near impossible to make relevant conclusions just based on how you think about two of your own papers or identify some "social effect" this way. And even if you could find (non-neuro/biological) issues by investigating gender citation patterns, the conclusion to suggest people artificially "diversify their citations" (based on race or gender) is wrong and harmful.

I think that doing and facilitating such citation investigations results in nothing (that is valuable) but: unscientific false hasty want-to-believe conclusions, wasted resources (time, effort, financial, work-force, expertise), polarization and division, a loss of trust in and reliability of science, and distraction from actual metascience issues (or interesting topics) like APCs, open access, reproducibility, next generations of scientists, widespread academic practices/system (referring to things like having studied in formal University etc), the education system, and so on.

Edited my above comment to add two links.

I don't think it's due or neutral to feature this tile on every page instead of having a dedicated page about gender-research and would support moving it to a page dedicated about this with gender-related queries for now; we could then reconsider adding it to all the pages in some way in the future.