(C) Gerry Lynch, 2006. Little Gulls winter in a London park.

"bird scholar" research

research question

Do scientific papers describing same-sex behavior in birds cite fewer prior sources than papers describing heterosexual or reproductive behavior?

hypothesis

In comparison of same-sex behavior avian papers and hetereosexual behavior papers, there will be less edges for same-sex behavior papers due to the smaller amount of research. In other words, there is less of a foundation for same-sex avian research, so the papers will cite less in-network sources.

methodology

Papers were collected using the Semantic Scholar API , limited to publications in the Biology category from 2015 onwards, with the queries below. Two separate datasets were built: one for avian same-sex behavior and one for avian heterosexual behavior. Each dataset was built independently, using distinct sets of search queries for each group. For each search query, a maximum of 100 papers were retrieved from the API. They were then filtered with specific keywords to make sure it was relevant to the research. From there, the papers' citations were collected through the same API and were filtered, to make sure they were also relevant. If the citing paper existed in the dataset, it recorded that connection as an edge between the reference and the citation. It drew edges between papers that were already in the collected dataset. In other words, it built a self-contained network rather than including citations to papers outside of the network/dataset. These networks were then visualized using NetworkX and ForceAtlas2, where node size reflects total degree (total number of citation links) and node color reflects in-degree (number of times a source was cited within the network).

See the code here.

same-sex behavior songbird, same-sex behavior avian,
homosexual behavior avian, homosexual behavior songbird,
same-sex mounting in birds, same-sex behavior bird,
bird same-sex parenting, homosexuality in birds,
homosexual bird mating, same‐sex partnerships in birds,
same-sex pair-bonds birds

heterosexual mating behavior birds, avian pair bond breeding,
bird reproductive behavior, avian monogamy breeding pair

same-sex, same sex, homosexual, homosexuality,
female-female, male-male, sexual behavior,
sexual behaviour, mating, pair bond, copulate,
courtship, mounting, pairing, sexual partner,
sex partner, reproductive behavior, reproductive behaviour

heterosexual, heterosexuality, opposite-sex,
opposite sex, female-male, male-female, mixed-sex,
mating, pair bond, copulate, courtship, mounting,
pairing, sexual partner, sex partner, cross-sex,
breeding pair, breeding behavior, mate choice,
mate selection, monogamous, monogamy, nesting pair,
clutch, fertilization, reproductive behavior, reproductive behaviour

bird, birds, avian, aves, songbird, passerine, 
ornithology, finch, sparrow, warbler, thrush, starling,
crane, parrot, owl, duck, goose, flamingo, kestrel, corvid, 
jay, wren, swift, macaw, dove, pigeon, raptor, seabird, 
shorebird, waterfowl, budgerigar, canary, zebra finch, 
quail, chicken, poultry

data

results

+ − Reset

Use + and − to zoom, then drag to pan.

+ − Reset

Use + and − to zoom, then drag to pan.

analysis

There are a few notable observations to take from this study:

• The most cited paper in the same-sex network was a review explicitly calling for more data: Same‐Sex Partnerships in Birds: A Review of the Current Literature and a Call for More Data. This implies that the field itself recognizes its limited foundation.
• The heterosexual network features a dominant hub, Extra-Pair Paternity in Birds, which cites 16 other papers in the dataset. This hub is a review paper that summarizes existing research in the field for the last couple of years. This does not exist in the same-sex behavior network graph.
• Generally, the same-sex network is a little more fragmented and less inter-connected in comparison to the heterosexual network, which is more centralized.

limitations

• Papers without abstracts are filtered using title alone, which may be less accurate.
• There are keyword filtering limitations; it is difficult to pinpoint what words are in relevant titles.
• Due to the small size of each dataset, the citation networks are sparse. Edges only form when both the citing and cited paper exist within the collected dataset, meaning the majority of real-world citations between these papers are not captured.
• The API having a cap of 100 returned results from each query limits how many papers could be collected.

conclusions

Both networks were similarly sparse (42 vs 50 citations, 51 vs 54 papers). There were no statistically significant differences in density to support or refute the hypothesis. However, with this small of a data set, the qualitative data (as stated in the analysis) is more telling than the quanitative.

The quanitative data would be more telling if this study was scaled up to better encapsulate more source data. Another way to collect the same-sex avian studies would need to be constructed, instead of retrieving multiple different queries of data through the API. Perhaps there is an API that has a queer study category that can be used instead? Additionally, this study would require a significantly larger dataset, with a more broad publication date range, and metrics created to be sourced from all cited papers in a specific source, not just in-network ones.

With that said, current queer-ecology research portrays that same-sex behavior is systematically underresearched. Anderson et al. found that while 76.7% of researchers in a survey had observed same-sex behavior in their study species, fewer than half collected data on it, and most of those who did never published it (4). The idea of same-sex behavior as a rare "Darwinian paradox" has withstood despite widespread reporting across all major animal clades (Anderson et al. 2). Specifically in birds, same-sex partnerships persist many years, occur in sexually dimorphic species where mistaken identity is impossible, and serve a purpose past reproduction (Gillies and Siddiqi-Davies 2-6). These missing data sets in same-sex research reveal scientific biases that determine "who counts and who does not," setting heterosexual behavior as the default framework for understanding animal behavior (Livio and Sánchez-Querubín 1787).

Challenging the binary and deconstructing hetreonormative frameworks that have molded and persisted in what gets researched is vital in order to ensure an accurate and all-encompassing understanding of animal behavior. Nonhuman animal behavior has long been used to justify human identities and sexuality, and scientific data that excludes same-sex behavior frames queerness as unnatural. If more diverse and representative data is collected, it can aid in diminishing negative social perception of queer humans.

tools

• Semantic Scholar API
• NetworkX
• ForceAtlas2

works cited

Anderson, Karyn A., et al. "Same-Sex Sexual Behaviour among Mammals Is Widely Observed, yet Seldomly Reported: Evidence from an Online Expert Survey." PLOS ONE, vol. 19, no. 6, June 2024, p. e0304885, https://doi.org/10.1371/journal.pone.0304885.

Gillies, Natasha, and Katrina Siddiqi-Davies. "Same-Sex Partnerships in Birds: A Review of the Current Literature and a Call for More Data." Journal of Avian Biology, vol. 2025, no. 3, 2025, p. e03452, https://doi.org/10.1002/jav.03452.

Livio, Maya, and Natalia Sánchez-Querubín. "Queer Ecological Data: Where Artificial Intelligence Meets the Avian." New Media & Society, vol. 28, no. 4, 2026, pp. 1773-91, https://go.exlibris.link/6V0nPH66.

Savulescu, Julian, et al. "Ethics of Genetic Research on Same-Sex Sexual Behaviour." Nature Human Behaviour, vol. 5, no. 9, Sept. 2021, pp. 1123-24, https://doi.org/10.1038/s41562-021-01164-y.

references

Brouwer, Lyanne, and Simon C. Griffith. "Extra-Pair Paternity in Birds." Molecular Ecology, vol. 28, no. 22, 2019, pp. 4864-82, https://doi.org/10.1111/mec.15259.

Brouwer and Griffith provide an overview of extra-pair paternity in birds. They summarize 500 studies over 300 species to determine mating outside the typical pair bond. This source was identified through the heterosexual citation network produced in this study; it is the dominant hub node due to its high number of outgoing citations to other papers in the data. In reflection, this makes sense as it is a large review article.