Monkey and Me: Students’ Perceived Similarity of Humans to Other Species Across Domains

class: center, middle, inverse, title-slide

.title[
# Monkey and Me: Students’ Perceived Similarity of Humans to Other Species Across Domains
]
.author[
### Thomas Pollet & Jeanne Bovet <br> Northumbria University <br><br> (<a href="mailto:thomas.pollet@northumbria.ac.uk" class="email">thomas.pollet@northumbria.ac.uk</a>)
]
.date[
### 2025-02-12 | <a href="http://tvpollet.github.io/disclaimer">disclaimer</a>
]

---

<div>
<style type="text/css">.xaringan-extra-logo {
width: 110px;
height: 128px;
z-index: 0;
background-image: url(NU2.png);
background-size: contain;
background-repeat: no-repeat;
position: absolute;
top:1em;right:1em;
}
</style>
<script>(function () {
  let tries = 0
  function addLogo () {
    if (typeof slideshow === 'undefined') {
      tries += 1
      if (tries < 10) {
        setTimeout(addLogo, 100)
      }
    } else {
      document.querySelectorAll('.remark-slide-content:not(.title-slide):not(.inverse):not(.hide_logo)')
        .forEach(function (slide) {
          const logo = document.createElement('div')
          logo.classList = 'xaringan-extra-logo'
          logo.href = null
          slide.appendChild(logo)
        })
    }
  }
  document.addEventListener('DOMContentLoaded', addLogo)
})()</script>
</div>

## Today

* Work in progress.

* Largely exploratory project.

---
## Back story

* Grant on Evolutionary Psychology text books and 'tokenism' (and perceptions).

* Picking certain cultures and species (but then not others).

* When coding examples of primates in textbooks - Example species chosen.

* For example, Buss (2019: 502) writes: _‘other primates besides humans, such as chimpanzees, baboons, and macaques, also engage in reciprocal helping (de Waal, 1982). Taken together, this evidence suggests a long evolutionary history of altruism.’._

???
Scorpionflies...
---
## Preconceptions about similarity to humans

* What are students' preconceptions about a species's similarity to humans?

* At a later stage perhaps link the usage of examples in textbooks to these preconceptions.

---
## This project

* How do students perceive similarity of species to humans - and what is the role of different domains (e.g., Diet vs. sociality)?

* Two studies: Primates / Other species

* Largely explorative project - but we pre-registered an analysis plan.

* 'Psychonetrics' to evaluate grouping of domains. More about that soon.

???

---
## Study 1 : Methodology - Online Survey (Qualtrics)

* **Prolific Sample**. To the best of our ability, balanced between 'Biology' (final _n_ = 220) and 'Psychology' (final _n_ = 243) degrees. Next, we attempted to balance based on Gender and country.

* Roughly equal split in gender categories (M/F), median age = 25 years. Around three out of four: UK/US resident (76%).

* Rated Baboon, Orang utan, Gorilla, Bonobo and Chimp  across 11 domains (Diet, Physical Anatomy, Brain Anatomy, Cognition, Sexual Behaviour, Disease, Physiology, Learning behaviour, Social Behaviour, Sex differences, and Genetics). On a 0 - 100 slider.

---
## Study 1: Methodology (cont'd)

* "_You will now be asked to rate the similarity of non-human primate species to humans with 0 indicating 'not at all similar', and 100 indicating 'totally similar' across a series of domains._" (Block randomised based on species, and then randomised domain within that block)

* Some sociodemographics and some background questions (Knowledge / Interest in Primates / zoo visits / interests in documentaries on primates).

* Focus is on how the domains cluster in this rating task: **Psychonetrics**

---
## Psychonetrics

* As an input we take the correlations between items. The method we use allows picking the appropriate correlation method (continuous, ordinal or tetrachoric).

* Implemented via `bootnet` -- here we used `EBICglasso` to estimate a 'Graphical Gaussian Model' (GGM). Default tuning parameter of .5, trade-off between sparsity vs. precision.

* In a GGM, nodes represent the variables (e.g., items), and edges represent partial correlations between variables, i.e. correlations after adjusting for all other variables in the network.

* Widely used to model symptom networks, etc. This is a method for variable selection. See [Constantini et al. 2015](https://doi.org/10.1016/j.jrp.2014.07.003) for a tutorial.

---
## Clique percolation method (CPM, Lange, 2011)

* Many algorithmic methods exist to find structure in networks ("clusters"/"modules"/"communities"): [Louvain method](https://doi.org/10.1088/1742-5468/2008/10/P10008) (Blondel et al., 2008) / [Leiden method](https://doi.org/10.1038/s41598-019-41695-z) (Traag et al.2019)

* Most algorithms force unique cluster membership, and also force that each node is part of a cluster. --> not case for CPM

* Very simply put: This method first identifies 'cliques' (in our case k = 3), next examine if these cliques overlap ("adjacent").  Next via "percolation", we grow 'communities' via joining up these adjacent cliques.

Implemented via `CliquePercolation` package in an algorithmic way. **Key point**: nodes can be part of more than one cluster. Isolates can exist.

<div class="figure" style="text-align: center">
<img src="https://upload.wikimedia.org/wikipedia/commons/f/f5/Percolation_p%3D0.51.gif" alt="Thierry Dugnolle, CC BY-SA 4.0 &lt;https://creativecommons.org/licenses/by-sa/4.0&gt;, via Wikimedia Commons" width="200px" />
<p class="caption">Thierry Dugnolle, CC BY-SA 4.0 <https://creativecommons.org/licenses/by-sa/4.0>, via Wikimedia Commons</p>
</div>

---
## 'Invariance testing'

Using a network comparison test, we can compare network structures between groups. [(van Borkulo et al., 2023)](https://doi.org/10.1037/met0000476).

--> **Permutations**

---
## Results

* Some descriptions, (exploratory) t-tests, (exploratory) signed rank tests, and (exploratory) ANOVAs.

* Mostly graphs.

---
## Unsurprising findings (exploratory)

* Biology students (_M_ = 49.97; _SD_ = 22.00) reported greater knowledge on primates than psychology students (_M_ = 40.73; _SD_ = 22.68), _t(458.82)_ = 4.45, _p_ < .0001, _d_ = .414.

* Biology students (_M_ = 54.53; _SD_ = 22.81) reported greater interests in primates than psychology students (_M_ = 46.09; _SD_ = 23.61), _t(459.03)_ = 3.91, _p_ = .0001, _d_ = .363.

* No significant difference in visits to zoo (W = 25986, _p_ = .588) or watching documentary on primates (W = 27045, _p_ = .823)

???
prim_documentary == "Monthly" ~ 6,
prim_documentary == "Weekly" ~ 5,
prim_documentary == "Multiple times a year" ~ 4,
prim_documentary == "Once every two years" ~ 3,
prim_documentary == "Once a year" ~ 2,
prim_documentary == "Never" ~ 1

---
## Exploratory ANOVA on ratings

| Effect                          | df               | MSE     | F        | η²   | p      |
|---------------------------------|------------------|---------|----------|-------|---------|
| Discipline Prolific             | 1, 441           | 14167.72| 15.94 | .017  | <.001   |
| trait                           | 6.92, 3052.98    | 1046.31 | 56.04 | .030  | <.001   |
| Discipline Prolific:trait       | 6.92, 3052.98    | 1046.31 | 2.34  | .001  | .023    |
| species                         | 3.74, 1648.59    | 864.36  | 74.10 | .018  | <.001   |
| Discipline Prolific:species     | 3.74, 1648.59    | 864.36  | 0.11  | <.001 | .974    |
| trait:species                   | 33.31, 14690.44  | 147.06  | 4.04  | .002  | <.001   |
| Discipline Prolific:trait:species| 33.31, 14690.44  | 147.06  | 0.82 | <.001 | .755    |

???
generalised eta squared

---
## Plot: Discipline

---
## Plot: Traits

---
## Plot: Species

---
## Network Invariance: Comparing across primates.

**Network invariance**: all _p_ > .2 --> Structure is roughly the same.

**Strength invariance**:

- Gorilla - Orang , _p_ = .031
- Gorilla - Chimp , _p_ = .073

--> Overall picture across all species.

---
## Psychonetrics: Overall

---
## Invariance: Biology vs. Psychology

They do differ:

**Network invariance**: _p_ = .0009

**Strength invariance**: _p_ = .609

--> but both have two clusters of 3 nodes.

---
## Biology

---
## Psychology

---
## Interim conclusion

* Clustering of traits: "learning, social, cognition" and "genetics, physiology, physical anatom"

* Different organisation between psychology and biology students but both seem to have the two above clusters.

---
## Study 2: 🐜, 🦆, 🐭, 🐬, 🐒

* Repeat the exercise with: 🐜, 🦆, 🐭, 🐬, 🐒.

* Final _n_ = 470. 228 men, 227 women, 15 non-binary or third gender. Approx. 80%. from UK and US.

* 'Biology' (final _n_ = 248) and 'Psychology' (final _n_ = 222)

---
## Unsurprising findings (exploratory)

* Biology students (_M_ = 58.02; _SD_ = 21.31) reported greater knowledge on animals than psychology students (_M_ = 50.08; _SD_ = 19.88), _t(467.19)_ = 4.18, _p_ < .0001, _d_ = .385.

* Biology students (_M_ = 65.27; _SD_ = 21.20) reported greater interests in primates than psychology students (_M_ = 58.50; _SD_ = 20.55), _t(465.05)_ = 3.51, _p_ = .0001, _d_ = .324.

* Significant difference in visits to zoo (W = 31456, _p_ = .005) or watching documentary on nature (W = 31284, _p_ = .008).

---
## Mixed model ANOVA (exploratory)

| Effect                          | df               | MSE     | F           | η²   | p      |
|---------------------------------|------------------|---------|-------------|-------|---------|
| Discipline Prolific             | 1, 455           | 12394.99| 1.60        | .001  | .206    |
| trait                           | 6.86, 3121.58    | 809.43  | 134.91      | .049  | <.001   |
| Discipline Prolific:trait       | 6.86, 3121.58    | 809.43  | 4.40        | .002  | <.001   |
| species                         | 3.52, 1602.37    | 1742.36 | 1287.30     | .352  | <.001   |
| Discipline Prolific:species     | 3.52, 1602.37    | 1742.36 | 0.29        | <.001 | .860    |
| trait:species                   | 29.68, 13506.48  | 266.54  | 89.76       | .047  | <.001   |
| Discipline Prolific:trait:species| 29.68, 13506.48  | 266.54  | 1.60       | <.001 | .021    |

---
## Plot: Discipline

---
## Plot: Traits

---
## Plot: Species

---
## Network Invariance: Comparing across species

**Network invariance**: some _p_ < .1; Dolphin - ant, p = .045; Duck - ant , _p_ = .056; Duck - Dolphin, _p_ = .051; Mouse - Duck, _p_ = .075, Mouse - Ant, _p_ = .045

--> None of these survive correction for multiple testing suggests we can broadly compare... . It's to be expected we'll find some different layouts given the different target species - but suggestive of roughly similar layout of domains.

**Strength invariance**:

Some of these are different at _p_ = .001 (Mouse vs. Ant; Duck vs. Chimp; Chimp vs. Ant).  This implies the strength of the edges between the nodes could vary between species.

<img src="duck_cute.png" width="250px" style="display: block; margin: auto;" />
---
## Ant - 🐜

---
## Duck - 🦆

---
## Mouse - 🐭

---
## Dolphin - 🐬

---
## Chimp - 🐒

---
## Overall picture

---
## Comparing biology / psychology students

They don't significantly differ:

**Network invariance**: _p_ = .457

**Strength invariance**: _p_ = .245

---
## Limitations

* Prolific sample

* Species chosen

* Strange task - anchored at humans

---
## Moving forward...

* Experts with same design.

* Just write the bloody paper ;).

---
## Any Questions?

[http://tvpollet.github.io](http://tvpollet.github.io)

Twitter: @tvpollet

---
## Acknowledgments

* I am greatly indebted to my collaborator(s). (Any mistakes are my own!).

* Funded by British Academy.

* You for listening!