+ - 0:00:00
Notes for current slide
Notes for next slide

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

Thomas Pollet & Jeanne Bovet
Northumbria University

()

2025-02-12 | disclaimer

1 / 40

Today

  • Work in progress.

  • Largely exploratory project.

2 / 40

Back story

  • Grant on Evolutionary Psychology text books and 'tokenism' (and perceptions).
3 / 40

Back story

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

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

3 / 40

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.

3 / 40

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: β€˜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.’.

3 / 40

Scorpionflies...

Preconceptions about similarity to humans

  • What are students' preconceptions about a species's similarity to humans?
4 / 40

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.

4 / 40

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.

4 / 40

This project

  • How do students perceive similarity of species to humans - and what is the role of different domains (e.g., Diet vs. sociality)?
5 / 40

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

5 / 40

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.

5 / 40

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.

5 / 40

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.
6 / 40

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%).

6 / 40

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.

6 / 40

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.

6 / 40

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)
7 / 40

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).

7 / 40

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

7 / 40

Psychonetrics

  • As an input we take the correlations between items. The method we use allows picking the appropriate correlation method (continuous, ordinal or tetrachoric).
8 / 40

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.

8 / 40

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.

8 / 40

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 for a tutorial.

8 / 40

Clique percolation method (CPM, Lange, 2011)

  • Many algorithmic methods exist to find structure in networks ("clusters"/"modules"/"communities"): Louvain method (Blondel et al., 2008) / Leiden method (Traag et al.2019)
9 / 40

Clique percolation method (CPM, Lange, 2011)

  • Many algorithmic methods exist to find structure in networks ("clusters"/"modules"/"communities"): Louvain method (Blondel et al., 2008) / Leiden method (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

9 / 40

Clique percolation method (CPM, Lange, 2011)

  • Many algorithmic methods exist to find structure in networks ("clusters"/"modules"/"communities"): Louvain method (Blondel et al., 2008) / Leiden method (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.

9 / 40

Clique percolation method (CPM, Lange, 2011)

  • Many algorithmic methods exist to find structure in networks ("clusters"/"modules"/"communities"): Louvain method (Blondel et al., 2008) / Leiden method (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.

Thierry Dugnolle, CC BY-SA 4.0 <https://creativecommons.org/licenses/by-sa/4.0>, via Wikimedia Commons

Thierry Dugnolle, CC BY-SA 4.0 , via Wikimedia Commons

9 / 40

'Invariance testing'

Using a network comparison test, we can compare network structures between groups. (van Borkulo et al., 2023).

--> Permutations

10 / 40

Results

  • Some descriptions, (exploratory) t-tests, (exploratory) signed rank tests, and (exploratory) ANOVAs.
11 / 40

Results

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

  • Mostly graphs.

11 / 40

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.
12 / 40

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.

12 / 40

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)

12 / 40

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 Ξ·Β² 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
13 / 40

generalised eta squared

Plot: Discipline

14 / 40

Plot: Traits

15 / 40

Plot: Species

16 / 40

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.

17 / 40

Psychonetrics: Overall

18 / 40

Invariance: Biology vs. Psychology

They do differ:

Network invariance: p = .0009

Strength invariance: p = .609

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

19 / 40

Biology

20 / 40

Psychology

21 / 40

Interim conclusion

  • Clustering of traits: "learning, social, cognition" and "genetics, physiology, physical anatom"
22 / 40

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.

22 / 40

Study 2: 🐜, πŸ¦†, 🐭, 🐬, πŸ’

  • Repeat the exercise with: 🐜, πŸ¦†, 🐭, 🐬, πŸ’.
23 / 40

Study 2: 🐜, πŸ¦†, 🐭, 🐬, πŸ’

  • Repeat the exercise with: 🐜, πŸ¦†, 🐭, 🐬, πŸ’.
  • Final n = 470. 228 men, 227 women, 15 non-binary or third gender. Approx. 80%. from UK and US.
23 / 40

Study 2: 🐜, πŸ¦†, 🐭, 🐬, πŸ’

  • Repeat the exercise with: 🐜, πŸ¦†, 🐭, 🐬, πŸ’.

  • 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)

23 / 40

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.
24 / 40

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.

24 / 40

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).

24 / 40

Mixed model ANOVA (exploratory)

Effect df MSE F Ξ·Β² 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
25 / 40

Plot: Discipline

26 / 40

Plot: Traits

27 / 40

Plot: Species

28 / 40

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.

29 / 40

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.

29 / 40

Ant - 🐜

30 / 40

Duck - πŸ¦†

31 / 40

Mouse - 🐭

32 / 40

Dolphin - 🐬

33 / 40

Chimp - πŸ’

34 / 40

Overall picture

35 / 40

Comparing biology / psychology students

They don't significantly differ:

Network invariance: p = .457

Strength invariance: p = .245

36 / 40

Limitations

  • Prolific sample
37 / 40

Limitations

  • Prolific sample

  • Species chosen

37 / 40

Limitations

  • Prolific sample

  • Species chosen

  • Strange task - anchored at humans

37 / 40

Moving forward...

  • Experts with same design.
38 / 40

Moving forward...

  • Experts with same design.

  • Just write the bloody paper ;).

38 / 40

Any Questions?

http://tvpollet.github.io

Twitter: @tvpollet

39 / 40

Acknowledgments

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

  • Funded by British Academy.

  • You for listening!

40 / 40

Today

  • Work in progress.

  • Largely exploratory project.

2 / 40
Paused

Help

Keyboard shortcuts

↑, ←, Pg Up, k Go to previous slide
↓, β†’, Pg Dn, Space, j Go to next slide
Home Go to first slide
End Go to last slide
Number + Return Go to specific slide
b / m / f Toggle blackout / mirrored / fullscreen mode
c Clone slideshow
p Toggle presenter mode
t Restart the presentation timer
?, h Toggle this help
oTile View: Overview of Slides
Esc Back to slideshow