NetLogo banner

Home
Download
Help
Forum
Resources
Extensions
FAQ
NetLogo Publications
Contact Us
Donate

Models:
Library
Community
Modeling Commons

Beginners Interactive NetLogo Dictionary (BIND)
NetLogo Dictionary

User Manuals:
Web
Printable
Chinese
Czech
Farsi / Persian
Japanese
Spanish

  Donate

NetLogo User Community Models

(back to the NetLogo User Community Models)

[screen shot]

Download
If clicking does not initiate a download, try right clicking or control clicking and choosing "Save" or "Download".

Try It in NetLogo Web

## WHAT IS IT?

This is an implementation of the Hierarchical Ising Opinion Model (HIOM).

van der Maas, H. L. J., Dalege, J., & Waldorp, L. J. (2020). The polarization within and across individuals: The hierarchical Ising opinion model. Journal of Complex Networks, 8, cnaa010.

The opinion of agents is modeled by a stochastic (cusp) differential function in which attention and information are the control variables of the cusp. Agents interact and influence both attention and information. Due to hysteresis within agents, this leads to new explanations of polarization.

The cusp model of the individual is derived from the Ising attitude model, described in

Dalege, J., & van der Maas, H. L. J. (2020). Accurate by being noisy: A formal network model of implicit measures of attitudes. Social Cognition, 38, S26-S41.

Dalege, J., Borsboom, D., van Harreveld, F., & van der Maas, H. L. J. (2019). A network perspective on attitude strength: Testing the connectivity hypothesis. Social Psychological and Personality Science, 10, 746-756.

Dalege, J., Borsboom, D., van Harreveld, F., & van der Maas, H. L. J. (2018). The Attitudinal Entropy (AE) Framework as a General Theory of Individual Attitudes. Psychological Inquiry, 29, 175-193.

Dalege, J., Borsboom, D., van Harreveld, F., van den Berg, H., Conner, M., & van der Maas, H. L. J. (2016). Toward a formalized account of attitudes: The Causal Attitude Network (CAN) model. Psychological Review, 123, 2-22.

## HOW IT WORKS

The algorithm has been slightly changed
and optimized compared to the original 2020 paper:

Iterate

* Randomly choose a set of agents, weighted with attention A (modified eq. 5).
* Iterate over these agents.
* Randomly choose a neighbour of the agent i as partner j in the interaction.
* Add attention to both agents i and j (modified eq.6a).
* Exchange information (modified eq.7).
* Apply decay in A to all agents (modified eq. 6b).
* Add noise and apply decay in I to all agents (modified eq. 7b).
* Update opinion O in all agents (eq. 4).

(4) dO<sub>i</sub> = -(O<sub>i</sub>^3-(A<sub>i</sub>+A_min)O<sub>i</sub>-I<sub>i</sub> )dt+s<sub>O</sub> dW<sub>i</sub> (t)
(5) Pr(select agent<sub>i</sub> )=A<sub>i</sub>
(6a) A<sub>i</sub>= min⁡(1,A<sub>i</sub>+ d<sub>A</sub>)
(6b) A<sub>i</sub>= decay<sub>A</sub> A<sub>i</sub>
(7a) I<sub>i</sub>=r I<sub>i</sub>+(1-r) I<sub>j</sub>, where r=rmin+(1-rmin)/(1+e^(-p(A<sub>i</sub>-A<sub>j</sub>)) )
(7b) I<sub>i</sub>=decay<sub>I</sub> (I<sub>i</sub>+Ν(m<sub>I</sub>,s<sub>I</sub>))

Where dt = .01, Amin = -.5

O is opinion, A is attention, I is information

and parameters (set by sliders):
m<sub>I</sub> = mean noise information (external field)
s<sub>I</sub> = sd noise information
dA = attention increase by interactions
p = persuasion
rmin = minimal resistance
decay<sub>I</sub> = decay information
s<sub>O</sub> = sd noise opinion

## HOW TO USE IT

SETUP sets the agents to values determined by mean-init-information, sd-init-information, mean-init-attention, sd-init-attention.
GO starts the simulation.

The sliders set parameters of the main equations for attention, information, and opinion (see HOW IT WORKS).
The %active-agents slider influences how many agents will be active in each interaction. %active-agents affects the decay of attention (decay_I).

Visualize can be set to opinion, attention, information.

The histograms show the distribution of opinion, attention and information.
The mean plot shows the change of the mean of opinion, attention of information.

setup-black-pete is used to set the parameters to the value required for the Black-Pete simulation described in the HIOM 2021 paper. In this scenario, activists are added to a conservative community with low attention. If attention grows too quickly, polarization will result. Click on add-activist to allow activists to be added with mouse clicks.

setup-meat-eating-vegetarian is used to set the parameters to the value required for the meat-eating vegetarian simulation described in the HIOM 2021 paper. In this simulation, the bound is set low, meaning that agents that are too different will not interact.
By clicking Perturb-Activists, some activists are perturbed to the meat-eating opinion, but their information position does not change. They can now influence the meat-eaters, which leads to a slow change to the vegaterian position.

## THINGS TO NOTICE

Polarization due to hysteresis (simulation 1 in the HIOM paper):

In the standard setup, you can see that high attention (set %active-agent high) leads to polarization. If you now decrease the difference in information (set decay_I to .5), the polarization remains even if the difference in underlying information is 0. Only if you also decrease attention (set %active-agents low), the polarization disappears.

Opposition to activism: the case of Black Pete (simulation 2 in the HIOM paper):
See HOW TO USE IT

A solution to polarization: the meat-eating vegetarian (simulation 3 in the HIOM paper):
See HOW TO USE IT

## THINGS TO TRY

The mean noise information function as an external field. Test if varying this paramter gives hystersis at the level of the mean opionion over agents.

Test the robustness of the 3 simulations by adjusting model parameters with the sliders.

In the Black Pete simulation activist are more succesful when attention changes slowly. Adjust da and %active-agents to test this.

In the meat-eating vegetarian simulation, the pertubation only works when attention is relatively low. Test this by setting %active-agents high.

## EXTENDING THE MODEL

Set-up the model in a preferential atttachment network

## CREDITS AND REFERENCES

van der Maas, H. L. J., Dalege, J., & Waldorp, L. J. (2020). The polarization within and across individuals: The hierarchical Ising opinion model. Journal of Complex Networks, 8, cnaa010.

(back to the NetLogo User Community Models)