NetLogo banner

 Contact Us

 Modeling Commons

 User Manuals:


NetLogo User Community Models

(back to the NetLogo User Community Models)

Model 2 - Social Conception of content

by Noam Drory (Submitted: 08/09/2017)

[screen shot]

Download Model 2 - Social Conception of content
If clicking does not initiate a download, try right clicking or control clicking and choosing "Save" or "Download".

(You can also run this model in your browser, but we don't recommend it; details here.)


This model studies how individuals on a network affect one another in regard to content delivery, and particularly, offensive content. each interaction of an agent with its neighbor, can, with some probability, cause the neighbor to change his or hers opinion and action regarding the content (Namely, being indifferent to it, forward it without taking a stand, or respond positively or negatively to it). Some agents, have fixed opinions and cannot be moved. Nevertheless, if an agent isn't fixed, it is harder to change his stand if his mind is set to a strong stand (Positive or Negative) rather than to a soft one (Indifferent or Distributer).


Each agent can have one of the opinions "I", "D", "P" or "N" (which stands for Indifferent, Distributer, Positive or Negative). For each tick of the simulation, one agent is randomly selected to be the speaker. This speaker then interacts with one of its neighbors, called the listener. once the speaker's voice is heard, he has a chance of influencing its neighbor and turn him to his opinion. If the speaker has a solid opinion (Namely positive or negative) and his neighbor has a soft one (indifferent or distributer), he then has 20% chance to convert him. on any other scenario the listener's chances of conversion are 5%. If the listener has a fixed opinion, the chances of conversion are 0%. Fixed opinion can be either positive or negative. The simulation will continue until the "system" is stabilized, meaning there were no changes on the last period.


Several parameters should be set prior to execution. The number of agents using the "agents" variable, the offensiveness of the content using "offensiveness" vaiable, the significant period of stability using the "stability-factor" variable and the precentage of commited agents using the "committed-agents-precent" vaiable. after "Setup" was made, one can adjust the network graph using one of the "Keep spreading agents" button.

The setup will create a graph with a number of agents according to the "agents" variable. with precentage of commited agents according to the "committed-agents-precent" variable. The number of "Negative" opinion holders is correlated to the offensiveness of the content which is defined by the "offensiveness" variable.

Once the "Go" function is started, agents will be picked randomly and will try to influence a random neighbor as discribed on previous chapter. The simulation will stop once the system is stabilized, which means, there were no changes on the last "significant period" - meaning on the last number of ticks as defined on the "stability-factor" variable.

A user can also run the simulation one tick at a time - using the "Go once" function, and thus, track the persuation efforts (Speaker's voice, Listener opinion and casted probability of influence) of each interaction.


Notice what happens when offensiveness is increased or decreased, and how the graph is stabilized on different stability factors - one can give the stability factor different interpretations, such as, time, the intensity of the content etc. Even more so, different interpretations could be given to the offensiveness scale. such as, how offensive is the content in the mind of the first distributer (or the content creator himself), what are the social norms of the network regarding these types of contents and more.


In order to make the model more realistic, few changes could be considered, such as:
1. Adding probability in which even a fixed speaker can be shifted from his opinion.
2. Opinion shifts generaly made more gradually. A model could take that into account.
3. In each interaction, both of the opinions could change and not just of the listener.
4. Opinion shift doesn't have to be to the one of the speakers, but to a more subtle one of the listener (e.g. Negative->Indifferent->Positive instead of Negative->Positive).
5. There are people that can be fixed on being only Distributers or Indifferent agents - it should fit better to the case of not-totally-fixed opinions as described in remark 1.
6. There is room for more of an academic study regarding the precentage of "fixed" opinions, so their distribution won't be totally random.


Social Consensus - by Daniel Diermeie (

(back to the NetLogo User Community Models)