Social Media Behavior Analysis
Inspired by and referencing George Hotz’s work
Social Media and Behavioral Engineering
The Modern Skinner Box
ATTENTION: This is a stream of consciousness exploration of:
- Anchoring
- The Availability Heuristic
- The Curse of Knowledge
- Confirmation Bias
- The Dunning-Kruger Effect
- Belief Bias
- Groupthink
- The Bystander Effect
The Platform Dilemma
Is it possible to create a policy that promotes fair and diverse platform participation while maintaining equitable evaluation methods? While algorithms optimized for financial returns based on third-party data may be efficient, we should examine whether uniform earning metrics serve all participants effectively.
When platforms implement new policies, participants observe and adapt their strategies, each seeking to optimize their approach within the given framework.
The Engagement Challenge
The focus on maximizing engagement can lead to strategies that may temporarily obscure or delay information sharing between participants.
Key Research Finding: In brief, Niv’s model states that the rodent should select the response rate that maximizes their average rate of food rewards while minimizing the effort of acting and the rewards potentially missed out on by inaction.
We applied a modified version of this reward learning model to the data from Instagram to directly test the prediction that social media engagement is driven by the same principles as animal reward learning.
This prediction was strongly supported. Qualitatively, our results showed that people spaced their posts in a way that maximizes the net rate of reward: they posted more frequently in response to a high rate of likes, and less frequently when they receive fewer likes. Quantitively, the reward learning model accounted for the precise distribution of posting times better than many alternative models.
The Game Theory of Social Media
Consider a scenario where a participant discovers strategic information that provides a competitive advantage. When this information includes the ability to predict other participants’ responses, it can create an imbalanced dynamic. This can lead to decreased engagement when other participants perceive unfair advantages, potentially resulting in platform abandonment and community discord.
The Corporate Reality
Social media companies (more accurately described as “digital advertising platforms”) operate within complex profit-driven structures. While regulatory oversight is important, we must learn to effectively utilize these tools despite their limitations. The opportunity cost of non-participation can be significant.
Current System Characteristics:
- Large advertising companies maximize data collection
- They employ sophisticated monetization strategies
- They utilize behavioral psychology in platform design
- They maintain significant influence over information flow
- In exchange, they provide:
- Information of varying relevance1
- Engagement-based feedback mechanisms
- While alternative approaches remain underexplored
The Current Paradigm:
- Big advertising companies extract everything they can from us
- They “rent seek and nickel and dime you on everything”
- They hire psychologists to manipulate our unconscious thoughts
- They maintain hegemony over perception
- In exchange, they provide:
- Useless information1
- New dopamine injections
- While people fail to realize other possibilities
Personal Reflections
I am curious about potential improvements to the current system. I envision content that’s more precisely tailored to individual preferences and needs, beyond simple algorithmic recommendations. There’s concern about how communities might respond when challenges become overwhelming, potentially leading to misattributed causation.
A perspective: I appreciate the innovations of market-driven systems, including new products, urban development, and technological conveniences. I value the efficiency of modern interfaces and content curation that helps filter out negative or irrelevant material.
A confession: I like capitalism, I enjoy new products, comfortable spaces in the city and luxury housing, I love to consume. It fills me with satisfaction when I don’t have to type in an entire query, or when I launch YouTube and don’t see it flooded with negative content.
Let me replace this with:
Personal observation: I appreciate the benefits of market-driven innovation, including technological advancement, urban development, and improved quality of life. The convenience of modern interfaces and intelligent content curation systems has notably enhanced user experience and productivity.
Reliability Theory and Trust Systems
The Evolution of Trust
Research on rodents from 80 years ago has evolved significantly, with mechanisms passing into human consciousness through both determined learning and osmosis. Despite scandals and revelations of harm, the system persists and accelerates through constant crises.
Modern Paradox: When a service worker doesn’t smile, they face immediate reprimand, yet we celebrate and financially reward those who simply perform rehearsed behaviors on camera.
The Theory of Credibility
The phenomenon of sensitization and desensitization is real. Many decisions rely on trusting strangers, forcing consumers to evaluate visions developed by specialized teams under specific management.
Key Theoretical Framework
The Theory of Credibility attempts to model situations that arise when individuals are uncertain about whether to trust the people they interact with. In this context, reputations for reliability hold significant value. An agent becomes credible by consistently providing accurate, valuable information or by always acting responsibly.
I analyze a series of simple, abstract models of information transmission. In the basic model, there are two agents: the Sender (S) and the Receiver (R), who will engage in the game a known, finite number of times. At each stage, both players learn the value of a parameter measuring the importance of that period’s play of the game (i.e., the value of making a correct decision), and S obtains information that is relevant to payoffs. S then sends a signal to R, who makes a decision that affects the welfare of both players. After both players learn the consequences of R’s decision, the process repeats. At each stage, R must decide what action to take, which requires assessing the credibility of S. The difficulty arises because R is uncertain about S’s preferences when the play begins.
I model this by assuming that, with positive probability, S has identical preferences to R (S is a friend), and with positive probability, S has completely opposed preferences to R (S is an enemy). This uncertainty creates a trade-off for enemy Senders. At each stage, providing honest information enhances S’s reputation but foregoes an opportunity for immediate gain by deceiving R. The possibility that S is a friend lends credibility to what he says, but this credibility may also be exploited by an enemy S.
The principal equilibrium has the following form. In the basic model, where S has no control over the importance of a given play, S typically conveys accurate information for the first several periods. R raises the probability he places on S being friendly after receiving accurate information if he believes that an enemy would have lied to him with positive probability. The more important the information, the more likely an enemy will attempt to deceive R. An enemy will eventually take advantage of R by misleading him. When this happens, the Sender loses all opportunities for deception in the future.
So why is this happening?
BEEP BOOP ONE MORE MINUTE
CHAT GPT
In the context of social media, the “Skinner cage” technique would refer to the use of behavioral reinforcement mechanisms to influence users and increase their engagement with the platform. Social media use a variety of techniques to keep users on their sites for as long as possible, as more active users translate into more advertising revenue.
The most common techniques are:
Rewards and reinforcements: Social media platforms often use mechanisms to reward users, for example by displaying likes, comments, or the like. These positive interactions reinforce feelings of satisfaction and encourage further activity.
Notifications and sounds: Social media uses notifications to grab a user’s attention, making them want to respond quickly to interactions with the site.
Personalization and algorithms: Platforms often analyze user data to personalize content and notifications. Algorithms draw conclusions about user preferences and present them with content that is more likely to pique their interest.
“Getting into the habit.” Social media is designed to keep users in the habit of repeatedly checking messages, scrolling through the page, or looking at the platform in their spare time.
Restricting access: Some features may be available only to users who have reached a certain level of activity or have spent a lot of time on the platform, which may encourage them to stay on the site longer.
Some of these techniques may be related to “Skinner’s cage” theory, especially when it comes to using positive reinforcement to encourage users to stay longer and use social media more regularly. It’s important to be aware of how these mechanisms affect our behavior and what consequences they can have on our mental health and time spent online. Being aware of these techniques can help us use social media more responsibly.
NATURE Social media as a modern-day Skinner Box?
Our computational reward learning model of social media engagement predicts that people maximize their average rate of social rewards - receiving likes - while minimizing the effort of posting and the rewards potentially missed out on by inaction. For details, see our paper in Nature Communications.
Because it is difficult to determine causality with observational data, we, together with PhD student David Schultner, decided to run an online experiment that mimicked key features of social media. Participants could post funny images with phrases (“memes”), and receive likes as feedback, on an Instagram-like platform as many times as they wanted during 25 minutes. Critically, we manipulated the average amount of likes the participants received and found that people posted more often when they received more likes. In other words, the experiment verified the causal influence of social rewards on how often people post, supporting our conclusion that social reward rates shape real social media behavior.
Together, our results establish that social media engagement follows basic, cross-species principles of reward learning. In extension, these findings may help us understand why social media comes to dominate daily life for many people and provide clues, borrowed from research on reward learning, to how maladaptive online engagement may be addressed.
PNAS
Significance Our brains draw on memories to predict the future; when our predictions are incorrect, we must update our memories to improve future predictions. Past studies have demonstrated that the hippocampus signals prediction error (i.e., surprise) but have not linked this neural signal to memory updating. Here, we uncover this missing connection. We show that mnemonic prediction errors change the role of the hippocampus, reversing the relationship between hippocampal activation and memory outcomes. We examine the mechanisms of this shift in neural processing, showing that prediction errors disrupt the temporal continuity of hippocampal patterns. We propose that prediction errors disrupt sustained representations and enable memory updating. Our findings bear implications for improving education, understanding eyewitness memory distortion, and treating pathological memories.
The brain supports adaptive behavior by generating predictions, learning from errors, and updating memories to incorporate new information. Prediction error, or surprise, triggers learning when reality contradicts expectations. Prior studies have shown that the hippocampus signals prediction errors, but the hypothesized link to memory updating has not been demonstrated.
SEARCH PROPOSALS AT THIS TIME ON THE SEARCH ENGINE.
The Science Behind Social Media Behavior
PNAS Research Insights
Key Finding: Our brains draw on memories to predict the future; when our predictions are incorrect, we must update our memories to improve future predictions.
Research Highlights:
- The hippocampus signals prediction error (i.e., surprise)
- Mnemonic prediction errors change the role of the hippocampus
- Prediction errors disrupt the temporal continuity of hippocampal patterns
- These disruptions enable memory updating
Implications: Our findings bear significance for:
- Improving education
- Understanding eyewitness memory distortion
- Treating pathological memories
The brain supports adaptive behavior through three key mechanisms:
- Generating predictions
- Learning from errors
- Updating memories to incorporate new information
Social Media as a Modern Skinner Box
Core Mechanisms of Engagement
- Rewards and Reinforcements
- Likes
- Comments
- Social validation
- Attention Capture
- Push notifications
- Sound alerts
- Visual cues
- Personalization
- Content customization
- Preference learning
- Interest targeting
- Habit Formation
- Regular checking
- Scroll behavior
- Time investment
- Access Control
- Feature gating
- Activity requirements
- Engagement rewards
Research Resources
Research Papers and References Bias Analysis Experimental Methods Persuasion Theory Modern Behavioral Science Game Theory Applications Neural Studies Cognitive Paradigms Psychological Research Statistical Analysis
