Social Psychology Research

How do Instagram "likes" influence politics?

Overview 1/3

Abstract

Social media serves as a platform for informing and sharing societal issues, with acceptance reflected in post "like" counts. This study investigates how these "like" metrics influence user sharing behavior and perceptions of climate change.

Participants viewed Instagram posts with varying like counts (high, low, or none) and completed a survey assessing their likelihood to engage and the perceived credibility of the posts.

While one-way ANOVA tests did not show significant relationships, further analysis revealed that high like counts increased user engagement and credibility perceptions. Additionally, participants' prior attitudes toward climate change affected their views on the like counts.

This study aims to enhance understanding of how social engagement metrics shape public discourse on critical issues like climate change. Read full paper here

TIME

3 weeks

DATES

May 2024

TOOLS

Qualtrics

TEAM

Chenchen Liu
Jose Rosado
Joyce Sitt

Overview 2/3

Context & Background

Social media has become a vital tool for communicating societal views on political issues and quickly consuming news, with over half the global population (62.6%) using it for an average of two hours and twenty minutes daily.

In particular, climate change has become a prominent topic on social media, with a recent poll showing that 56% of 14 to 18-year-olds learn about it through these platforms, suggesting that both the content of posts and the number of "likes" and comments significantly influence public opinion on the issue.

Theoretical Framework: The ELM Model

The Elaboration Likelihood Model (ELM) examines how both deep cognitive processing and superficial cues, such as post "likes", influence user perceptions and engagement with climate change content, while considering the user's existing concern and cognitive engagement with the topic.

Existing Research

Previous research in China has shown that content likeability positively influences social media engagement and credibility, particularly in mobile social networks, but its focus on product placement and a predominantly Chinese user base limits generalizability. Therefore, we propose collecting data from a more diverse group of Instagram users in the United States to better reflect the platform's global audience.

Overview 3/3

Hypotheses

Hypothesis 1

Instagram posts about climate change with higher like counts will lead to increased user engagement, enhanced perceptions of truthfulness, and potentially influence changes in concern for climate change among users.

Hypothesis 2

Individuals who already possess a high level of concern for climate change may exhibit less variability in their responses based on like counts, suggesting a potential moderation effect of pre-existing attitudes towards the issue

Methods 1/8

Participants

We recruited 28 participants for an online study by posting invites to social media web pages and direct messaging through Instagram, Snapchat, GroupMe, and iMessage. This initial sample size was established based on practical and budget limitations.

From this initial pool, we excluded 10 participants who failed the stimulus check item.

Methods 2/8

Procedure & Materials

This study was a between-subjects experimental design with 3 levels (Like Metrics: high, low, none). We assessed participants’ perceptions regarding: (1) user engagement and (2) perceived credibility.

Informed consent

Participants were first informed about the study to prevent them from being coerced or misled into agreeing to something they don't fully understand.

Demographics

Participants then completed a questionnaire collecting information on 6 different demographic variables to understand subgroup analyses.

Age
Gender Identity
Ethnicity
Education Level
Social Media Usage
Content Engagement

Methods 3/8

Cognitive Assessment

Determined each participant’s current attitudes surrounding climate change, as an aspect of our hypothesis is grounded in the idea that individuals who care more about climate change may not care about like metrics.

Methods 4/8

Attention Check

Participants were given an attention check item to exclude those who answered the survey carelessly (N=0).

Methods 5/8

"Like" Count

Participants were randomly assigned to view one of three social media post conditions regarding climate change with varying "like" counts (high, low, control). Each group viewed the same 3 posts with varying "like" counts.

Methods 6/8

Perceived Credibility

For the perceived credibility items, participants rated the post they were given on perceived truthfulness. This was done to assess how the participants perceived the credibility of each post that varied in "like" count.

Methods 7/8

User Engagement

Participants rated each post based on how likely they were to like, comment on, or share it. The questions were designed to assess how "like" counts can influence the participant’s desire to engage with a post.

This tested our hypothesis that the posts with higher "like" counts were more likely to be engaged with than the posts with lower "like" counts.

Methods 8/8

Stimulus Check

We presented a multiple-choice question that read “According to the last post you viewed, what is climate change killing?” with 5 options:

Polar Bears
The Arctic (correct response)
The Earth
People
Marine Life

Participants who answered incorrectly to this question were excluded from the study (N=10).

Overview
Methods
Results
Discussion

Results 1/5

Association Between Participant Cognition Assessment Scores and Perceived Credibility Ratings

We conducted a linear regression analysis to predict the outcome on participants' rating of the perceived credibility of the social media posts based on their cognitive assessment scores on climate change.

We found no evidence that participants’ cognition assessment scores predicted participants’ perceived credibility ratings for the social media posts, b = .09, t(16) = 0.32, SE = .30, p = .756.

Results 2/5

Association Between Participant Cognitive Assessment Scores and User Engagement Ratings

We conducted a linear regression analysis to predict the outcome on participant’s ratings of likelihood to engage with the social media posts based on their cognitive assessment scores on climate change.

We found no evidence that participants’ cognitive assessment scores significantly predicted participants’ user engagement ratings for the social media posts, b = .53, t(16) = 1.41, SE = .38, p = .177.

Results 3/5

Association of Post Like Count Metrics on Perceived Credibility Ratings

We conducted a one-way ANOVA to evaluate whether the like count metrics of a post (i.e., high, low, no metrics) influenced the perceived credibility of a social media post. We found no evidence of an association of like count on perceived credibility ratings, F(2, 15) = 2.51, p = .115.

We then followed up with a Tukey post-hoc test to determine which pair of like count metrics there was a difference amongst. We found the greatest difference in perceived credibility ratings between posts with high like metrics and no like metrics, p = .100, and less differences between no like metrics and low like metrics, p = .453, and high like metrics and low like metrics, p = .489.

Results 4/5

Association of Post Like Counts on User Engagement Ratings

A one-way ANOVA showed no significant effect of like counts on user engagement with climate change-related Instagram posts, F(2, 15) = 3.36, p = .063.

However, post-hoc comparisons indicated that participants were slightly more likely to engage with posts having high like counts compared to those with low or no like counts. High like count posts had higher engagement than low like count posts (p = .350) and control posts (p = .052). Low like count posts also saw higher engagement than control posts, though less pronounced (p = .395).

Results 5/5

Summary of Findings

The tests did not show any significant effects. However, post-hoc comparisons indicated the following results:

Credibility

HIGH > LOW > CONTROL

Engagement

HIGH > LOW > CONTROL

Discussions 1/4

Findings Overview

Despite our hypothesis that higher like counts would lead to increased perceived credibility and user engagement, the results did not show strong significant effects. However, some patterns in the follow-up analyses suggested differential effects of like counts on user engagement and perceived credibility, providing nuanced insights into social media dynamics.

Discussions 2/4

Contribution to Theory

Echo Chamber Effect

The limited effect of "like" counts may stem from climate change's polarizing nature. Users may distrust high "like" counts, attributing them to bots or manipulation, reflecting echo chamber dynamics. Our study suggests "likes" have little influence on opinions in polarized contexts, highlighting issues like confirmation bias and selective exposure.

"Likes" Suggest Engagement

Although "likes" didn’t significantly boost engagement, posts with high "like" counts showed a slight increase in interaction. This suggests "likes" may subtly influence users' decisions to engage, aligning with research that shows "likes", while not decisive, can contribute to a cumulative impression favoring interaction.

These findings highlight the complexity of influencing public opinion on social media, especially for polarized issues. The study emphasizes the need for a nuanced understanding of how features like "like" counts interact with psychology to shape behavior and beliefs.

Discussions 3/4

Limitations

Varied Participant Commenting Behavior

‍Our study did not account for the idea that not all users engage with posts by commenting or sharing. Oftentimes, those who do not comment or share on a normal basis will not comment or share even for a post they like, which decreases the validity when finding an effect on user engagement.

Narrow Participant Diversity

Due to constraints regarding research resources, this study primarily gathered data from students in the United States who were exclusively exposed to simulated posts linked with Instagram. Hence, the study's findings may not be broadly applicable to other countries or alternative social platforms. Varying results may occur due to differences in social norms and influences; however, further studies are recommended to be done to investigate other social media platforms in other countries.

Misrepresentation of Real World Effects

In our study, we explicitly instructed participants to “pay attention to all aspects of the post (text, imagery, like count, etc.),” aiming to isolate the effect of like counts. However, in real-life scenarios, users may not direct their attention as deliberately to the number of likes. By directing participants to pay attention to likes, we may have disrupted the natural effects of these metrics, leading to results that do not reflect typical user behavior.

Discussions 4/4

Future Directions

Future studies should further investigate the significance of social media platforms and “like” metrics. Since the control group rated lowest in engagement and credibility, it appears that "likes" influence information processing to some extent. Exploring this could reveal the psychological mechanisms behind user attitudes and behaviors.

Additionally, examining how user demographics and contexts interact with "likes" may provide insights into the broader implications of social media engagement on information dissemination.