How Social Media Posts Affect People

Running head: SOCIAL COMPARISON 1

SOCIAL COMPARISON 37

Social Comparison and Priming: How Social Media Posts Affect People’s Self Perception

Student Name

Florida International University

Abstract

The social comparison theory explains how people’s self-views are affected through comparisons with others, while priming effects show how people’s perceptions change when they are informed or given warnings about something they are about to do. Among our two studies, we used 179 and 547 participants, respectively, consisting of undergraduate students and their family and friends. We analyzed individuals’ perception of themselves when exposed to either upward, downward, or an average social comparison condition (study one, conducted in person ), and then again when given a warning beforehand (study two, an online survey). We predicted that individuals in the upward condition would feel worse about themselves, while those in the downward condition would feel better. Warnings beforehand would make individuals score higher in positive self-esteem, with those who received no warning scoring higher in negative self-perception. Participants in the upward condition did report higher scores in negative self-views, while those in the downward condition showed higher self-esteem. However, priming effects showed no difference in individuals’ self-perceptions when it came to upward or downward social comparison. These results suggest that although social comparison can have an impact on people’s view of themselves, forewarnings have no effects when it comes to people’s self-views.

Keywords: social comparison, priming effects, forewarning, self-perception, self-esteem.

Social Comparison and Priming: How Social Media Posts Affect People’s Self Perception

It is often acknowledged that people choose to post on social media only the positive things that happen to them. Social media consumers might use how others are doing and what they are doing as a rule of thumb in order to reflect on how they should be doing in their own lives. This can be detrimental in the long run because most people do not necessarily post their hardships on social media, as there is a common social desire to always display the self in a positive manner (Liu et al., 2016). While 30 years ago our social circle might have included only a few selected people, today we are exposed to thousands of individuals that are just one click away, and many of them share aspects of their lives online. Additionally, since most of those people tend to be friends and family, the potential for social comparison is much higher when information about their lives is so easily available (Liu et al., 2016). The main focus of this study is to analyze how different social comparison conditions can alter a person’s view of their own aptitudes and self-esteem.

The social comparison theory explains how people develop a sense of opinions, abilities, personality, and self-esteem through comparing themselves with others around them (Festinger, 1954). This comparison can therefore affect people’s expectations about themselves. The two types of social comparison conditions in this study are upward and downward social comparison. Downward social comparison is defined as when people compare themselves to those that are in a lower position than them; in other words, they are ‘looking down’ at the other person, which in turn makes them feel superior; while upward social comparison is defined as the opposite, when people are comparing themselves to someone they see as better than them, so they are ‘looking up’ to them. This latter type of social comparison usually results in more negative feelings from people (Wang et al., 2017), as it implies that others are better off than them.

The types of posts others make can affect people directly when they decide to compare themselves to those they see on social media, and studies have shown that social media posts have the ability to influence people’s emotions. In one particular study, Liu and colleagues (2016) showed that Facebook users tend to engage in upward social comparison when exposed to posts from distant friends rather than close friends, which in turn generates more negative emotions. Especially in the case of individuals with a high self-esteem, they experienced negative emotions when reading positive posts from distant friends, meaning they engaged in upward social comparison. Additionally, those individuals reported experiencing positive emotions when they read negative posts, meaning they engaged in downward social comparison. These findings relate to our study because they provide an insight into how looking down at others through downward social comparison can make people feel better about themselves.

Exposure to mostly upward social comparison on social media can have short term as well as long term effects on a person’s self-esteem and self-evaluation (Vogel et al., 2014). Moreover, people who spend greater amounts of time making social comparisons in social media in turn have higher symptoms of depression, higher self-discrepancy, and lower overall wellbeing. Results from the study showed that people who reported spending more time on Facebook engaged in more upward than downward social comparison, and in turn revealed to have lower self-esteem. Other studies have also found that upward social comparison can have a direct effect with symptoms of depression, especially on individuals that fall low on levels of optimism (Liu et al., 2017).

Some say that adolescents and young adults represent a demographic that is affected by early exposure to all the forms of social media available, and this exposure can have strong effects on individuals’ vulnerability at such a young age (Comer et al., 2008). In a study performed by Lee (2014), a sample of college students was chosen because they are more likely to engage in social comparison than adults, more susceptible to influence from peers, and use social media at higher rates. The study found that young people engage in social comparison as a form of self-evaluation, especially when it comes to topics like academic performance and identity development, which are at the forefront of their main priorities at that age. Their findings showed correlations between self-esteem and the frequency of social comparison, with a likeliness of people feeling bad about themselves when they engage in upward social comparison. Online profiles can additionally affect someone’s view of their own body image. In a study performed by Haferkamp and Kramer (2011), findings showed that looking at people deemed as attractive on social media can lead viewers to experience negative emotions and dissatisfaction with their own body. This all ties together with our studies because in order to feel bad about one’s own image after looking at someone deemed as attractive means that there is a degree of upward social comparison taking place.

To get some insight into how social comparison relates to social media posts, we designed a study that shows how other people’s posts on social media can affect a person’s own view of themselves. Participants in this particular study are college students, and they were asked to look at a stranger’s Facebook profile. Depending on the profile they saw, participants fell in one of three conditions, where they would either engage in a downward, upward, or average social comparison.

Our hypothesis was that participants in the condition for upward social comparison would score higher in negative self-perception than participants in the downward and average social comparison condition, so they would agree more with negative statements as: “I feel inferior to others at this moment”, and “I feel like I am not doing well”, while disagreeing more with positive statements. On the other hand, participants in the downward social comparison condition would score higher in positive self-esteem, as they would agree more with positive statements: “I feel good about myself”, and “I feel confident in my abilities” and disagree more with negative statements when compared to those in the upward and average social comparison condition.

Study One Methods

Participants

One hundred and seventy-nine participants took part on this study, out of which 91.6% (n=164) reported being students from Florida International University, while 8.4% (n=15) reported not being FIU students. Of these 179 participants 55.3% (n=99) were female and 44.7% (n=80) were male. Participants ranged in age from 17 to 59 with an average of 24.52 years (SD=7.71). Our sample consisted of 49.2% Hispanics (n=88), 24% Caucasians (n=43), 13.4% African Americans (n=24), 5% Asian American (n=9), 1.7% Native Indian (n=3), 5% Mixed race (n=9), and 1.7% Pacific Islander (n=3). Out of all our participants, 81.6% reported English as their first language (n=146), while 18.4% did not report their first language as English (n=33). Of these, 13.4% reported Spanish as their first language (n=24). See Appendix A.

Materials

Each of the three surveys contained five parts. In part one, participants looked at a fake Facebook profile of Pat Masters, an FIU student, along with some neutral general information, a made-up advertisement, a section for “Friends”, and four posts that Pat had recently made. Pat’s name was chosen as gender neutral, and the profile included a cover picture of FIU’s library, while the profile picture showed three people, one male and two females, so that participants did not rely on gender norms while completing the survey.

For each of the three surveys, the only difference was that two of Pat’s posts fell into a social comparison condition (either upward, downward, or average social comparison). In the condition for upward social comparison, Pat reported to have done extremely well in a job interview, claiming to have been considered the best candidate out of everyone else. In another post, Pat posts about receiving an A on a tough exam while most students in the class received a B. In the condition for downward social comparison, the job interview went really bad, stating he was not a good candidate, and Pat received a C on the tough exam while most students received a B. In the average social comparison, Pat states that he did ‘okay’ in the interview, being an ‘okay’ candidate, and he received a B on his tough exam while most students also received a B.

In part two of the survey participants had to rate their impressions of Pat Masters by reading seven statements and answering through an interval scale from 1 (strongly disagree) to 6 (strongly agree). Some examples of these statements include “Pat seems like a highly motivated person”, “Pat seems like a good job candidate”, “Pat seems like a likeable person”, etc. Part three of the study consisted of 13 statements where participants had to rate their level of agreement about themselves. Once again, an interval scale from 1(strongly disagree) to 6 (strongly agree) was used, and the statement included a mix of both positive self-statements: “I feel I am a very bright person”, “I feel I am a likeable person”, “I feel good about myself”, etc., and negative self-statements: “I feel inferiors to others at this moment”, “I feel like I have trouble understanding things”, and “I feel I am not doing well”. When creating our first dependent variable, we decided to call it ‘negative self-perception’, and we computed it using the mean score from participants’ negative self-views by selecting the following statements form part three of the survey: “I feel I am a lazy person”, “I feel inferior to others at this moment”, “I feel like I have trouble understanding things”, and “I feel like I’m not doing well”. The higher the score on this new variable, the more participants agreed about negative statements, meaning they felt worse about themselves. Our second dependent variable was called ‘positive self-esteem’, and it was based on the mean score for participants’ positive self-view by selecting the following statements: “I feel I am a very bright person”, “I feel I am (or will be) a good job candidate”, “I feel good about myself”, “I feel as smart as others”, and “I feel confident about my abilities”. A high score on this new variable would mean that participants felt good about themselves. Part four of the survey required to provide participants’ demographic information, which include gender, age, race or ethnicity, first language, their status as a student at Florida International University, and their current relationship status. In this part, participants were made aware of their right to leave any question blank if they did not feel comfortable answering it. Part five of the study consisted of a final question used as an attention check, which asked the participant to recall how Pat had done in the interview (great, bad, or okay) in order to tell if the participant was attentively completing the survey.

Procedures

Participants for this study were randomly chosen from the campus at Florida International University. They were approached in person and were orally asked if they were willing to participate in this research study. Participants were informed of how long the survey would take to complete, that they would experience no risks or discomforts from the study, and they were reminded that participation was voluntary. Once verbal consent was obtained, participants were randomly given one out of three surveys and they independently read and answered the questions on their questionnaire. It took them approximately 5 minutes to complete and they answered the questions on paper and pencil. The end of the study consisted of the debriefing of the participants, where they became aware of our hypothesis on the effect of social comparison in social media.

Study One Results

In this survey, social comparison condition (upward, downward, or average) was our independent variable and whether or not the participants recognized which condition they were in was our dependent variable. We then ran a chi square test as a manipulation check which showed significant results, X2(4) = 275.2, p < .001. Most participants in the upward social comparison condition reported that Pat did great in the interview (94.9%), most participants in the downward condition recalled he did bad (91.6%), and most of those in the average condition said he did ‘okay’ (88.3%). This shows that participants were paying attention when reading Pat’s posts on the survey. See Appendix B.

Our first analysis consisted of a One-Way ANOVA Test for social comparison condition (upward, downward, or average) as our independent variable and participants’ negative self-perception as our dependent variable. Results from this test showed a significant difference F(2, 176) = 25.7, p < .001 and Tukey post hoc tests showed that participants felt worse about themselves in the upward social comparison condition (M = 3.94, SD = 0.61) than in the downward social comparison condition (M = 3.12, SD = 0.74). Results showed that those in the average social comparison condition (M = 3.48, SD = 0.49) also felt worse about themselves than people in the downward condition. See Appendix C.

Our second analysis was another One-Way ANOVA Test, this time using social comparison condition (upward, downward, or average) as our independent variable and participants’ positive self-esteem as our dependent variable. Results showed a significant difference F(2, 176) = 45.7, p < .001. Tukey post hoc tests showed that participants felt better about themselves in the downward social comparison condition (M = 4.81, SD = 0.48) than in the upward (M = 4.08, SD = 0.38) social comparison condition. Also, those in the average (M = 4.48, SD = 0.38) condition felt better about themselves than participants in the upward social comparison condition. See Appendix D.

Study One Discussion

Our predictions were that participants in the upward social comparison condition would score higher in negative self-perception than participants in the downward and average social comparison condition, so they would agree more with negative statements about themselves and disagree more with positive statements compared to participants in the downward and average social comparison condition. In addition, we hypothesized that participants in the downward social comparison would score higher in positive self-esteem and lower in negative self-perception when compared to those in the upward and average social comparison.

Our results supported our hypothesis as we found that people in the downward social comparison condition were more likely to score higher in positive self-esteem, so they felt better when reading about the outcomes of someone who was doing bad, while those in the upward condition would feel worse about themselves when reading about someone who was doing great as they were more likely to score higher in negative self-perception. These findings support the idea behind other studies on the matter, which state that upward social comparison can negatively affect people’s views of themselves (Liu et al., 2016). To better understand social comparison, we wanted to dive deeper and see what would happen if participants were informed about the effects of social comparison before being administered the survey, so we decided to do some research on priming effects.

Study 2: Priming and Social Comparison

A lot of people consider knowledge to be power, and that once someone is aware of something, it is impossible not to let it affect how we react to everything else, especially if that knowledge is directly related to what we are doing in the moment. Many people believe to have a strong sense of will and independence when it comes to making their own decisions; however, priming research has shown that people’s actions and attitudes can be manipulated by presenting them with certain stimuli and concept which do not necessarily need to be in their conscious minds (Bargh et al., 1996). Our second study brings into question a new independent variable that includes priming, which will help us analyze how receiving a warning about social comparison effects before participating in the study will affect individuals’ perceptions of themselves when presented with different social comparison conditions.

Priming effects explain how a link between a stimulus and a prime can activate behaviors in memory, sometimes even by changing the self-concept of individuals (Loersch & Payne 2012). In a study by Eimer and Schlaghecken (2002), priming was compared against response inhibition, and their findings suggest that response inhibition happens only when primes cannot be identified. In other words, once individuals are aware of the fact that they are being primed it is harder for them to act indifferent to these and to resist distractions in the study. These results show that once knowledge about priming is acquired, it has a tendency to affect people’s perception and therefore the results of the study as well. This research can be connected with our own study because it suggests that receiving a warning beforehand might affect individuals’ reaction to their survey and by default provide different results than those from people that receive no warning.

Once something is learned, it is against human nature to act as if it never happened. Loersch and Payne (2012) showed that when participants were aware of the presence of subliminal messages, priming effects ceased to work. Giving people a warning seemed to raise their caution levels, which in turn stopped any subliminal effort from being effective. This phenomenon can be explained by the fact that being conscious about priming effects makes people develop strategies that make them process information with the purpose of the prime in mind (Lombardi et al., 1987), which would then make them immune to any attempt of subliminal manipulation. Other studies on subliminal advertising have found that simply telling participants that they were being subconsciously influenced was enough to erase the priming effects (Verwijmeren et al., 2013). In other words, just raising awareness on the presence of subliminal messages eliminated the effectiveness of priming on people. All of this research directly connects to our study because it shows how providing a warning before subjecting participants to a study can in a way render invalid the effects of that study altogether; as participants have adopted a different mindset than those who would otherwise receive no warning beforehand, they are in a way defending themselves from any potential prime that they could come across.

As priming can be found in many branches within the field of psychology and many aspects of individuals’ lives, it has also been seen in relationship with social comparison. For example, in a study performed by Dijksterhuis and colleagues (1998), participants were primed with thinking about Albert Einstein, which led the individuals to engage in upward social comparison and report themselves as less intelligent. The results of this study connect with ours because it shows the impact that priming can have on how people feel about themselves by subconsciously making them engage in upward social comparison when exposed to people who are considered smarter than them. This resulted in participants reporting themselves as less intelligent because upward social comparison has been shown to have effects on both self-evaluation and self-esteem (Vogel et al., 2014).

We have seen how social comparison can affect someone’s self-views, but what would happen if people were already familiar with social comparison before completing the survey? To gain more insight into priming effects and its connection with social comparison, we designed a second study with a new independent variable, forewarning, that shows how priming can affect an individual’s results in a social comparison task. Participants chosen for this study were randomly selected, and they were asked to complete a survey. Our new independent variable had two levels. If chosen for the warning condition, participants were informed on the effects of social comparison and would then receive the survey for either the upward or downward social comparison condition. Participants in the no warning condition would be directly given one of the two social comparison profiles. In other words, the no warning condition is simply a replication of study one without the average social comparison.

We hypothesized that participants that were warned about social comparison before completing the survey would overall score higher in positive self-esteem when compared to participants who did not receive a warning; and participants that received no warning would overall score higher in negative self-perception. Additionally, for participants that were not warned about social comparison beforehand, we hypothesized that those in the upward social comparison condition would score higher in negative self-perception than participants in any other condition; and those who received no warning but were in the downward social comparison would score the highest in positive self-esteem out of all other participants. We also predicted that participants who were warned about social comparison would show similar scores in negative self-perception between the downward and upward social comparison conditions, and similar scores in positive self-esteem between the downward and upward social comparison conditions.

Study Two Methods

Participants

This second study consisted of 547 participants, out of which 49.4% (n=270) claimed to be Florida International University students, and 50.6% (n=277) consisted of a mixture of family members and friends of FIU students. Of these 547 participants, 58.9% (n=322) were female and 41% (n=224) were male, with 0.2% identifying as Other (n=1). They ranged in ages from 11 to 69 with an average of 23.94 years (SD=7.59). The sample consisted of 72.4% Hispanics (n=396), 13.9% Caucasians (n=76), 8.6% African Americans (n=47), 2.7% Asian American (n=15), 0.2% Native Indian (n=1), and 2.2% identified as Other (n=12), out of which 0.8% consisted mostly of mixed races (n=4). From our participants, 67.5% reported that English was their first language (n=369), and 32.5% did not report English as their first language (n=178). Of this latter group, 29.8% reported Spanish to be their first language (n=163). See Appendix E.

Materials

As our second study consisted of two independent variables, forewarning versus no forewarning and upward versus downward social comparison; participants were randomly assigned to one of four possible conditions for this study: forewarning and upward, forewarning with downward, no forewarning with upward and no forewarning with downward. In part one of the survey, participants in the forewarning condition were presented with a paragraph that explained the effects of social comparison on people, while those in the no forewarning condition were directly shown the fake Facebook profile of Pat Masters for either upward (where Pat had done really well on his job interview and exam) or downward social comparison (where he had done badly on both the interview and exam). We used the same manipulations for social comparison conditions from study one, except we excluded the average condition for this second study.

Parts two and three of the survey were the same from study one. In part two the participants had to rate their agreement to the seven statements on impressions of Pat Masters from 1 (strongly disagree) to 6 (strongly agree). In part three, participants had to rate their level of agreement about themselves in the same thirteen statements. We decided to call our first dependent variable ‘negative self-perception’, and we created it using the mean score from participants’ negative views of themselves by selecting the statements used for the dependent variables in part one of the study: “I feel I am a lazy person”, “I feel inferior to others at this moment”, “I feel like I have trouble understanding things”, and “I feel like I’m not doing well”. A higher score on this variable would indicate that participants felt worse about themselves. Our second dependent variable was called ‘positive self-esteem’, and it was based on the mean score for participants’ positive views of themselves, again by selecting the same statements from study one’s dependent variable: “I feel I am (or will be) a good job candidate”, “I feel good about myself”, “I feel as smart as others”, and “I feel confident about my abilities”. The higher the score on this new variable, the better participants felt about themselves.

Part four of the survey requested participants’ demographic information, including gender, race or ethnicity, first language, whether they were a student at Florida International University, and current relationship status. Participants were reminded of their right to leave any uncomfortable question blank. Part five of the study was used as a manipulation check, where participants were asked to recall how Pat had done in the interview (great, bad, or unknown) or whether they were reminded of the purpose of the study, that is, the effects of social comparison (yes, now, unknown). The purpose of this section is to tell if participants were attentively completing the survey.

Procedures

Participants for this study were chosen to be either students from Florida International University, or friends and family of FIU students. They were provided a link to a Qualtrics survey that randomly assigned them to one of the four conditions and initially showed them an informed consent page, where they were informed of the purpose of the study, the number of participants in it, how long the survey would take to complete, any risks or discomforts from the study, and the benefits of participating. As the survey was online, participants were not required to complete it at a specific time or setting. Once participants consented to participate, they were allowed to move on to the rest of the survey, which consisted of five parts. Participants were debriefed at the end of the study and informed about our two independent variables and our hypotheses for the effect of social comparison and priming in social media. They were also provided contact information for any questions regarding the study.

Study Two Results

For our first manipulation check, we used social comparison condition (upward or downward) as our independent variable, while our dependent variable was whether or not the participants recognized which condition they were in. The chi square test we ran showed significant results, X2(2) = 455.9, p < .001. Most participants in the upward social comparison condition reported that Pat did great in the interview (95.6%), and most participants in the downward condition recalled he did bad (94.5%). These results show that participants were attentive while completing their surveys.

For our second manipulation check, forewarning (warning or no warning) was our independent variable, and whether participants recalled the purpose of the study (social comparison) was our dependent variable. Our chi square test showed significant results, X2(2) = 197.1, p < .001. Most participants in the warning condition reported being reminded of the effects of social comparison (83.6%), while most participants in the no warning condition reported not being reminded (62.5%). This shows that participants were paying attention when reading the survey. See Appendix F.

Our main analysis consisted of a 2 X 2 factorial ANOVA Test for social comparison condition (upward or downward) and forewarning (warning or no warning) as our independent variables and participants’ negative self-perception as our dependent variable. Results from this test showed a significant effect for social comparison F(1, 543) = 67.6, p < .001. Participants scored higher in negative self-perception in the upward social comparison condition (M = 3.66, SD = 0.89) than in the downward condition (M = 3.06, SD = 0.82). There was no significant main effect for forewarning F(1, 543) = 0.17, p = .68. Participants that received a warning beforehand (M = 3.33, SD = 0.89) showed similar scores on negative self-perception as participants that received no warning (M = 3.4, SD = 0.92). There was also no significant interaction for social comparison X forewarning, F(1, 543) = 0.98, p = .32. Participants did not differ when it came to scores on negative self-perception between the warning and upward comparison condition (M = 3.61, SD = 0.89), the warning and downward comparison condition (M = 3.08, SD = 0.82), the no warning and upward comparison condition (M = 3.71, SD = 0.89), and the no warning and downward comparison condition (M = 3.04, SD = 0.82). See Appendix G.

We ran a second 2 X 2 factorial ANOVA Test that used social comparison condition (upward or downward) and forewarning (warning or no warning) as our independent variables and participants’ positive self-esteem as our dependent variable. Results showed a significant effect for social comparison F(1, 543) = 96.7, p < .001. Participants scored higher in positive self-esteem in the downward social comparison condition (M = 4.76, SD = 0.64) than in the upward condition (M = 4.22, SD = 0.63). There was no significant main effect for forewarning F(1, 543) = 1.43, p = .23. Participants that received a warning beforehand (M = 4.54, SD = 0.68) scored equally in positive self-esteem as participants that received no warning (M = 4.44, SD = 0.69). There was again no significant interaction for social comparison X forewarning, F(1, 543) = 0.84, p = .36. Participants did not differ when it came to scores in positive self-esteem between the warning and upward comparison condition (M = 4.28, SD = 0.62), the warning and downward comparison condition (M = 4.76, SD = 0.66), the no warning and upward comparison condition (M = 4.16, SD = 0.63), and the no warning and downward comparison condition (M = 4.75, SD = 0.62). See Appendix H.

Study Two Discussion

We predicted that participants warned about social comparison beforehand would score higher in positive self-esteem than those who did not receive a warning; and those who received no warning would overall score higher in negative self-perception. Additionally, for participants in the no warning condition, we hypothesized that those in the upward social comparison condition would score the highest in negative self-perception and those in the downward condition would score the highest in positive self-esteem. We also predicted that warning participants would show no difference in negative self-perception scores between the downward and upward social comparison conditions, and no difference in positive self-esteem scores between the downward and upward conditions.