Data Analysis
Data Analysis
After viewing the Instagram profiles of 36 men and 36 women and collecting data on the 5102 verified accounts that they followed, we concluded that there were clear differences in gender based content consumption that mostly supported our initial hypothesis. We analyzed this data on three levels: overall trends, single-gender trends, and individual case studies.
First, when the data is viewed collectively, we can see that the overall verified followings of these 72 students were male dominated , with 39.36% of the accounts being men, 35.57% being non-binary or non-gendered, and only 25.07% being women. In terms of raw data, this amounted to a total difference of 729 accounts , or more than 10 accounts per person . The most drastic difference was in sports, where almost 86% of followed verified accounts were male and under 7% of followed verified accounts were female. This difference supports the predictions based on our initial research, which showed that athleticism tended to be rewarded more in men and that female athletes were often considered to be “lesser” than their male counterparts, as studied in Ellie’s individual commentary . However, another large gap existed in the music realm, with close to 65% of followed verified accounts being male and just over 30% being female, which was more surprising considering that three of the five most followed women on Instagram are music artists .
The only following category where verified women outnumbered verified men was in the Model, Influencer, or Personality category, with around a 9% percentage gap, which definitely plays into the stereotypes that were predicted by Ayşe Lara’s commentary in terms of what content was most rewarded for women versus men .
However, in line with Angeliki’s commentary, when this data is examined for each gender separately, we can also see clear differences in what content is rewarded by men versus by women .
The verified followings of the 36 men were less than 15% women and almost 50% men, with the rest being non-binary or non-gendered accounts. On the other hand, the verified followings of the 36 women were much more equal, with 36.14% women, 35.15% non-binary or non-gendered, and 28.71% men. It was interesting to note that overall, 25% of the men did not follow any verified women on Instagram. On the other hand, under 6% of women did not follow any verified men on Instagram. In general, the verified followings of women tended to be much more gender-balanced than the verified followings of men. Excluding the Educational Institution sector (which was entirely non-gendered), men followed more verified men than women in all seven of the remaining categories, whereas women followed more verified women in four of these categories and more verified men in three of those categories.
As we are all studying at an educational institution focused on politics, it is notable to look at the followings of Sciences Pistes in the field of Politics/Current Events. Students of both genders tended to follow politicians and news accounts from their own countries as well as politicians and news accounts from other countries, reflecting the global perspectives on campus, and both genders followed the most non-binary or non-gendered accounts in this category. However, men tended to follow more than twice as many male political accounts compared to female political accounts — on average, five men and two women — while a woman’s average following was just under three male political accounts and between three to four female political accounts. This trend supported our initial ideas that men would be more likely to fit in with the bias of the political field toward men, while women would be more likely to respect and elevate female voices in politics, as Lilith’s commentary elaborated upon. Even at a school where the majority of students tend to identify as feminist and in support of feminist politics, this is not always reflected (whether knowingly or unknowingly) in the media that students consume.
Interestingly, the large gap in sports followings existed for both genders — both women and men tended to follow more male athletes than female athletes. However, the number of sports accounts followed by each gender differed: the average man followed around 11 male athletes and fewer than 1 female athlete, while the average woman followed about 3 male athletes and fewer than 1 female athlete. Therefore, there was a much more dramatic difference in the followings of men. One consideration for why this gap persisted for both genders, however, could be that male sports leagues tend to receive more funding, publicity, and prestige compared to their women’s counterparts. For example, F1 accounts and drivers (all of whom are male) were heavily featured in the athletic followings of both genders, while female racing drivers in W Series or other similar racing leagues were virtually nonexistent; however, F1 receives much more media coverage and is regarded as significantly more elite compared to other leagues that have female competitors.
Looking at the Model/Influencer/Personality category, the data was almost opposite between the genders: the average man followed around 6 men and 2 women in this category, while the average woman followed around 9 women and 4 men in this category. We were initially unsure of what to make of this trend, since at first it did not seem to align with our netnography unless we considered internalized sexism. However, this category was actually best analyzed by looking at individual accounts and seeing what specific Model/Influencer/Personality accounts were most followed. The women followed by women in this category tended to be famous women (for example, the Kardashians or Jenners) or social media personalities. Although men tended to engage in much less of this content, the female accounts followed in this category were significantly more sexually explicit, to the point where they could not really be compared to the average Instagram comedian. For this reason, we do not feel that placing all of these types of content in the same category best represents the data, as these nuances were not entirely captured on a larger scale, and if we were to do future studies on the matter, we would make a new category to capture this distinction.
While looking at this data on the individual level, we wanted to highlight some individual case studies that we saw as “extremes” of each trend in each gender.
One male student followed 19 verified men including many athletes, some music artists, a couple of lifestyle accounts, and an actor. However, he only followed two verified women: an actress and an account that shared sexually promiscuous content. Another male student followed 53 verified men, yet only eight verified women: two musical artists, two actresses, and four accounts sharing sexually promiscuous content. There was one man who followed 234 verified accounts, but only four of them were women, and as mentioned above, there were also nine men out of the 36 who did not follow a single verified woman. However, on the other hand, there were some men who did the opposite: one man followed 25 verified men and 38 verified women. Another followed 123 verified men and 123 verified women — perfect statistical equality.
Looking at some case studies of individual women, there were two women who did not follow any men, and more who followed a very low percentage of men, such as one woman who followed only 3 verified men out of the 45 verified accounts that she followed. There were some women who were on the opposite end of the spectrum, with one woman following 44 verified men, most of whom were athletes, and only two verified women, a lifestyle account and an influencer. Another woman followed 21 verified men, more than half of whom were musical artists, and again only 2 women. In general, when women followed significantly more men than women, those men tended to be athletes or musical artists. However, women as a whole did tend to have much more balanced followings. One woman’s verified following count was exactly 20% men, 20% women, and 60% non-gendered, to point out another impressive ratio of equality.
Based on the individual case studies, gender-based analyses, and overall data, we found that the vast majority of our data aligned with our hypothesis that social media reinforces gender stereotypes for both genders, and that users (especially male users) will engage with content that aligns with these stereotypes. We saw that the stereotype of male athleticism was present in the followings of both genders, although the sexualization of women was seen much more in the followings of men than the followings of women. We saw that men were more likely than women to align with certain pre-identified biases, such as the bias toward men in politics. Finally, we saw that in terms of overall content, men tended to have less equal verified followings compared to women.
As in any study, there were some limitations to our data. One has already been mentioned: the Model/Influencer/Personality category was too broad. We also recognize that we are only looking at the information that is publicly available to others on Instagram, and that it does not necessarily include content that people consume on their Explore feeds but do not follow; it also does not include accounts that were not verified on Instagram. However, we chose to rely on the verified accounts that people followed for a few reasons. Firstly, it provided a clear basis for us to compare followings, as it is harder to categorize non-verified accounts of people we do not know (are they following this person because they post fitness content or because they know them from high school?) and erased this ambiguity in what accounts we would consider for data. We also felt that if we asked people to share their Explore feeds with us, they would be able to curate what part of their feed they shared with the knowledge that we would be viewing it attached to their name. Although people’s followings can also be viewed publicly and are attached to their names, our hypothesis relied on the fact that people do not always think critically about the accounts that they are following and how they express subconscious biases toward certain types of content, whereas people would be consciously thinking about the content that they wanted to share if they were asked to self-report. This is why we chose to conduct a small range of interviews once we had already collected this data, so that we could compare people’s conscious responses (without being able to look at their feeds) to what we saw on their profiles.
Finally, we did not conduct research on non-binary or gender non-conforming students at Sciences Po because the comparatively smaller group of non-binary and gender non-conforming students at our campus would result in the data not staying anonymous in effect, and because there would not be sufficient data to form statistically sound conclusions. However, we do want to acknowledge that gender and gender identity is not limited to the binary of this project.
Links