Abstract
Sexism is discrimination, stereotypically against women, on the basis of their gender and is socially encoded into our language. Many studies have found that language is a predictor of how sexist a culture can be. Also, many studies have found that sexist language can prime women to do worse in certain areas.
It is important to understand the effects that language like this can have in order to fight it and stop it from doing harm. Many studies that have been done only look at men or women but very few look at both. For this study we wanted to fill in the gaps by looking at both men and women together and we do this by using social media to deliver a survey to prime our participants and measure both their amount of sexism and their mathematics scores.
In order to prime our participants, we used an article that originally called women girls and a second measure where we altered then original articles word girls to women and the word girl to woman. We believed that women’s scores would lessen when given the article that referred to them as ‘girls’ and that men’s scores would stay constant. However, we found during the course of our study that the priming we provided created no difference in mathematics scores.
Sexist Language and its Effects on Men and Women
Women around the world are faced with language that is meant to make them feel lower and degrade them. When we address a group of women, we refer to them as ‘girls’, and this takes women from a place of being adult to making them feel like little girls.
However, when we refer to a group of men, we refer to them as ‘men’ or ‘guys’. This use of language is important to recognize because although women have rights they are still not treated equally. That’s what makes this topic so important, because it will help people recognize when they are using sexist language and help them see how it affects the people around them.
The use of sexist language affects women all around the world. Women are missing from many different careers around the world because those careers are considered male jobs. If language is adjusted and presented differently in those careers, then it could attract more females and fill those gaps where women are missing.
In language gender is used to identify male from female but in most languages it does more than that. In a study conducted by Prewitt-Freilino, Caswell, and Laakso (2012) they wanted to look at how different languages use gender and if they have any gender equality.
They believed that countries who have a heavily gendered language would have more sexism than countries that do not. During their study they categorized 134 countries and sorted them into seven categories; either gendered, natural gender, genderless, gendered and natural gendered, gendered and genderless, natural gender and genderless, and other.
They found that that their hypothesis held true. The researchers believe that current findings show a difference in gender equality however their research cannot prove what role language plays in this gender equality. The overall limit of this study is that it cannot show causation.
Gendered language is dangerous because it not only effects one group but both. The previous articles show us that it is so ingrained that using sexist language doesn’t affect men at all and it makes women feel less than. In a study done by Sczesny, Moser, and Wood (2015) they wanted to look deeper into how we use gendered language and even how we use it to manipulate certain situations.
Their motivation for this was that they found there was a gap in previous research between the difference of habitual and deliberate use of gendered language. To look at this they used two different studies. During study one the researchers used a questionnaire in order to measure; gender inclusive language, deliberate predictors of language use, and habitual predictors of language use.
Then during study two they repeated of the first study, however the questionnaire contained several subtle points of sexism throughout. Though the authors of the article didn’t point out any limitations I found that the questionnaire could limit a participant’s answers and ultimately add bias. The author found that there was a way to promote more gender inclusive language, and that they found that their research shines alight on what causes sexist ideology and language.
In our society today, gender is a more objective concept and, in a study, conducted by White and Gardener (2009) they look at whether the prominence of a person’s gender identity affects they way they process gender stereotypes. So, if a woman encounters a gender stereotype will she start to identify differently to protect herself.
The authors felt this research was important because they found that there was a gap in the research on gender identity and stereotypes. For their study they had participants look at words and see what they identified with using a Stroop task. They also conducted a second study where they did the same thing minus the Stroop task. The study had a few limitations including the sample size which held more women then men and they also found it hard to interpret results.
Though everyone laughs at jokes about women there is a relationship between sexist language and attitudes towards women. In a study conducted by Douglas and Sutton (2014) they try to look at what factors can predict a gender gap that happen in language. During this study the participants were asked to complete a survey that contained an attitude toward women scale, SDO scale, and inventory of attitudes toward sexist/nonsexist language.
They found that women were in support of gender inclusive language and men where not. Also, they found that use of sexist language may not be a product of sexism but rather show an acceptance of it. The authors of this study had a few limitations including a lack of diversity in their small participant pool. All of this study’s participants were college student which lacks a representation of the larger population.
Our attitudes towards women effect how we subconsciously present information to people. In an article by Steele and Ambady (2006) they explore how gender priming can affect performance. The authors wanted to explore this because they found that priming had only been done to participants not part of the stereotyped population, so in their study they wanted to include the stereotyped group. they really wanted to see if the positive or negative priming would affect women’s math and art related abilities.
The author had a few limitations in their study including a small participant size which started out small at 50 and became smaller when four more participants dropped out. They also felt that they had a confounding variable of interpersonal goals. Even with those limitations the authors felt that their study still showed that priming affects the attitudes of a stereotyped group.
Sexist language is important to identify, and it is also important to see what can cause us to have that sexist language. In a study conducted by Rudman and Borgida (1995) they looked to see if people primed with negative attitudes toward women are more likely to show discriminatory behavior. The authors wanted to look more closely at this because they felt that there was a gap in research on the effects of institutionalized sexism.
The authors found that there was a limitation in the questions in the interview portion because they felt the questions were too ambiguous. Also, while reading the article it becomes apparent that there may be some experimenter bias. Overall the authors of this study found that men don’t tend to take derogatory thoughts against women seriously, and instead think of it in more of a joking manner.
Previous research is missing studies on how sexist language affects both in and out groups. I believe that priming will make people feel a certain way toward women in any type of job environment. In my study I want to look at both men and women’s confidence in getting a job when I prime them with a sexist article.
My independent variables will consist of a sexist and non-sexist article. Then my dependent variable will measure confidence in themselves and whether or not they feel they can reach their dreams also it will evaluate whether or not they read the article. I believe that I will be hiding the purpose of my survey by pretending that it is a memory survey.
Methods
Participants
For our study we ended up with 150 participants total, however 60 participants did not finish the study and had to be dropped, this left us with 90 participants total. Of that there were 30 men and 60 women. Our participants ages ranged from 18 to 74 (M= 3.26, SD= 1.195). More than half of our participants had done some sort of college. The income of our participants was 47.8% for under 60,000 and 52.2% for over 60,000 a year.
There was not much diversity while 70% of our participants were White, 15.6% were Hispanic/Latino, 4.4% were Asian, 2.2% were Black, 2.2% were American Indian/Alaskan Native, 1.1% were Native Hawaiian/Pacific Islander, and 4.4% were Other. There were 48 women articles and 42 girl articles. Of the participants, the women article had 20 men and 28 women, and the girls article had 10 men and 32 women.
Materials
Level of sexism was measured with 6 items adapted from the Neo-sexism scale (Campbell et al., 1997), such as “I consider the present employment sysem to be unfair to women” (ɑ = .608) and 11 items adapted from the Ambivalent Sexism Inventory (ASI) (Glick & Fiske, 1996), such as “Most women interpret innocent remarks or acts as being sexist” (ɑ = .673). Responses on the Neo-sexism scale were rated on a 7-point scale (0 = Disagree Strongly, 6 = Agree Strongly).
The middle number that would be labeled as “neither agree or disagree” was removed to avoid acquiescence bias. Questions 1 and 6 were reverse coded. Responses on the ASI were rated on a 6-point scale (0 = Disagree Strongly, 5 = Agree Strongly). Questions 3, 6, 8, and 10 were reverse coded. Mathematics score was measured using 6 items adapted from the practice ASVAB test, such as “What is the name of a quadrilateral with four equal sides?” (ɑ = .232).
Procedure
The study consisted of an online survey that was built using Qualtrics. The survey was distributed over the researchers’ social media accounts (i.e. Facebook, Instagram, Snapchat). Once the participants clicked on the link, they were then asked to read and agree to the consent form.
In order to cover up the study’s true purpose, participants were told that the researchers were looking into adults’ education, including mathematics and reading scores. Afterwards, they saw the introduction page, which let them know what steps they were about to participate in and continued to encourage the cover story.
The participants were then asked to read one of two randomly assigned articles. One article was an original article we found on Psychology Today, which had already referred to women as “girls” (see Appendix A). The second article (see Appendix B) was identical, except for the changing of the words “girl” to “woman”, and the words “girls” to “women”. Both articles were cut down to prevent participant mortality.
After the article, the participants were asked to complete a math test, which was taken from the math section of a practice ASVAB and was cut down from 16 questions to 6, also in order to prevent participant mortality. An attention check was included in this section of the survey, in which participants were instructed to answer “5” if they were paying attention.
From the math test, the participants were asked to complete a reading check as a part of our cover story. To check the sexism levels of our participants, there were two sexism scales: The Neo-sexism Scale (Campbell, Schellenberg, and Senn,1997) and the Ambivalent Sexism Inventory (Glick and Fiske 1996). Once finished with all above parts, the participants were asked to complete a demographics section.
Questions included gender, age, education, income, and ethnicity. Following the demographics, there was a manipulation check, where the participants had to answer how they felt about calling women “girls”. The response was measured using a 7-point Likert scale (1 = Strongly Disagree, 7 = Strongly Agree). This question concluded our survey, and the participants were then able to read the debrief, where they were told what the study was truly about. Contact information was included if they had any questions about the survey.
Results
- Correlation- ASI & NEO
Results of the Pearson correlation indicated that there was a significant positive association between both of our sexism scales, (r(90) = .46, p < .01). This means that our scales were similar in the results we measured.
- Correlation- Sexism Score & Mathematics score
Results of the Pearson correlation indicated that there was not a significant association between the participants sexism score and their mathematics score, (r(90) = .60, p >.005). This proves the null hypothesis that there is no direct association between the sexism score and the mathematics score.
Gender, Article Received, and Math Score
Results of the MANOVA indicated that there were not significant differences in math scores for gender based on the article they were given (F(1,90) = 0.67, p = .797). This means that our hypothesis is null and that the manipulation we presented did not alter our participants scores.
Males who read the article using the word women (M=4.05, SD= 2.50) opposed to males who read the article using the word girls (M=4.10, SD= .54) and female who read the article using the word women (M=3.93, SD=.212) opposed to females who read the article using the word girls (M=3.84, SD= .198).
Gender, Article Received, and Sexism Score
Results of the MANOVA (see figure. 1) indicated that there were not significant differences in sexism scores for gender based on the article they were given (F(1,90) = 1.398, p = .240). This means that our participants sexism did not change whether they were presented with a sexist article or not.
Males who read the article using the word women (M=32.75, SD= 2.61) opposed to males who read the article using the word girls (M=35.40, SD= 3.69) and female who read the article using the word women (M=32.46, SD=2.21) opposed to females who read the article using the word girls (M=28.69, SD= 2.06).
Conclusion
Overall our findings were not significant. Even though our sexism scale was valid, our overall hypothesis was null. The reason that our findings were not significant could be due to the fact that or mathematics measure was so short and not comprehensive enough to capture whether or not the manipulation had any effect. Also, our results could have been because our manipulation was not effective, or people were not reading our article manipulation thoroughly.
Discussion
Sexism is very harmful whether it is done deliberately or habitually and there are many studies to prove this to be true. However, our study suggests that there is no difference in scores based off of sexist priming. We found that the scores between men and woman stayed constant no matter the article they received. This could be because of the fact that our priming was not done well enough and possibly because the source its self was sexist so our ‘woman article’ didn’t have the neutral effect that we desired.
Although our research suggests there is no harm in the way that we refer to women and that sexist language has no impact on women’s performance. There are countless studies that suggest otherwise. In in a study conducted by Wasserman and Weseley (2009) these researchers look at how countries with gendered grammatical rules effect women and they have found that those rules have has an impact on women.
Likewise, in a sexist language study conducted by Sarrasin, Gabriel, and Gygax (2012) they found that the implementation of gender-neutral language may have push back but overall it is important to implement that kind of change because it reflects the societies view on women and the amount of sexism. The study also found that although there may be push back at first that push back will fade over time.
Though we looked for research that would agree with the findings in our study, overall there was none that said that sexist language doesn’t affect women. It could be possible that where we looked didn’t have those results available to us. However, with the overwhelming amount of articles that disagreed with these results it is safe to assume that our research had limitations that caused us to reach these conclusions.
There were many limitations to our study, because we only had a few weeks to design, implement, and collect data. One big problem of our study was participants including size, demographics, and drop out. We had a large drop out rate due to the fact that participants were required to read an article. Our demographics were limited to who the four of us knew because we distributed the study via our private social media accounts.
In the future there are ways to improve upon the limitations in our study. Future research could create a stronger mathematics test, because our test was only 6 questions long and did not cover a wide area it was not a good measure of mathematics ability.
Another way to improve on our study is to get a lager diversity within the demographics, more than half our participants were white and this is not an accurate depiction of the population. Lastly, a way to improve on our study is implement it in a lab, this is because our study was so long that it had a large drop out rate. If you were to move the study to the lab it would lessen the dropout rate.
Though our study had many limitations that gave us a slew of insignificant results. It is important to note that this kind of research is very important. Studying sexism gives us a good look at what happens to a oppressed group of people. Women are extremely marginalized, and we need to understand how that affects them.
There is a difference between calling a grown woman a woman and calling her a girl and according to current research it can harm women’s performance. As a society we should continue to collect research, so we are able to understand the harmful effects of sexist language and take the steps to change it.
