The Relationship Between Major and Social Media Usage

The Relationship Between Major and Social Media Usage

        Our study asks the question “does your major influence which social media sites you visit or how long you spend on them?” We thought that maybe there might be some correlation between the types of sites someone uses and their major, for example perhaps more Arts majors would use artistic inspiration sites, like Pinterest or Tumblr. We also wondered if we could learn anything about total hours spent on the Internet.

Methods

        To gather our data, we conducted an online survey, using survey monkey. We asked for some basic background information, class standing, gender and, most importantly, major. Then we asked how many an hours a week students spent on Facebook, LinkedIn, Pintrest, Tumblr and Dating sites. We also asked for any other recreational sites used and how many hours a week for those sites. We received 56 responses.

        After we gathered our data, we needed to edit it a little so we could use it. First we threw out a few data points that didn’t fit within our research. We got a couple of responses that listed majors not offered at Lewis & Clark College, because we wanted to focus the survey on the Lewis & Clark community.  We also threw out any data points that were undecided.

        Since we didn’t get enough responses per major to make a study of each individual major plausible, we grouped them in to four different categories.  Our “Arts” major is mostly composed of Theatre majors but there were a couple Art and Music majors. The “Humanities” major had a few Rhetoric and Media Studies majors and several people involved in the foreign languages and a single English major. “Social Sciences” include quite a few Psychology majors and a few Sociology and International Affairs Majors among several others. “Math/Science” majors were mostly Biology and Computer Science majors.

        We also had a few double majors. We decided to count them twice, if the two different majors were in different categories, even though it would compromise the integrity of our simple random sample a bit. We didn’t feel like there was a better way to handle them. Since we didn’t ask for a preference between majors in the case of doubles. Our other option would be to throw them away, but we decided the smaller sample size would negatively affect our study more than the double counting of double majors.

        After gathering our data, we realized we made a few key errors that compromised our data. We decided to do an electronic survey and we distributed it the best way we knew how, via Facebook. It was only after when we were noticing that all of the people we surveyed used Facebook that we realized we only distributed the survey via Facebook and therefore, people who didn’t have Facebook profiles couldn’t have gotten access to the survey. This could have compromised our data a lot more than it did, because a large portion of the Lewis & Clark community is on Facebook. It still compromised our Simple Random Sample and we should not expand our results to the Lewis & Clark community as a whole, rather the Lewis & Clark community with a Facebook. We also don’t have a way to tell if everyone who took the survey is, in fact a Lewis & Clark student. Since we posted the survey on Facebook, any number of people could have taken it regardless of whether or not they were a student. We were able to edit out a couple of these people because they listed majors that Lewis & Clark doesn’t offer, but if someone from a different school took the survey and they had a major that Lewis & Clark offered, we’d have no idea that they weren’t part of the community we were trying to survey. A survey about Internet usage, like this would be much more accurate done in print and distributed around campus in hard copy. It would also be a lot easier to make sure everyone who took the survey was currently enrolled at Lewis & Clark.        

        After all of our edits we ended up with 66 usable data points.

Data Analysis

First we looked a the data in terms of how many people per major used each site, as shown in the figure below.  We noticed that the bars within each major

[pic 1]  seemed similar proportionally to each other.

        Then we did an Analysis of Variance on the hours spent by each major on three of our most popular sites, Facebook, Tumblr and Twitter. In ANOVA, the distribution of data should be normally distributed. ANOVA also assumes the assumption of homogeneity, which means that the variance between the groups should be equal. ANOVA also assumes that the cases are independent to each other or there should not be any pattern between the cases. In our survey, we didn't know any of these assumptions to be true, but we decided to go ahead anyways,

Facebook

Our null hypothesis that major doesn’t affect how long a student spends on Facebook and we are testing this null hypothesis with an alpha of 0.1

We got a very high P-Value of 0.9, which means we have no evidence to support the rejection of our null hypothesis.

Tumblr

Our null hypothesis that major doesn’t affect how long a student spends on Tumblr and we are testing this null hypothesis with an alpha of 0.1

Our P-value was very high again so we cannot reject the null hypothesis.