This project was undertaken to further understand and investigate the possible correlation between a person’s gender and what they like to read. There is a common stereotype in America that women like to read more than men. Everyone knows that old lady that makes frequent trips to the library to get romance novels, but does that old lady have a male counterpart? So let me ask you, is there a correlation between the type of reading someone participates in and a their gender?
The objective of my original study was to see if there is a correlation between a person’s gender and if they like to read. To investigate this, I received a school roster that has every current teacher and student on it. I segregated the list into gender and I had Minitab choose 60 names from each list, giving me 120 names total. I picked 120 names because it’s a very close number to 10% of the total sample size and it’s a nice, easy number.
While Minitab was picking names out of a digital hat, I started constructing my survey. I made 2 lines in the header, one for the person’s gender and the other for their grade level. Teachers were instructed to put a “T” for teacher. I then asked a series of six multiple choice questions. The first asking if the person calls themself a reader. I figured people would react differently to this question because the word “reader” is so subjective and broad, so I asked 5 more reading related questions. The questions were yes or no questions based on what a person might read, for example a question was “Do you read blogs such as 4chan, Reddit, and Tumblr?”. If a person answered yes to this, they would be a reader but they may of answered no to the original question.
After I finished with the surveys and Minitab finished doing it’s thing, I headed to the Guidance office to do a very tedious search through people’s schedules. I was looking for studyhalls or a class they had Ms. Caso as an instructor for. Almost half of the students didn’t have a studyhall, so I had to find a flexible teacher in their schedule. This process took over 3 non consecutive man hours. I then gave the surveys to Ms. Caso so she could put them in the teacher’s mailboxes.
About a week later, I received 86 out of the original 120 back. As I was sorting through them gathering data I noticed a very frightening error, 46 people left the “Grade level” and “Gender” tabs blank. This was a code red situation. Those two pieces of data were crucial to my study. Without them I couldn’t test my hypothesis. 40 complete surveys weren’t nearly enough. At that point I knew I either had to make new surveys or do something different with my existing data. I decided to change my hypothesis to something very similar but I could test this new hypothesis with much more certainty than my original. My hypothesis went from “Is a correlation between a person’s gender and if they like to read?” to “Is there a correlation between the type of reading someone participates in and a their gender?”. My new hypothesis offered a lot more variability. I am now comparing 2 lists of 5 variables compared to originally only comparing 2 single variables.
In order to test my hypothesis I needed to do a Chi-Square Test for Association. I had to do this test because I’m comparing 2 sets of variables. I ran into another snag, I didn’t have enough data for one of my expected values. For the male blog expected value, I got 4.696. This value has to be at least 5. I decided to round up because the value was close enough to 5 and I didn’t have any other options. I ended up getting a P-Value of 0.457, which is greater than my alpha at .05. Therefore, I fail to reject the null hypothesis at the .05 level of significance, concluding that there isn’t sufficient evidence to say there’s a relation between gender and the type of reading one does.