At the beginning of this class, I was primarily apprehensive about writing literature reviews, finding relevant data that I could analyze to test my hypothesis, and which statistical method I would use in my study.  Throughout the course of this class, I have found that the first two fears turned out to be non-issues, but the third fear was somewhat well-founded.  There turned out to be a copious amount of literature on the topic of video games and academic achievement, with researchers finding evidence to support both sides of the argument.  The second fear actually turned out to be one of my favorite parts of the class.  I had never conducted my own survey before, and I developed a greater appreciation for the approval process researchers must undergo, as well as how they craft their questions and distribute their surveys to their test subjects.

Since I took COL Foster’s IS-460: Research Design for Political Science class last semester, I already had an idea of the major concepts covered in this course, such as the main parts of a political science research paper.  In that class, the professor briefly mentioned some of the most common statistical methods of analysis.  However, one of my favorite parts of this class was being able to actually practice analyzing data using these statistical methods.  I especially enjoyed learning about the different types of variables (interval, continuous, etc.) and how researchers used different statistical methods based on what type of variables they were dealing with.

Although the SPSS homework assignments did feel a bit monotonous at times, I enjoyed being able to work with statistics again for the first time since freshman year.  After eight weeks of SPSS practice, I now feel very comfortable with program itself.  Navigating the program and knowing how to do technical parts of obtaining the correct charts and graphs was never difficult for me, thanks in large part to the helpful visuals and explanations in the SPSS textbook.  Looking back on my time spent working on the SPSS homework assignments, I now realize that at times I became so focused on simply crunching the numbers that I did not always take a minute to step back and take note of the large concepts going on.  For example, I know from memory how to run a Cramer’s V test, but I struggle to remember when to run a Cramer’s V test, why I would want to run the test in a certain situation, and what exactly the test results tell me about my findings as a whole.  I actually mentioned this fear in my blog post at the end of week seven of the class, so it turns out that my fears were somewhat well-founded—and I am relying significantly on the SPSS textbook for guidance on when to run certain statistical tests.

            Nevertheless, I found the SPSS homework assignments to be helpful and informative.  I found the content of the later chapters to be the most interesting, because they reviewed the concepts that I had learned in freshman-level statistics and took then took them a step further.  For example, I already had a general idea of how to find statistical significance using p-values and confidence intervals, but by learning Somer’s d, I learned how to find the actual strength of this relationship.  So, I found these chapters to be the most rewarding ones of the SPSS segment of the course.

I remember feeling significant frustration at my fairly poor performance on the in-class quizzes.  Going into the quizzes, I felt confident that I knew all the terms, especially since I had taken COL Foster’s IS-460: Research Design for Political Science class last semester.  When I saw my first quiz results, I was surprised to find that I did not actually know the meanings of terms and concepts such as deductive and inductive logic, interval and ordinal data, and positive and negative skewness.  I was also frustrated that the second quiz did not go well, either.  I confused the definitions of additive and interactive relationships.  However, getting the answers to these questions wrong helped me to remember them for the future.  I now know these meaning of those terms today because I defined them incorrectly on the quizzes.

With our group’s research paper almost complete, I can safely say that, despite some challenges, this segment of the course has been my favorite.  We ran into some of the typical challenges of group projects, such as trying to make sure that all four people in the group are on the same page, especially with regards to how to interpret the data analysis.  For the data analysis, we were no longer working with nice, clean data from datasets in the textbooks that came with exact, step-by-step instructions on how to analyze them.  The task was intimidating at first, and I was nervous that the other groups were far ahead of us in this regard (some of them were).  Nevertheless, running tests on our own data that we collected and analyzing the results turned out to be the most rewarding experience of the entire course.  Even though none of our tests turned out to be statistically significant, I now feel like I have a solid understanding of every component part of a political science research paper.

Now that this course is under my belt, I can turn to tackling my Honors Thesis with the skills I have learned here.  That may prove to be my most rewarding intellectual experience yet.

I found the SPSS homework assignments to be useful in that they introduced me to the basics of the computer techniques of statistical analysis. For the most part, I found them to be fairly easy, although sometimes they felt a bit repetitive. They were grade-boosters as well, but I am glad to be done with them and moving on writing an actual research paper.
Of course, my group already submitted drafts of our literature review and our section on our theory, hypothesis, and conceptual definitions earlier in the semester. I feel good about our work on them, although we need to add several more sources to our literature review and do a little more work on the causal mechanisms in the theory.
Right now, our immediate concern is collecting our data. I took COL Foster’s IS-460: Research Design for Political Science class last semester, so I am familiar with the main parts of a political science paper. However, in COL Foster’s class, all of us relied on existing datasets already available online. This is the first time that I will be collecting data myself.
In my past several years as a cadet at VMI, I have received numerous requests from other students, most of them via email, asking me to participate in a survey for their research paper. I had noticed that on all of those surveys, the first page was always an “Informed Consent” form.
As I am now filling out the IRB application for this class, I realize now the importance of these forms. I understand the seriousness of violating the privacy of subjects who participate in our study. I have learned techniques on how to maintain the privacy of the subjects, such as by not attaching their names to the surveys themselves at any point, and by destroying the surveys after we have gleaned the data from them and have the data stored securely on a spreadsheet in a database.
In the IRB application, our group had to decide what questions we wanted to ask our subjects, as well as how we were going to select participants in the study. The two most obvious questions that we absolutely needed to ask corresponded to our independent and dependent variables: “How much time do you spend playing video games?” and “What is your GPA?”. However, there were problems with simply asking these two questions directly. With regard to the question “How much time do you spend playing video games?”, people may not really know or may intentionally misreport how much time they invest into playing video games. Popular social thought often associates playing video games with laziness, unproductivity, and a lack of ambition. People may not want to admit to a researcher or themselves how many hours they truly spend gaming, or they simply may not keep track of the hours in the head very accurately. To combat these problems, our survey broke the main question down into parts by first asking the subjects how many days per week they played video games, and then asking how many hours per day they played. We hoped that this approach would guide subjects to provide more accurate estimates. With the regard to the question, “What is your GPA?”, similar problems seemed present. People with lower GPAs may be less willing to report to a researcher their actual GPA, or they may unintentionally misreport the information. To combat this problem, we provided a several ranges of GPAs. The subjects simply had to select the range into which their GPA fell, instead of writing down their actual GPA out to several decimal places.
After that, we submitted our IRB application, with survey attached, for review. We had to iron out a couple things: First, we had originally framed the question about video game usage as “Do you own and have access to a video gaming platform?” However, COL Sanborn pointed out that this did not really reflect the point of our research paper, which was to discover if actually playing more video games led to a decrease in GPA. So, we adjusted our original question to the questions noted in the paragraph above. Second, our survey had no control variables. So, we added several control variables such as whether the subject held rank or was an NCAA athlete. Third, we had to decide what to do if the first person on the door card was not in the room. (We decided to simply proceed to the next room.)
We agreed 60 subjects was a sufficient number for our survey. We surveyed 15 subjects each. Our method was to go to every fourth room in barracks and survey the person whose name came first on the door card. There was some trouble with this method since people where not always there or chose not to respond to the survey. (Most cadets are very tired of taking surveys at this point in the semester.) So, I had to adjust my survey method to going to every other room in barracks, and soon I had collected all the data I needed.
Once we had our data, the coding process was straightforward. I had never coded variables myself before. It was actually an interesting experience, as I changed people’s answers to basic questions into “1’s” and “0’s” on a spreadsheet in SPSS. COL Sanborn showed me how to change the names of each column of data in the SPSS spreadsheet to reflect the questions that each column corresponded to.
At this point, all we need to do is decide what type of statistical analysis to use and then run it on SPSS—and then write the 5,000-word paper.