Overview
Summary: There are four opportunities to conduct short media analyses over the course of the semester that build on our labs to give you additional experience using research methods to answer a research question about media. For this first paper, you will focus on Content Analysis.
Details: Please write this paper using Google Docs and share the document with me as an Editor.
As we know from our work on the #fitspiration lab assignments, content analysis requires coding a large number of artifacts. To complete a content analysis, you will therefore need to create a research question that is answerable through a content analysis (whether quantitative or qualitative). I'm giving you a few options here, including one where you propose to collect your own data from a set of texts.
Option 1: #
Complete a content analysis modeled on the #fitspiration but using a completely different hashtag. You'll have to come up with the coding sheet yourself and I recommend reviewing Rose for information on how to create codes that are exhaustive, exclusive, and enlightening (p 155) and the #fitspiration article for some inspiration on how to apply content analysis to social media (doesn't have to be Instagram, it could also be X/Twitter, or TikTok, though both require some modifications). You will want to code 30 posts (images in Insta, videos in TikTok, etc.). Turn in a screenshot of the images, your code spreadsheet, and a summary of the results (whatever patterns you notice) and conclusion (what you think the pattern means). This is harder than the original option, since you have to figure out the code sheet yourself. If you want to use the AI App we used in the lab to do a second automated coding (only works for Instagram pics formatted as a row of three posts), that's fine. You can also use the same AI tools to do a more complex statistical analysis. If you use these, please make their use explicit in the writing of your essay.
Option 2: YouTube Scraping
If you feel somewhat computer savvy, then I recommend trying to scrape the comments on a Youtube video for analysis. To do this, it helps to have some familiarity with Python and feel somewhat comfortable using a terminal environment. I am not personally very familiar with either, but was able to use AI to create a scraping tool that I can share with you: https://colab.research.google.com/drive/1KfprBtaXzRmSNR-VjxzEXwSJSPDU_wG2?usp=sharingLinks to an external site.. What you'll do is go to the website, click the play button on the top left side of the code window to run the application. It will take a few seconds to run, at which point you'll be prompted to put in the URL for a youtube video you want to scrape. Try to choose a video that doesn't have more than 200 comments (newer videos often have fewer comments). You will then be given a spreadsheet that includes the usernames, comment text, number of likes, number of replies, and when the post was written. You need to then code these responses based on your research question. Since you'll have a lot of comments to sort through, you should try to have a code sheet that only examines a few things. One could be positive vs. negative comments, and seeing which kind of comments have more likes. You may also want to look for certain types of keywords based on the type of video. You must do the initial coding manually. If you'd then like to use Chatgpt or another AI tool to do a second automatic coding, or to do some more advanced statistical analysis, please talk to me to let me know, but it should be fine. You will need to explain your use of these tools in the essay.
Option 3: Real Vs. AI Youtube Comments
This option is very similar to option 2, but your research question is actually about biases in AI training. At an early point in engaging with AI around content analysis, I used Gemini's AI Studio to create a Youtube Comment Scraper AI that simulates comments. It wouldn't allow me to create an actual scraper app, so it instead created a simulation based on the type of video. In other words, the AI will create simulated (i.e. fake) comments based on the kind of video you would otherwise want to scrape. So, for this option, complete option 2 above but after you've generated the csv file from the real video, open this YouTube Comment Scraper AI and input the same url and the number of comments from the original video, and see what simulated comments it spits out. Then compare the comments from the original (actual video) with the simulated (fake) comments to see how AI changes their nature. Some things to consider might be whether the AI generated comments are less negative or when they are negative, have fewer replies. If the original file had offensive comments, did the AI generate similarly offensive comments?CriteriaRatingsMethod/Artifact Alignment
view longer description
Complete
The method and artifact are well aligned and there is an appropriate number for analysis.Incomplete
Failed to select an appropriate method and artifact combo. Artifact is not suited for analysis by method or method cannot optimally be used to read artifact. Or there is an inappropriate number of examples of the artifact to analyze using this method.Method Explanation and Comprehension
view longer description
Complete
The method is concisely and well explained, allowing for a full analysis without relying too heavily on description or quotes from the book.Incomplete
The method is not explained for a basic application. It lacks enough depth or detail, is superficial or general, and/or lacks grounding in the text.Method Application
view longer description
Complete
The method is well applied to the artifact(s). Appropriate vocabulary and/or concepts are used to create an analysis.Incomplete
The method is not applied to the artifact(s). Appropriate vocabulary and/or concepts are absent and hinder the creation of an analysis.Organization and Mechanics
view longer description
Complete
The organization is good. Intro and conclusion are clear and too the point, with little extraneous information. The thesis/argument outlines the analysis that results from the application of the method to the artifact. Most of the ideas in the sentences and paragraphs flow logically. Writing is largely free of typos, spelling and grammatical errors, and poor constructions (no more than 4).Incomplete
The organization is confusing. Intro and conclusion do not serve their purpose, are irrelevant, or no present. Essays lacks any thesis/argument. Ideas in sentences and paragraphs lack logic. Writing contains 5 or more typos, spelling and grammatical errors, and poor constructions.