Week 3: Getting into the Grind

Krish S -

Welcome back guys! After spending the past two weeks developing a methodology and making significant revisions following a miserable initial attempt at carrying it out, I’m finally able to say I’ve made some progress! The process is very tedious, but it is what it is. Finding the reviews is the only hard part about this whole data collection process. I actually enjoy reading about what the reviews say about the actors/actresses. It’s interesting comparing how different news outlets view the performance of a certain actor/actress. It’s also fascinating looking at how the style of writing for critics differs when talking about male actors versus female actors. However, these are just my observations from the naked eye. The sentiment analysis will have the final say of whether or not gender bias exists from the perspective of these film reviews. 

If I am being honest, while I was in the midst of collecting my 120 film reviews, I got really bored. And so, I began researching about sentiment analysis. Based on the research, I actually don’t think it will take too long. There are a lot of resources online that will help me build my sentiment analysis model. Additionally, I know a few people who are also conducting a sentiment analysis for their respective research project. I thus am able to ask them for help if I get stuck and use them for insight on how to perfect my own sentiment analysis model. The only concern that I have for sentiment analysis as it pertains to my research is whether I’ll get good results. Because my sentiment analysis is being done to evaluate differences between how male actors and how female actors are evaluated, I fear that my model might not be able to distinguish between the gender types. That means I will have to input specific criteria based on gender, which will be a big headache. However, I can worry about that when the time comes. See you next week! 

More Posts

Comments:

All viewpoints are welcome but profane, threatening, disrespectful, or harassing comments will not be tolerated and are subject to moderation up to, and including, full deletion.

    sahib_b
    What kind of challenges might you face when dealing with mixed-gender reviews? How can you make sure that the sentiment analysis will tell the difference between performance-based judgements and language targeted towards certain genders?
    March 23, 2025 at 3:25 pm - Reply
    krish_s
    To clarify, my research doesn't have any mixed-gender reviews. Either the review focuses on the actor or the actress, never both. Now, something related that you might have been getting at with your question is the reviews (although they mostly talk about the actor and the character they embody) also reference the director and supporting actors. To address this limitation, I will run each review through a name entity recognition model to ensure that the sentiment analysis will run only on actors/actresses instead of other parts of the review. For your second question, the results of the sentiment analysis will be representative of language targeted towards certain genders, rather than performance-based judgements. This is because all the actors in my study (for their respective films) won the Oscars. So, the expectation is that all reviews should talk about these performances in the same (positive) way because they are ALL Oscar-winning performances. For this same reason, any differences would indicate bias.
    March 24, 2025 at 11:47 pm - Reply

Leave a Reply

Your email address will not be published. Required fields are marked *