Cathy O’Neil discusses the problems and ethics with using data science in her book Weapons of Math Destruction. In one chapter she discusses the U.S. News college ranking, it’s development, and it’s consequences. It was the choice of what was deemed important for a college to have in order for the student to succeed that I see as the problem. A problem with this is that people were trying to cheat the system and raise their score without really deserving it. Those who weren’t trying to cheat the system rarely saw an improvement while prestiges schools didn’t have to do much to stay on top because they are seen as really good in the eyes of the people and since that was also a factor, they hardly got booted down at the beginning of it’s implementation. The main idea from this chapter that I got was that not all models are great because they might not take into account everything or they scale certain things to be more important than others. The choice of what should be seen as important is subjective and should be changed based on the person’s intention with the ranking and with what they need. They talked about the people in charge of their
In another chapter she talks about the development of a survey that clears the way for the application that will best suit the companies needs. It stops applicants from progressing to the next stage in the hiring process, which is actually meeting a human person, by filtering out those who give answers that when grouped together prove they are a bad fit for the position. The example provided by O’Neil was of a man diagnosed with Bipolar Disorder who was screened out because his answers suggested that he had tendencies that weren’t desired by the people he was applying for. His honesty when answering those questions hurt him without ever having to tell him the reason he didn’t move onto the next level. He didn’t learned the reason for not getting a call back and was just ghosted. It was the idea of his behavior due to his mental illness that stopped him and that’s not allowed. Because they didn’t straight up tell him and he never met a real person, they were able to get away with this, until he was told about the survey screening process by a friend that already worked there. This survey is one of many ways take out the human interaction at the beginning of the application process in order to not have to pay HR more money to interview them. Capitalism is dehumanizing. Corporate heads don’t see people as people and instead are trying to find a way to make the most profit. And it’s not just that the people aren’t a good fit, it’s because the programs are set on figuring out which questions are being graded as “does this person seem to have a personality that will leave if they get bored with the job” or “ is this person too much of a freethinker, do they consider themselves to be unique” which is a con because they think that being unique means that they are too conceited - I think I’m going on a tangent so I’ll stop here with this, but it still applies.
In regards to the Chicago Crime project, I would say that when/if looking at how often crimes happen in certain counties and making a statement about which counties seem safer could be miss leading because it would be taking into consideration the amount of people that live in those counties. It is safe to say that in counties with larger populations we would expect there to be more crimes than a county with a smaller population. But without making that difference in population obvious, it could lead to skewing people’s perception about the counties.