I was not a fan of this study to be honest. The methods of information gathering seemed extremely subjective and I’m not sure that anyone can take it that seriously. Even the author admits that they have to make a lot of estimations and concessions just for the sake of completing the study. User “Omaji” spoke well when they said the data has a large human social depth. I’m just not sure that it’s that reliable. Underreporting is something that can’t be accounted for as well as the other variables mentioned and not mentioned by the author. In summary, I just find the whole thing misleading.

I found two articles of the selection for the week quite interesting.  The extract on psycho-history gives a bit of a solution to how we can make data collection and use less biased. If studying the human sources and collectors of the data alongside data analysis was a norm, then maybe the individual perspectives and opinions that generate bias could be understood and the extremeness of these biased views could determine whether or not the data collected is valid enough for public use. The article by Michael Martell on the wage gap between men of different sexual orientations was also very insightful. The methods the author used just to ensure that other factors, that were worth considering in the general analysis of the entire situation, were looked at alongside the data are other ways that we could move away from the collection and use of biased and general data.

In “Differences Do Not Matter: Exploring the Wage Gap for Same-Sex Behaving Men” Michael E. Martel explores the previously studied statistics that confirm gay men make noticeably less than heterosexual men, and emphasizes that the data proves that this wage gap is in no way created by difference in skill level. In this article GSS and Census data is used to make initial claims that the existing wage gap is present, there is a table that references multiple studies that find that a penalty in earnings comes with being gay. I think the presence of the table really emphasizes the wage gap and is an effective use of data to ensure a picture is painted for the reader of what the issue is. It is stated in the reading that “[they]  are interested in differential treatment that gay men experience at work. Those who identify as gay are more likely to indicate to employers and coworkers that they are gay [Carpenter 2005],” (Martel 39). I think this claim really makes there assumptions of the sources of the wage gap more valid because they are using data that has obvious correlation. They then go on to compare it to the wage gap with minorities which also emphasizes the problem that is at stake. The conclusion reiterates the point that the wage gap is clearly evident between gay men and heterosexual men. I think this article used data in a really straightforward way that supported their claims and their assumptions about where this wage gap comes from, they were able to support their reasons for this claim, that were not clear, with data that was correct. One of my classmates made the claim “misses out a big chunk of potential data from people who aren’t in a cohabiting relationship” and although I think this is a good point that even if a gay man is not in a cohabiting he should be recognized in the workplace. Although I think this is a valid claim I believe that they left that group of people out because they wanted to ensure that the data was from a group of extremely similar people and a cohabiting relationship is what they decided would be a factor of a similar group of people that would provide good supporting data.

Hey again, checking in here to see if this thing is even working.

In the Pew analysis we see the importance of not only mentioning one source of data. The decrease in recidivism in the United States was the topic and main focus of the analysis made. However, the Pew article mentioned important data trends that both directly and indirectly influence the argument made. The article mentions both the decrease in crime rates on the state level and the decrease in incarceration which is vital data information to consider when analyzing the recidivism in the U.S. So how does this relate to the wage gap of gay employees? In the study hundreds of pieces of information are mentioned that standing alone would have no relevance. The study doesn’t only look at the wage gap between different-sex males and same-sex males, but it analysis how that wage gap is compared and influence by other wage gaps — such as women or black workers. Looking back at Julia Middlebrook’s post from last week, she mentions how we can answer macroeconomics questions by looking at microeconomic trends. Approaching data from different angles, as she mentions, and as we see in the studies discussed above is vital to understanding data at a deeper level.

The Martell article “Differences Do Not Matter” highlights the struggles of collecting data for under-represented/discriminated groups. Something that was very crucial in collecting accurate representational data of homosexual men was the defining what a homosexual man is. An example they gave was avoiding misclassifying heterosexuals who have had same sex as gay man, and instead only look into data where individuals had self-identified themselves as gay. This reminds me of something that Michael Somkuti had pointed out in his “Psychohistory” post that spoke about modelers having to be conscious of the type of data and parameters they are using in order to produce sound outputs. I think this perfectly fits with this article and being conscious here is even more important than it would otherwise be, as you are dealing with a sexual identity. Data thus has to be consciously collected because the ramifications on the population that identify with this sexual identity are at stake.

The PEW study on recidivism rates dropping from 2005-2015 paints a wonderful picture of an improving prison system, without enough data to back it up. There are plenty of states where the 3 year recidivism rate has dropped over the past 10 years, but they only showed/used data from 23… which is less than half of the country. It is hard to completely dismiss the findings though, and because of that we can see that programs aimed to help inmates upon release seem to be working. The goal of these programs is to save the taxpayer money by changing the behavior of inmates, so that once they are released, they are able to reintegrate into society. I think the article is made less trustworthy by showing how little data they used, which is unfortunate because it’s a feel good story, and you’d like to believe we’re making progress.

Reading the Pew research on recidivism, and seeing loads of important info missing (tons of states, people who died in the examined period, people who got arrested in other states etc.), my first instinct was to shout “False! The data and the research is shady! Case closed!” (a bit like Jonathan in his response, but less eloquently). But then I thought of something that came up in a group discussion in class in relation to the Eviction Lab reading, namely, that sometimes you just have to do the best you can with the data that’s available, and maybe having stats from lackluster data is better than not having any.

Therefore I think we should think about the responsibility of government institutions in compiling and publishing reliable data on important issues such as recidivism.

The Martell and Pew articles, each gathered mass amounts of information from databases to analyze and make claims. However, in both instances there were major flaws in; the spread of the data, location of data collection, the follow-through of time related data (primarily seen in Pew) and failing to control all confounding variables. The Pew article stood out to me because the data collected for these prisons showed such high rates of recidivism, and there were still somewhere between 12-20 states that did not report. I think the percentage of released prisoners that reoffend is much higher than what the article explained, because of the underreported data. If a prison has extremely high rates of recidivism, they might not report that data because it is possible that their prison could lose funding, or be absorbed into another prison system.  Also, in both the methodology and the article, the authors acknowledged that Pew did not keep track of people who reoffended in a different state, which lessens their overall percentage.  Furthermore, the purpose of the article was to show the steady decline of recidivism, but in my opinion, there were too many uncontrolled variables to make that claim boldly enough to put in the title.

However, while mass data bases often have large information gaps, they do provide valuable information when used with caution, and with wide (and as random as possible) spreads of data.