In the journal article “The effects of low-pay and unemployment on psychological well being: a logistic regression approach” Theodossiou explores the correlation between unemployment and mental health issues. They explored the issue mainly based on people’s answers to a series of questions, which makes me wonder of people are unhappy because of their unemployment or if they are unhappy because they are most likely surrounded by other unemployed people so maybe it has become a consensus that they are unhappy instead of individual feelings of unhappiness. I think is is important how the article addresses why unemployment might make people so unhappy, it reads “it may be a source of prestige and social recognition, a basis for self-respect and sense of worth, an opportunity for social participation or merely a way of earning a living.” This proves that if people are unemployed they are typically unsatisfied and allows data to explore to what level they are unsatisfied. I think it is also important that they look at the subgroups, as we know from the Simpsons Paradox reading they can paint a different picture, in this case they seem to get the same result but with different reasoning, for example young people are more unhappy out of boredom and lack of purpose instead of out of stress. I think this article does a complete job of addressing how unemployment can affect many different aspects of mental health, I think they use there data well to prove multiple different correlations that are present between this data set. I think this was one of the more comprehensive articles in collecting data and addressing the concrete conclusions that can be made from it. In the simpson’s paradox blog from last week the question was raised “Is this a connection that humans are forcing or something that happens on its own?” In regards to correlation and I think in this particular study they were able to present accurate and enough information that the correlation between unemployment and unhappiness is one that is really evident and not one that humans are forcing.

In this paper published by Timothy Guinnane and Cormac Grada, “Mortality in the North Dublin Union during the Great Famine”, the authors intend to show a correlation between the mortality rates in workhouses in Ireland, and the mortality rates due to the Great Famine, an incident that only plagued essentially people of lower class. While reading this study, I found it quite eye opening that they only used the mortality rates from only one workhouse, instead of many to minimize outliers. One of my fellow classmates brought up an interesting point as well, writing that, “Data, here too, could be vary depending on perspective”. This brings up a new point that I hadn’t really thought about until now. If the people directing the study wanted to make a point, couldn’t they just choose one specific workhouse in order to prove their point?

In the article,  “Mortality in the North Dublin Union during the Great Famine,” the authors bring up the idea that underlying factors should be considered when we interpret data given to us. They give the example of workhouses during the famine. While one workhouse could have an impecable survival rate, it could be because they had less severe cases to attend to. On the other hand, a workhouse that had a high percentage of deaths could have had people near-death come to them, making their jobs a lot harder. An idea like this- that data is dependent on the individual’s situation- should be considered when we analyze all data. Background information should always be given so we have context and can approach the data with full knowledge on the situation. Without it, people could assume that those workhouses with better survival rates are much better than those with lower rates, when in actuality we don’t know if they are better due to the differences in the cases they received.

Pressure is one of worst emotions to exist because it can only create stress  levels to increase and people not being able to properly function. Without the ability to feel free and liberated, productivity and living a healthy life is not attainable. Guinnane and O Grada piece on Mortality in the North Dublin Union talked about the conditions of the workhouses in Dublin.  In the article it explicitly talks about how these working conditions explicitly affect productivity and mortality rates. While reading this article is became very apparent to me how similar this event in history was to Lewiston around the same time period. I found the post, “Guinnane and O Grada…The Power of Data in history”  to be very informational because it really honed in on the power of data  and  why people should study and research these sets. Data also is very critical to these types of situations because it may offer some solutions to enhance performance and answers the question “why?” to a lot of problems that are way beyond the surface. 

This article got me thinking a lot about how history can be used as data. The journals, facts and numerical data from the past gives an understanding of what life was like, even though some of it may just be numbers or names on a piece of paper. This article gave another example of categorized data based on the past, similar to the data set we are using in class. It helped me understand the power of a historic data set, and was very interesting to learn about.

Data can get incredibly complex. Especially when different perspectives present countering data points, it gets difficult to identify which information is most accurate. Simpson’s Paradox is an example of how data, recorded from two ends, spits out completely different results. Studying the acceptance rates for graduate school, data shows that less women get accepted than man. However, from the perspective of the university, the acceptance rate for women is higher than that for men. Simpson’s paradox is a prime example of why data should never be considered from just one perspective, nor should it be recorded to a limited extent. From the study on mortality rates in North Dublin workhouses we see similarities in data complexity. During the height of the famine a high proportion of “immigrants” were dying in workhouses compared to “Dubliners”. Data, here too, could be vary depending on perspective. Immigrants higher mortality could be defined by the poor treatment and malnutrition of the outsiders. But what if we looked at their origin? Through that we can look at how foreign genetic pools may disadvantage immigrants’ immunity against the given virus circulating Dublin during the famine. Here, different perspectives give different results; and these are trends we see in many data models. Last week Elizabeth Cullen discusses how the development of data models is often overlooked. Mentioning the influence data recorders can have on data outcomes she writes: “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. Considering this and Simpson’s Paradox, as individuals absorbing data, we must remember the complexities and biases of data models.

I find the Idea of a Simpon’s Paradox fascinating because it represents one way that humans can incorrectly interpret data. It may simply be counterintuitive and not a paradox, but most people are incapable of telling the difference. Something that is true when the whole population is considered seems as though it should remain true within subgroups, which tends to be the case with most things encountered in everyday life. This casual inference is hard to combat because Simpon’s Paradox is rare enough that most people do not consider, but common enough to be more than a statistical anomaly.

 The authors argue that one should look for Simpon’s Paradox in fields they believe to be most relevant, but I did not see anything in the paper to suggest that I should trust that their claim of which studies are the most vulnerable. I believe most snapshot psychology studies to be inherently flawed because people are not static things and change both day-to-day and year to year. Psychological studies that about cognitive subjects like memory are acceptable at the snapshot level, but still vulnerable to Simpon’s Paradox.

I believe using programs like R to detect Simpon’s Paradox are useful tools but are not a viable solution. The issue does not stem from the data or programs that organize them, but the human that processes the data. A human brain can metaphorically be compared to a computer, but one has to remember that the human brain was created by accident. Learning how to compensate for the errors in human software would be more beneficial than trying to create a program to do it for us. For a program to absolve data of human errors and biases could be possible with a neuro-network, but both collection and interpretation of data would have to be done by AI, which as a general human enthusiast I believe to be more problematic than simply compensating for our species defects.

The Simpson’s Paradox is interesting when related to large corporations or news sources, because many times the paradox is manipulated in their favor. This paradox basically says that when many subsets of data are expressed, the individual subsets may have specific characteristics, but expressed as one big data set they may have the opposite characteristics.  When you don’t consider individual variables, there are pieces of the picture missing from the data set. News sources do this all the time when reporting data for state or even national and political reasons. Relating this back to the Pew article about recidivism rates, if the database reported that the recidivism rates were declining nationwide, they may be completely be eliminating the fact that they had been increasing in half of the states across the country. I like how in this source about the paradox, the author acknowledged that there needs to be treatments to fix the biased data in a good statistical model. Many of the treatments involve separating the variables into subsets or clusters. This changes the entire outcome of the data, but it is more accurate.