Overview

The past week I have tried to complete the task where I have to find examples of the different types of bias in data. As said last time, they are good to know so you can be aware of the possible issues that can be made.

What was learned

These following bias’s have been have been talked about last week. But basically the following is a real world example of where this has occurred. These are also not all of the possible bias’s that could occur (because that would be heaps of different names/examples). Some examples of such:

• Response Bias: Twitter tweets were used to study people’s behaviour after Hurricane Sandy struck the US northeast. Hoping to understand the behaviour of users in the worst hit areas. However, those researchers later discovered that most of the data came from Manhattan. Very few tweets came from the more severely affected regions in New York. It happened because over time, power blackouts set in and then the phone batteries drained and thus even fewer tweets came from the worst hit areas.
• Selection Bias: Participants included in an influenza vaccine trial may be healthy young adults, whereas those who are most likely to receive the intervention in practice may be elderly and have many comorbidities, and are therefore not representative. The reason that this is not likely to occur often is because most in that sector know that targeting people that are heathy doesn’t mean the same as the unhealthy.
• Omitted Variable Bias: There was a case where the goal was to predict the probability of death (POD) for patients with pneumonia so that high-risk patients could be admitted to the hospital while low-risk patients are treated as outpatients. Realizing that it admitted people weirdly seemed anomalous, the researchers investigated further and learned that patients with a history of asthma who exhibited symptoms of pneumonia usually were admitted not only to the hospital but directly to the ICU (Intensive Care Unit).

Reflection

What lessons were learned from failure?

Some of the lessons that were learned from failing (or getting annoyed) were trying to find examples of bias. It was hard to find actual examples where an incident had occurred. But after some that patience I learned that patience is the best/only way to succeed when having issues. Luckily, I wouldn’t say that I have ‘failed’ yet but was more on the path to fail if I didn’t do something to overcome it.

What made you ?

I got curious about the case studies when learning about how many things can contribute to bias (when I actually found some). It is very intriguing how simple factors can affect a basic graph/study/experiment. Some things that are important when trying to overcome the bias when making your data representation is to become aware of the possible factors and then so your best to not become victim to it. Another thing that made me curious was thinking about how they could have fixed it (assuming it wasn’t intentionally done).

How did I hinder others?

Unfortunately, I think I may of hindered others in the week by not really answering their questions. At the start of the week some others were trying to find examples of the bias’s, however at that point of time I wasn’t bothered to do it. And then when I was bothered it was to late because it was the weekend. So to improve I could actually start the task on-time and when others are asking for help, so that I can possibly help them and have a better understanding for myself.