All entries for Tuesday 16 April 2013
April 16, 2013
RDM: Correlation vs Causation
When making a decision, we should always ensure that we are not confusing correlation with causation. This happens more often that we think it does, and even some newspapers are susceptible to interchanging the two in their articles. This, in some cases, can be extremely controversial, as we don't know their underlying motives. For example, many newspapers have written about studies that apparently show people who eat breakfast are less obese. This essentially points towards causation, however, it is more likely that in the given sample, this was just a mere correlation between the two things. Informed readers always think critically, and they might identify this as a deliberate ploy from the breakfast cereal companies in getting more people to buy their product.
The most that such articles can do is state that they found a correlation between eating breakfast and less obesity. And that even though, they APPEAR to be related, we cannot be sure. Hence, the statement that "people who eat breakfast are less obese" might be true or it might not be. We simply don't have enough information to be able to confidently support or reject that assertion.
The significance of this is that we should always be careful in judging whether there is a causal relationship between two things or whether its just simple correlation before making a decision based on it. In the case of a confirmation bias, this becomes even more important. We have to resist from blindly using data that confirms our idea, as it may only be showing the correlation between two completely unrelated things.
Hasan Khan
16 Apr 2013 04:22
| 
Tags: Rdm
| 
Comments (0)
| 
Report a problem
Please wait - comments are loading
RDM: Confirmation Bias
In my last blog, I wrote about some of the biases that affect decision-making. In addition, I stated that improved self-awareness can minimize the effects of different biases. However, it is also important to note that a well thought out decision-making process can equally have a positive effect of reducing the overall bias.
From my experience of the presentation, I have learnt that we should always have strategies in place that help us avoid any bias in the future. These, in most cases, need not be complicated. For example, in order to avoid availability bias, we should consciously check that we are not solely relying on memory and try to get any data that could be relevant. Similarly, to counter anchoring bias, I have learnt that we should always try to examine the problem from a number of perspectives. By analyzing the differences, we will become relatively more objective in our decision making.
However, there is one bias that I would like to highlight specifically. That is the confirmation bias. As students, we can recall falling in this trap a number of times. For example, we often think of something really clever to write in our assignments. Then, as soon as we find studies that even partly supports our proposition..we use it and portray it as solid evidence that proves or justifies our point. Whilst this sort of practice has a comparatively low risk attached to it in our academic lifes, I strongly believe that this sort of behaviour can lead to drastic consequences in the professional world. Companies can really end up making huge losses and it could lead to many employees losing their jobs. Hence, we should be particularly careful about our motives. We should always use data or any other evidence to better inform us rather than justify what we believe. We should also not stop gathering information that may or may not support our plan. This would ensure that we don't fall victim to the confirmation bias.
Hasan Khan
16 Apr 2013 03:51
| Tags: Rdm
| Comments (0)
| Report a problem
Loading…