Skip links

5 Mistakes To Avoid When Analyzing Data

5 Mistakes To Avoid When Analyzing Data

If you have ever been in a dark room trying to feel your way around, and someone turns on the light, then you can relate to what insights discovery does for a decision maker. Good insights is to a decision maker what a compass is to an explorer. If insights are so awesome in “lightning up” your decision making path, you may ask, “How come many people don’t profit from it?” Good question. It turns out that drawing insights, even from good data can be subject to many errors and biases. The good news is that once you know these biases, you can be careful to avoid them, thereby drawing profitable insights from data.

Biases You Need To  Avoid

 

Survival Bias

Imagine that you are a second year student at your university, and you want to make a decision; Whether to complete your program, or to toe Bill Gates’ path, and that of many other illustrious names in silicon valley and drop out of school (BTW, the second option is destiny-truncating in present day Nigeria). If you go further, and think that many successful entrepreneurs dropped out, and hence you decide to drop out, you have just committed a survival bias. This is because you did not bother to think of the many determined and hardworking youths that dropped out and became a failure.
The survival bias  is the logical error of concentrating on the people or things that made it past some selection process and overlooking those that did not, typically because of their lack of visibility. In order to steer clear of the survival bias, you need to look beyond successes or survivals, and consider the cases that were failures or non-survivals.

 

Correlation is not Causation

This is known as the third-cause fallacy. Because two quantities move together such as the number of cars on the road, and cost of a bottle of coke does not mean that one is responsible for the other. When two quantities seem to be in agreement with each other, the presence of other quantities is usually responsible for the agreement. For instance, the increasing cars on the road, and the increasing cost of coke, can be explained by other quantities such as changing economic conditions, increased population, increase in the availability of cars, e.t.c.
Many people make this error when trying to explain the reason for an event.  It is important that you ask yourself which other quantity can make two quantities or variables move in agreement before stating that A causes B.

 

Anchoring Bias

In a widely read book, “Thinking fast and slow”, Nobel Laureate Daniel Kahneman documented the anchoring bias extensively. The “anchoring effect” names our tendency to be influenced by irrelevant numbers.  As an example, most people, when asked whether Gandhi was more than 114 years old when he died, will provide a much larger estimate of his age at death than others who were asked whether Gandhi was more or less than 35 years old.
The anchoring bias can be difficult to avoid, but a good practice is to compare your value of interest to reasonable and flexible bench marks as opposed to hard numbers you are used to.

 

Availability Bias

Traveling by air is the safest means of travel according to statistics. It is safer than motor vehicles, trains and boats. However, many people think is dangerous when they remember two or three plane crashes. When there are a few reports of  something deemed rare happening such as a miracle or winning a lottery, it is human nature to think that the rare event is not so rare after all, and is more possible than was thought to be. The availability bias is an important tool in the hands of the media to influence the society.If a story about child abductions is aired, people will begin to panic, thinking child abductions are happening with more frequency.
The availability bias makes us overestimate probabilities and end up with poor decisions. This bias can be overcome by basing your decision on enough examples of the event in question, instead of just taking a few positive examples and rushing to make a poor decision.

 

Illusion of Validity

If a dude is always on his laptop and likes mathematics you may think he is a geek. Right here is the root of the illusion of validity. The said dude could equally be a maths teacher or a graphics designer. Just because something matches a stereotype doesn’t mean we should make a judgment based on the stereotype.
Avoiding this bias is possible if you are aware of it and you actively resist it,  and then you go on to get enough information about your event of interest.

 

I must confess, even after considerable expereince with analyzing data, I still get blinded by theses biases sometimes. It takes effort and increased awareness to get rid of the biases and analyze with clarity. The only way is to practice practice practice !

Join the Discussion