Good Bad And Ugly: (mis)Use of data


Image Credit:

My readings through the month takes me to some fascinating destinations. This month I focus on the types of research I come across.

Almost every other coffee table conversation takes you to some topic where one or the other person tells you about something that is substantiated by research. Somehow, I find myself starting with cynicism at the mention of research.

This started with something a doctor mentioned to me 5-7 years ago, “Salt intake has nothing to do with your blood pressure. That is old school thought, new research proves that it is XYZ that matters and not salt intake.” 15 years ago same doctor might have scoffed at you for your pickle consumption due to the salt content.

Elsewhere, I read that the whole right-brain/left-brain theory that one side is responsible for logic, practicality etc and the other for emotions, creativity, etc is not true.

What am I as a consumer supposed to believe? Everything was based on research. This is still tolerable, at least new studies are being conducted and we are learning.

The Bad: Random studies that assume they are solving world problems

The scary thing is with big data, there is so much data available that you can draw pretty much whatever conclusion you have in mind. The more generic the conclusion, the easier it has become to find data to support it. There is all kinds of random analysis being made under the name of research.

For example, this study correlates how mobile phone data reveals food consumption in Central Africa. 

The main finding of this study is as follows –

expenditure in mobile phone top up is proportional to the expenditure in food in the markets

This conclusion is based on just the fact that the mobile phone top ups are the same in the geographical areas where high-value food items are consumed.

As if the randomness in the correlation isn’t bad enough, the article ends with “useful stuff!” in terms of finding solutions to world food shortage problems.

The relatively good: A study that says you can remember more by using technology.

Then you have studies that at the very least won’t harm anyone. This one goes on to say that you can use forgetting to help you remember more. 

Clicking ‘save’ on a digital file makes your memory worse for that information, but improves it for what you learn subsequently.

So, at the very least this one encourages you to make lists and use your phone alarm instead of trying to remember every little thing and over-taxing your brain. Maybe, some other study somewhere will tell you that by not remembering without a reminder you closed down nerver pathways that were responsible to sharpen your memory. But, at the very least you will have a “saved for later” folder to help your failing memory.

The ugly or the beautiful: Analysis of genre data from Netflix gives you a wealth of information

A 36-page training document teaches paid people at Netflix to watch and rate movies on a whole range of parameters – from parental guidance to one of 76,897 micro-genres. Considering that categorizing films is a Herculean task, I was happy that someone is doing the job with such attention to detail. Only to find out that this data is being used to reverse engineer movies to suit the viewer’s taste. While it is great that more movie-makers will make movies as per my taste, isn’t it a death of creativity and growth? Will this also mean fewer filmmakers will follow their heart and passion and bring a world to me that I had never imagined?

The eyerolls

And then there is research you see and wonder if there could be better use of your time reading through the study, let alone of the time, effort and energy of the people conducting the study. Cases in point –

This study wants to prove that “1 human year = 7 dog years” is a myth. The only question that I had was, “how does it matter?” 

This one analyses the game of “rock-paper-scissors”. I am just going with, I want to be less cynical this festive season and so I hope this one is useful in understanding the human mind better and/or finding something revolutionary in game/probability theory. Oh wait, that will mean more research and studies that might or might not be based on a good sample size and might just mean means to an end of some pre-decided conclusion.

Of course, I’m not saying research shouldn’t be done. I am just wondering about the misuse of analysis just because data is available.

So, have you come across studies that sound frivolous to begin with but led to a conclusion that was not so? Have you been frustrated with research that has led to exact opposite conclusions?

Meeta Kabra

Meetu is a Chartered Accountant and an MBA but she’d rather not keep books or run a business. She deployed her analytical skills to reviewing movies instead and, along the way, rediscovered her sense of humor. By doing this she gets to exploit both her love for movies and writing. She took it upon herself to write reviews, "Without Giving the Movie Away" @ Wogma. She also writes short stories and poems at Minus i.

Leave a comment

Your email address will not be published. Required fields are marked *