Has data credibility become a big problem?

Is data credibility something we should worry about?

It’s in our nature as humans to make sense of the world and environment surrounding us. We have a phylogenetic predisposition to categorize and prioritize the objects around us. Based on this information we make predictions about the future. Sometimes the predictions are correct, sometimes they are not, so who is to blame? If we think about the human cognitive system we know that all the information from the external stimuli is processed through our experience filter, the data is described, inferred and evaluated. Sometimes our cognitive system fails us and makes inaccurate and false assumptions regarding a future event because our prior events had a strong emotional impact on us and unbalanced our logical, empirical and pragmatic system.

Now this is just an analogy regarding an imperfect system like our brain, but what about the digital data that governs and surrounds our world, the data you find in a financial report, a medical journal, a product spec, a laboratory measurement, how come we find errors in these perfect systems, systems that do not feel emotions, perfect systems that are built by an imperfect being. The answer is really simple, it’s human nature, we use our machines, soft-wares and programs to calculate, analyze and organize the data, but we are the ones that introduce it, we make inferences and predictions using them and we are the ones interpreting the outcome. Sometimes the initial data is correct, but the interpretation is not and the consequences can be critical, for example an incorrect laboratory measurement could kill a patient, we can make a bad investment, an airplane can crash because the weight ratio was calculated incorrectly, we can get lost if the information in the maps data base is wrong etc. Now we live in a very complex world where we are technology dependent, the biggest corporations in the world rely on digital data, but this is not an exact science and sometimes errors appear that affect the development and progress of an institution and people running it.

If there are repeated incidents regarding the data credibility problem, managers start to lose faith in data and they start relying on intuition in order to make good and smart decisions and implement the necessary plan of action. The probability of success is reduced because instead of using statistical methods like correlation or regression to analyze the data, they start relying on their intuition, experience, sometimes this strategy can be effective, but not so often than we think, because we have a expectancy bias mechanism in our brain that enables us to be more subjective to the data in front of us instead of being objective, calm and calculated. So we have two choices, try to clean up the existing bad data, correct all the errors, replace the garbage data with the correct one or try something more pragmatic like improving the way new data is created.

We could use some strategies to avoid these issues. For example we should focus on the problem in hand, identifying and correcting errors and seeking confirmatory sources can be exhausting for the workers, it’s time consuming and difficult to handle because sometimes the workers don’t understand how the data is used and implemented, so when the error occurs they resort to a problem focused approach and try to work around it without assessing the underlying causes. If we want this issue resolved we should focus more on the people working in our company, we should invest in creating an interface language that every employee understands and can be implemented while respecting the technological standards. So we don’t really need better technology, better hardware or faster and advanced software but better professionals.

Categorised in: Did you know