While making decisions, people often say it happened “naturally”; however, there are complex mechanisms in our brains at the time. The human brain works like a computer processor that receives, processes, stores, modifies, and transmits information but sometimes freezes. Dr. Gerda Ana Melnik-Leroy, a researcher at the prestigious École Normale in Paris and a researcher at the Data Science and Technology Institute (DSTI) of the Faculty of Mathematics and Informatics (MIF) at Vilnius University (VU), believes that serious and comprehensive research requires analyzing a problem or object of study from the perspectives of different sciences.
In 2020, the COVID-19 information graphs and illustrations have been used extensively around the world as data scientists and journalists have shifted to tracking and presenting information about the pandemic - from infection and death rates to vaccination data and its variables. Policymakers also relied on COVID-19 data and charts to make important decisions. Dr. G. A. Melnik-Leroy points out that the general public has significant gaps in their understanding of statistics and the interpretation of graphs, reinforcing the tendency towards cognitive biases, which gives grounds for numerous threats. which gives grounds for numerous threats.
To better understand how the human brain perceives and processes incoming information, where processing errors arise from, whether the processing is highly accurate, and how this can affect behavior, the VU researcher is currently conducting a study entitled Cognitive Mechanisms of Information Processing: Numerical and Linguistic Information. The research has already landed her a Lithuanian Academy of Sciences (LAS) scholarship.
“There are, indeed, many dangers. One is that people are too confident in their own decisions and knowledge. Research has shown that people tend to overestimate their abilities. The fact that we are shown a lot of data, it is explained to us, and then in a few months, we feel like experts - that seems to be the problem, because we probably still know very little. Especially since even the experts don’t know everything about a new virus like COVID-19,” the researcher names the first threat.
Another major danger, she says, is that it’s very easy to manipulate both numerical and graphical data. Marketing has been using it for a very long time, sometimes without even knowing the mechanisms behind it but knowing it works: “For example, displaying a graph on one scale or another can paint a particular picture. When we choose a certain design, our brains are automatically triggered and pick up the information without us even realizing it. When we see a graphic, we rarely think it should have been illustrated this way and not that way. Our brains don’t have time for that. We see something and come to conclusions.”
Why do cognitive biases exist?
One hypothesis as to why cognitive biases exist is that they are not just an error of nature but a special mechanism that facilitates our daily lives. Humans have to make many decisions with split-second precision every day, but it would take a lot of time if we did things at 100% every time. In many situations, the most effective way is to make decisions without overthinking; otherwise, we would freeze and not get on with our daily work.
“The assumption is that our cognitive mechanisms are designed to calculate things roughly, and in most situations in life, this is perfectly sufficient. However, there are rare cases where this does not work. When we have those cases with data, with statistics, we need that precision, but our mechanisms don’t work that way unless we make some reasonable, concrete efforts,” says Dr. G. A. Melnik-Leroy.
Interestingly, these biases are neutral, meaning they are cross-cultural, not differentiated by gender or age category, and education has little impact on them. According to the VU researcher, the consequences of the bias were particularly evident when people were simply bombarded with data on COVID-19, climate change, and war.
“We have become a data-driven society as if we were all data scientists. In this context, these biases have become very pronounced,” she says.
For years, scientists have been trying to find out if there are patterns in people’s behavior and decision-making. In short, at what point in performing a specific and repetitive task does the human brain, compared to a computer, seem to get stuck, to freeze.
“For example, people will choose to have surgery in a clinic with a 95% chance of success over one with a 5% mortality rate. Even though they are, in fact, exactly the same,” Dr. G. A. Melnik-Leroy gives an example.
Is objectivity even possible?
Much of the data visualization that bombards us today is sometimes just decoration and at worst a distraction or even misinformation; however, some cases highlight the scale of the problem and draw public attention.
One example is a graphic by Simon Scarr, Senior Designer at Thomsoms Reuters. The graph shows the number of deaths in Iraq each month from 2003 to 2011. It is also an inverted bar graph: the higher the number of deaths in a given month, the further down the bars go. S. Scarr has chosen the color red, which means that the whole graph looks like blood running down the page. In case the message was ambiguous, the chart was titled Iraq’s Bloody Toll. But another data visualization expert, Andy Cotgreave, saw this chart and did a little experiment. First, he recolored the graph by presenting the same columns in a cold blue color. Then, he turned the chart upside down. Eventually, he changed the title to Iraq: Deaths on the Decline.
The change in emotional impact when looking at the graphs is drastic. Which chart is better? This depends on the message behind it. But there is another question. Is objectivity even possible when we present graphical information to the general public?
“I think knowing how human perception of information works, and having the goodwill to use that knowledge, it would be possible to create something approaching the optimal option,” Dr. G. A. Melnik-Leroy does not rule out the possibility of complete objectivity.
“We are currently researching different types of graphs and looking at how people respond to them to understand where one type of graph is more useful than another. There are two things. One is natural perception. For example, we perceive red as a warm color, blue as a cold color, etc. If, for example, we show low temperatures as red and high temperatures as blue, the perception seems to freeze there.
Some are innate reflexes and some are cultural. But when we talk about statistics and data, we are talking about knowledge: math, statistics, etc. That knowledge is often very scarce. So, mathematical knowledge can help us cope with these biases if we have a solid cognitive mechanism, but if we don’t, our natural perceptions take over and can distort the information in certain situations. After that, everything happens in a chain reaction based on how we take in information ad how it affects our behavior. Studies show that behavior is strongly affected. Suppose you have seen the same information several times and misinterpreted it. This will only reinforce the tendency to behave in a certain way, which may not necessarily be rational,” the researcher says.
Challenges for women in science exist
Dr. G. A. Melnik-Leroy returned to Lithuania with the aim of using her knowledge of cognitive science to enrich mathematical algorithms, software, models, and even artificial intelligence systems. In addition to the desire to promote interdisciplinary interaction between the social sciences and the exact sciences, the researcher also notes other cross-cultural differences between Lithuania and France.
One of the main challenges is balancing career and motherhood: “When I returned to Lithuania, I was positively shocked by the joy surrounding women having children. In France, I worked in a high-level laboratory, but this topic was taboo. One PhD student got pregnant, and I saw them looking at her like she was a leper. Talking about wanting children was like betraying science.
Maybe it’s a cultural thing, but in Lithuania, at least personally, I haven’t seen or heard such things. My supervisor always supported me, and I knew it would be no tragedy if I announced I was pregnant, which would have been the case in France, for example. This aspect is psychologically significant,” she recalls.
However, the researcher does not hide the fact that there is still room for improvement in Lithuania, and there is still pressure on women in society to choose between having children and having a career. But women are often more likely than men to convince themselves they have to choose. This just goes to show that the problem still exists in society.
“There was a study on knowledge of math. The groups of young women and men were given identical math exam papers. In one case, the groups were told nothing at all, and in the other, it was mentioned that the exam is difficult and some will find it hard to pass. The study showed that the men were unaffected by the cue, but the differences between the women groups were striking. I think it cuts across the board,” she says.