A few months ago, a senior from my class told me that MBTI (Myers Briggs type indicator) can be a good choice to lead the conversation when going for a dinner with seniors. MBTI is a statistics-based table that distinguishes characteristics into 16 categories. It could be a very nice ice-breaker in an awkward situation between people meeting for the first time. Likewise, many Koreans use MBTI to start a conversation or to make a consensus between people who have the same MBTI. This “MBTI Fever” has been the main social topic in Korea for several years, but there are certain drawbacks; some people have too much faith in MBTI. They treat MBTI as science and judge people by MBTI, forming prejudices and stereotypes of a person. It begs the question, is it really scientific?
One’s MBTI is determined by answering several questions that relate to the precision of our mind and emotions. Sometimes there are questions under specific circumstances. These cognitive variations are then converted into four scores, each from one of the two styles of human characteristics: The attentions or attitudes are separated into “extraversion” or “introversion.”; Perceiving functions are separated into “sensing” or “intuition.”; Judging functions are separated into “thinking” or “feeling.”; Lifestyle preferences are separated into “judging” or “perceiving.” The scores of each question signify where each respondent is close to these human characteristic features.
MBTI is based on statistics and probability, partitioning people’s features into 16 types. So, the first question is, can a statistic-based technique be a science? Many types of science are based on statistics and theories. For example, statistical mechanics, quantum mechanics, and other fields use probability and statistics. But we must consider the fact that “Four pillars of destiny” and “Feng shui-geological study” in eastern Asia are also formed from statistics. Then, what is the difference between statistic-based science and statistic-based mythology? 
I believe this distinction originates from the difference in probability types. Usually, there are three types of probability: theory-based mathematical prob-
ability, observation-based statis-
tical probability, and geometrical probability. Similar to probability, a statistic also can be classified into theoretical and observational. The theoretical statistic is like quantum mechanics, predicting the micro world’s particles by the Schrödinger equation, calculating and indicating their location as a probability density. Based on this, data is calibrated, which means that there should be certain relationships and that coefficients must exist. However, the observational statistics are different. We cannot declare that our measured data always have a relationship with each other. In short, theoretical statistics are made from “causation” and measured statistics are made from “result”. So, observational statistics come from reverse inference. Inferring cause and theory from the result is not always accurate. So, we cannot confirm MBTI as a science.
MBTI can be a nice choice to establish relationships with people but that does not mean that we can judge people by MBTI. Inferring the cause from an effect is “wag the dog”. Of course, it is based on big data from all over the world. So, in most cases, MBTI might work. But we cannot find out the exact reason and relationship between MBTI’s every feature. In short, I want to say: Do not trust MBTI too much. Sometimes it can be harmful to your social life.