The challenges of Big Data study Thinking about personality types



[ad_1]

New research using Big Data suggests that psychological paradigms built on personality types need to be revised.

In the study, researchers at Northwestern University analyzed data from more than 1.5 million respondents to the questionnaire. The review found that there are at least four distinct groups of personality types: average, reserved, egocentric and pattern.

The results, which challenge existing paradigms in psychology, are published in the journal Human behavior nature. The study leader, Dr. Luís Amaral of the McCormick School of Engineering, believes that new perspectives may be of interest to hiring managers and mental health providers.

"People have been trying to classify personality types since the time of Hippocrates, but the previous scientific literature found this absurd," said co-author Dr. William Revelle, a professor of psychology at the University of Toronto. Weinberg College of Arts and Sciences.

"Now these data show that there are higher densities of certain personality types," said Revelle, specializing in personality measurement, theory, and research.

At first, however, Revelle was skeptical about the premise of the study. The concept of personality types remains controversial in psychology, with scientific evidence difficult to find. Previous attempts based on small research groups have produced results that are often not reproducible.

"Personality types only existed in the self-help literature and had no place in scientific journals," said Amaral. "Now we think that will change because of this study."

The new research combined an alternative computing approach with data from four standardized questionnaires with more than 1.5 million respondents from around the world.

The data was obtained from John Johnson's IPIP-NEO questionnaire with 120 and 300 elements respectively, myPersonality and BBC Big Personality Test projects.

The questionnaires, developed by the research community over the decades, have between 44 and 300 questions. People voluntarily take online questionnaires attracted by the opportunity to receive comments about their own personality. These data are now available to other researchers for independent analysis.

"What's really, really cool, is that a study with a data set of this size would not have been possible before the web," said Amaral.

"In the past, perhaps the researchers would recruit undergraduates on campus and perhaps receive a few hundred people. Now we have all these online resources available, and now the data is shared.

From these robust data sets, the team drew the five broadly accepted fundamental personality traits: neurosis, extroversion, openness, agreeableness, and dedication.

After developing new algorithms, four groups emerged:

Average
The average people are rich in neurotism and extroversion, but not very open. "I would expect the typical person to be in this group," said Martin Gerlach, a postdoctoral fellow at Amaral's lab and the journal's first author. Women are more likely than men to fall into the average type.

Reserve
The Reserved type is emotionally stable, but not open or neurotic. They are not particularly extroverted but are rather pleasant and conscientious.

Role models
The models have a low neurosis score and a high level in all other traits. The probability that someone is a model increases considerably with age. "These are people who are reliable and open to new ideas," said Amaral. "They are good people to take care of things. In fact, life is easier if you treat more with models. More women than men are likely to be role models.

Self-centered
Self-centered people score very high on extroversion and below average in terms of openness, friendliness and awareness. "These are people you do not want to hang out with," said Revelle. There is a dramatic decrease in the number of egocentric types as people age, both in women and in men.

The large dataset needed fine-tuning because the group's first attempt to sort the data using traditional clustering algorithms yielded inaccurate results, Amaral said.

"At first they came to me with 16 personality types, and there is enough literature that I'm aware of that says it's ridiculous," Revelle said. "I thought there was none at all."

He challenged Amaral and Gerlach to refine their data.

Their algorithm first looked for many clusters using traditional clustering methods, and then reconverted them by imposing additional constraints. This procedure revealed the four groups reported.

"The data came back and they continued to have the same four clusters of higher density and higher densities than you'd expected by chance, and you can show by replication that this is statistically improbable." said Revelle.

"I like the data and I believe in these results," he added. "Methodology is the essence of the journal's contribution to science."

To ensure that new type groups were accurate, researchers used a notoriously self-centered group – adolescents – to validate their information.

"We know that teenagers behave in an egocentric way," said Amaral. "If the data was correct and sorted according to demographic data, it would turn out to be the largest cluster of people."

Indeed, young men are overrepresented in the self-centered group, while women over 15 are significantly underrepresented.

In addition to being a tool that can help mental health providers assess personality types with extreme traits, Amaral said the study's findings could be useful to managers looking to ensure a potential candidate or to people who go out together. for a suitable partner.

And good news for teen parents everywhere: as people grow up, their personality types change often. For example, older people tend to be less neurotic but more conscientious and pleasant than those under 20 years of age.

"When we look at large groups of people, it is clear that some trends may change some of these characteristics over time," said Amaral. "This could be the subject of future research."

Source: Northwestern University

Related Articles

[ad_2]
Source link