Can Psychological Traits Be Inferred From Spending? Evidence From Transaction Data

Abstract

The automatic assessment of psychological traits from digital footprints allows researchers to study psychological traits at unprecedented scale and in settings of high ecological validity. In this research, we investigated whether spending records—a ubiquitous and universal form of digital footprint—can be used to infer psychological traits. We applied an ensemble machine-learning technique ( random-forest modeling) to a data set combining two million spending records from bank accounts with survey responses from the account holders ( N = 2,193). Our predictive accuracies were modest for the Big Five personality traits ( r = .15, corrected ρ = .21) but provided higher precision for specific traits, including materialism ( r = .33, corrected ρ = .42). We compared the predictive accuracy of these models with the predictive accuracy of alternative digital behaviors used in past research, including those observed on social media platforms, and we found that the predictive accuracies were relatively stable across socioeconomic groups and over time.

Using a supervised machine learning algorithm for detecting faking good in a personality self‐report

“…Machine learning algorithms were specifically developed for processing high dimensional, linear/nonlinear data like these (Marsland, ). Such approaches have been already adopted in psychology (e.g., Gladstone, Matz, & Lemaire, ; Park et al, ; Youyou, Kosinski, & Stillwell, ), but—to our knowledge—the present study is among the first attempts at using them in the evaluation of faking good, taking a similar approach of other recent papers (e.g., Dua & Bais, ; Goerigk et al, ).…”

Personality Research and Assessment in the Era of Machine Learning

“…First, ML models have been used to predict individuals' Big Five personality traits from a wide range of data sources; these sources include digital footprints from social media platforms (e.g., Facebook Likes, status updates, Kosinski, Stillwell, & Graepel, 2013;Youyou, Kosinski, & Stillwell, 2015), language samples (Park et al, 2015;Schwartz et al, 2013), spending records (Gladstone, Matz, & Lemaire, 2019), music preferences (Nave et al, 2018), and mobile sensing data (Chittaranjan, Blom, & Gatica-Perez, 2013;De Montjoye, Quoidbach, Robic, & Pentland, 2013;Hoppe, Loetscher, Morey, & Bulling, 2018;Mønsted, Mollgaard, & Mathiesen, 2018;Schoedel et al, 2018;Stachl et al, 2019;W. Wang et al, 2018).…”

“…In contrast to the general R 2 , which can be roughly thought of as a normalized version of the MSE, the squared Pearson correlation is a special linear rank measure. The Pearson correlation between predictions and observed criterion values, which is reported in a substantial number of publications (e.g, Youyou et al, 2015;Gladstone et al, 2019), mainly captures whether observations with higher values in the outcome also receive higher predictions and vice versa. It only weakly reflects how much the predicted scores differ from the observed criterion values.…”

“…The correction for attenuation formula, which is used in many ML applications in personality science (e.g., Gladstone et al, 2019;Youyou et al, 2015), implies that using manifest sum scores as criterion variable leads to underestimation of predictive performance. However, overestimation is also possible when systematic method bias is contained in the psychological measurements.…”

Untitled

“…Для прогноза личностных черт используются также особенности интернет-серфинга человека (Kosinski et al, 2014), структура социальных связей в социальных сетях (Quercia et al, 2011), характер его финансовых транзакций (Gladstone et al, 2019), а также данные, касающиеся использования смартфона (Stachl et al, 2019).…”