Data and Code for the 2019 SIOP Machine Learning Competition

Overview

• Competition Overview and Awards Presentation
• Competition Portal

Objective

• Predict self-report personality traits from open-ended text of 5 SJIs.

Data

• Data was especially created and collected for this compeition.
• Predictors. Predictors are text responses to five open-ended situational judgment items (SJIs) designed to elicit trait-relevant behaviors. Exact question text in the (full_data readme)
• Critera. The criteria are aggregate trait scores on a Big Five personality inventory (link to items?)
• Data splits. During the competition there were three splits: training set (n = 1088), dev set (n = 300), and testing set (n = 300).

Winners

First Place: Natural Selection

Josh Allen @ Walmart
Matthew Arsenault @ Walmart
Blaize Berry @ Walmart
David Futrell @ Capital One
Private Test Set Average Correlation = .26021

Second Place: Team Procrustination

Feng Guo @ BGSU
Nick Howald @ BGSU
Marie Childers @ BGSU
Jordan Dovel @ BGSU
Sami Nesnidol @ BGSU
Andrew Samo @ BGSU
Sam T. McAbee @ BGSU
Private Test Set Average Correlation =.24784

Third Place: Logistic Aggression

Ross Walker @ Michigain State University
Jacob Bradburn @ Michigain State University
Jeff Olenick @ Michigain State University
Private Test Set Average Correlation =.23252

Fourth Place: PI-RATES

Wes Barlow @ USAA
Fred Shumate @ USAA
Fabian Castro @ USAA
Frank DeVilbis @ USAA
Private Test Set Average Correlation =.2293

Organizers

Isaac Thompson @ Modern Hire
Nick Koenig @ Modern Hire
Mengqiao (MQ) Liu @ Amazon

How to Cite Data

Thompson, I., Koenig, N., & Lui, M. The 2019 SIOP Machine Learning Competition. Presented at the 34th annual Society for Industrial and Organizational Psychology conference in Austin, TX.