Open Conference Systems, ITC 2016 Conference

Font Size: 
PAPER: Detecting Who Is Going to Innovate
Achim Preuss, Katharina Lochner, Richard Justenhoven

Building: Pinnacle
Room: Cordova-SalonD
Date: 2016-07-03 03:30 PM – 05:00 PM
Last modified: 2016-05-21

Abstract


Introduction

Innovation has become something like a “Holy Grail†in economics since innovative products and services are a competitive advantage in rapidly changing international markets (Maier, Streicher, Jonas, & Frey, 2007). Thus, companies strive to establish an environment that facilitates innovation (Amabile, Conti, Coon, Lazenby, & Herron, 1996) and try to recruit innovators. It is desirable to measure applicants’ potential to innovate at an early stage in the recruitment process, i.e. in unsupervised online mode. Whether someone will be an innovator is determined by cognitive ability, certain personality characteristics, and creativity (Farr, Sin, & Tesluk, 2003). To date, personality and cognitive ability can be measured in unsupervised online mode, but for creativity there are no standardised instruments for unsupervised testing with automated reporting available.

Objectives

Creativity is measured by three components fluency, flexibility, and originality (Guilford, 1967). The aim was to develop an online creativity test that assesses these scores using an automated machine-learning-based scoring algorithm. The task is to draw pictures on a scratch board using given shapes and to name them.

Design/Methodology

In two subsequent studies the instrument and scoring was developed. In a third study with N = 470 participants it was validated.

Results

Test-retest reliabilities were .82 for Fluency, .67 for Flexibility, .71 for Originality, and .72 for Overall Creativity When scoring the pictures according to the Torrance Tests of Creative Thinking (TTCT; Torrance, 1974) manual, correlations between this and the automated scoring were .99 for Fluency, .85 for Flexibility, .88 for Originality, and .93 for Overall Creativity.

Conclusions

The machine-learning-based scoring algorithm for the online creativity test provides reliable and valid creativity scores. A limitation so far is that there is no proof of criterion-related validity in the sense that the test really predicts creative performance in workplace settings. A study is currently being planned.


An account with this site is required in order to view papers. Click here to create an account.