Is it okay to deliver the PI Behavioral Assessment on a mobile device?
The Predictive Index (PI) fully endorses the use of mobile devices for taking the PI Behavioral Assessment.
PI has compared the results of thousands of PI Behavioral Assessments taken on computers, tablets, and mobile phones. Two studies have verified that the choice of device has no effect on the number of adjectives a respondent chooses or the resulting factor scores.
Does the device matter?
In some assessments, the device used by the respondent can make a difference. For example, it would be difficult to write out an essay response on a mobile phone.
In 2014, PI selected a random sample of 3,000 results from our database: 1,000 results taken on a computer, 1,000 results take on a tablet, and 1,000 results taken on a mobile phone. We then compared the number of adjectives selected from each of the four behavioral factors: Factor A (Dominance), Factor B (Extraversion), Factor C (Patience), and Factor D (Formality).
Analysis of this sample showed no meaningful difference between the number of adjectives checked for respondents using different devices. There was a small effect (Cohen’s D = 0.27) between the number of Factor A adjectives checked by respondents on computers versus the respondents using mobile phones, but this small effect was not observed on the other factors. Given the within-person scoring of the assessment, this small difference was not expected to make any difference in the interpretation of the reported factor scores. No effects were observed for the other factors or for tablet users, and there was no significant difference between the total number of adjectives selected (Factor M) by device type.
A second study was conducted with 97,739 respondents. The sample was primarily computer users (only 4.5% used a mobile phone and 2.3% used a tablet). In this study, the researchers looked at whether there might be an interaction effect between the device type and the job level of the respondent. The hypothesis was that some job levels might be more comfortable interacting with certain devices. Once again, no meaningful effect was found in this sample, with Cohen’s f2 values falling below 0.02 for every factor and job level. These findings support the initial study’s findings that the device type does not impact the number of adjectives a respondent selects.
What does this mean in practice?
Respondents are free to use computers, tablets, or mobile phones to take the PI Behavioral Assessment online. The choice of device has no impact on the accuracy of their scores or the resulting behavioral patterns.