In February of 2019, The Predictive Index conducted a survey of executives, managers, and individual contributors across diverse industries. We secured respondents through market research firm Critical Mix (now part of Dynata).
PI started with a pool of 94 Likert scale items, asked on a 5-point strongly disagree to strongly agree scale. All questions were required, and a not applicable option was provided as an “out.” In an attempt to avoid survey fatigue, two versions of the survey were created—each contained about 65 items. Responses were collected over the course of two weeks in February 2019.
The following psychometric analyses were run:
- Basic score calculation. PI assessed the mean, % favorable, % neutral, and % unfavorable of each item. Items that scored extremely high, or that had a large amount of neutral scores, were flagged as potential removals.
- Factor analysis. PI assessed the way the items naturally wanted to group together so we could ensure the way we presented the items in categories accurately captured the experiences survey takers were considering when thinking about their day-to-day work.
- Reliability analysis. Confirming the factor analysis, PI ensured that the items grouped together in each category were related but not overlapping. The objective was to cover as many aspects of each category as possible without redundancy.
- Correlations. Which items correlate most highly with engagement? As one of the intents of asking items beyond engagement was to provide items upon which leaders and managers can take action to influence engagement, PI ensured that all survey items had a linear relationship with engagement.
Following those analyses, the final survey was created, consisting of 10 Job items, 10 Manager items, 10 Team items, 16 Organization items, and 4 Engagement items.
Through those means, we collected 3,010 completed survey responses. Here’s some information about the respondent pool: