The best predictor of workplace performance and potential
Every item in Wave® earned its place because it is a proven predictor of performance at work. Nothing else got in.
Why would you accept anything less than the best?
What psychometric assessment actually means here
The measurement of behavior, ability & motivation
And the evidence that those measures predict something worth predicting. Most questionnaires are built first, then checked afterwards to see whether they predict anything. Wave® put that evidence at the heart of how it was built.
We had people complete the initial questionnaire version and asked their managers, colleagues and others to rate independently how well they performed at work. Then we tested 214 candidate measures of work behavior against those ratings and kept the 108 that predicted performance best. The rest were cut.
That decision, confirmed in a second, independent sample, is why the numbers further down this page hold up.
27/30
Independent BPS quality rating across six areas
0.57
Criterion validity, where 0.6 is rarely exceeded
1 in 50
Serious mis-hire risk, down from 1 in 5 to around 1 in 50
108
Work behaviors measured, not five broad traits
How Wave® was built
Two approaches dominate questionnaire design. One groups items by how they cluster statistically. The other measures what the author considers important.
Both leave validity until the end. We added a third, and it changed what came out.
Build the criterion
Alongside the questionnaire, we built a model of work performance itself: what people are actually rated on, drawn from organizational competency frameworks, published models and performance criteria across occupations.
That model of performance, rather than personality theory alone, is what Wave® was aimed at.
Write more than you need, then let the ratings choose
A team of experienced psychologists, including our founder Professor Peter Saville and our current Chief Science Officer Rab MacIver, spent the equivalent of more than two full-time years writing and reviewing items against strict rules: short, direct, behavioral, positively phrased, free of idiom and of content that favors one group.
Items were then selected first and foremost on how well they predicted independently rated performance, rather than on how neatly they correlated with each other.
Cross-validate, then keep validating
One study isn’t enough.
Item choices from the development trial were re-tested on a separate standardization sample of 1,153 participants, and the validation has been extended ever since across countries, languages, sectors and role levels.
"Questionnaire developers should prioritise validity, using every available method to create more powerful assessments."
Rab MacIver
Chief Science Officer, and one of the creators of Wave®
One model, aligned from prediction to performance
Wave® measures behavior at four levels of detail, from four broad clusters down to 108 specific facets. The depth matters because broad traits blur useful distinctions. Two people can score identically on a five-factor measure and behave very differently at work.
Each level has a matching equivalent on the performance side. The Inventive dimension was built from the items that best predicted independent ratings of Generating Ideas. Every predictor is tied to the criterion it was selected to forecast, so the link to performance is built into the report, not left to interpretation.
Four clusters
Solving Problems, Influencing People, Adapting Approaches, Delivering Results.
The broad shape of how someone works
12 sections
More detailed than clusters, while still giving an overview of someone’s strengths and challenges.
36 dimensions
Specific enough to be actionable, broad enough to be reliable. The level we use to map Wave® to client competency frameworks.
108 facets
The detail beneath each dimension, used in feedback to explore what underpins someone’s styles.
Why we ask twice about every behavior
Motive and talent, measured separately
Every one of the 108 facets is measured twice: once for what someone is driven to do, once for what they are good at. “I enjoy generating ideas” and “I produce lots of ideas” are different questions with different answers.
Where the two differ, it can be one of the most useful insights in the profile. It shows where someone may be working against the grain, and where their motivation isn’t yet matched by their talent.
In a development conversation it helps separate what someone cannot do yet from what they can do but may not want to.
Rating and ranking in one pass
Rating scales tell you how positive someone is about themselves overall. Ranking tells you what they choose when they cannot claim everything. Each has known weaknesses on its own, and most questionnaires pick one.
Wave®’s own rating-and-ranking format can capture both from a single response: candidates rate statements, and the system derives the ranking, only asking the candidate to rank genuine ties.
Combining them raises validity, sharpens the distinction between behaviors and makes distortion harder to sustain and easier to spot.
What independent reviewers found
The British Psychological Society reviewed Wave® and rated it 27 out of 30 across six technical areas. The review is available through the BPS website, which matters more than the score. A rating you can look up yourself is worth more than one we describe.
Wave® Professional Styles produces in a single test the information that might be gained from the administration of several other instruments and will be a valuable tool in the kitbag of trained professionals.
Fairness is designed in, then monitored
Written out at item level
Your most exposed step:
Measured and published
Group differences by age, gender and ethnicity are reported in the technical documentation across different samples.
Your most exposed step:
One consistent norm group, one consistent method
Generally small or negligible differences do not justify separate norms by age, gender or ethnicity, and we do not recommend them. Consistency is what makes a decision defensible later.
Your most exposed step:
Transparent by construction
Items are deliberately direct, so candidates see what is being asked and are not surprised by their results. Opaque items need more of them to measure the same thing, and they cost candidate trust.
Written out at item level
Your most exposed step:
Measured and published
Group differences by age, gender and ethnicity are reported in the technical documentation across different samples.
Your most exposed step:
One consistent norm group, one consistent method
Generally small or negligible differences do not justify separate norms by age, gender or ethnicity, and we do not recommend them. Consistency is what makes a decision defensible later.
Your most exposed step:
Transparent by construction
Items are deliberately direct, so candidates see what is being asked and are not surprised by their results. Opaque items need more of them to measure the same thing, and they cost candidate trust.
Where this science shows up
In the assessments
The behavioral range runs from a short sift to a full in-depth read. Depth changes; the model underneath does not.
On the dashboards
Wave Connect turns the model into a Success Profile for the role, then shows how each person measures against it, with the link back to the behavior every measure was chosen to predict.
In the training
The same model is what accredited users are trained on, so interpretation is more consistent across your team and ours.
Go deeper
Validity and Research
Coefficients, reliability, fairness data, norm groups and the comparative research program in full.
Research and resources
Whitepapers, webinars and podcasts from the people doing the research.
Security, data and compliance
Certifications, data handling and the documentation your security team, procurement lead or legal reviewer will ask us for.
The Wave® methodology is validation-centric questionnaire design. Instead of writing items, grouping them statistically and testing validity at the end, we modeled work performance alongside the questionnaire, wrote 214 candidate measures against it, and kept the 108 that best predicted independently rated performance. Item selection was driven by prediction, rather than internal structure.
Assessment validity is measured as a correlation between assessment scores and an external measure of performance, usually independent ratings by managers or colleagues. The higher the correlation, the better the assessment predicts performance: 0 means no better than chance, and a perfect 1 is never achieved in practice. Around 0.3 is considered good for a personality measure and it is extremely challenging to achieve higher than 0.6 validities for any selection method.
Our most in-depth behavioral assessment reaches a criterion validity of 0.57 against independently rated global work performance. In practical terms, moving from no assessment to that level of validity takes the risk of a serious mis-hire from roughly 1 in 5 to around 1 in 50.
Yes. The British Psychological Society reviewed Wave® and rated it 27 out of 30 across six technical areas. The review is available from the BPS, so you can read the assessment of our science.
Group differences by age, gender and ethnicity are generally small or negligible and are published in the technical documentation across multiple samples. Because the differences do not justify different treatment, we recommend one norm group and one consistent method for a given role, which is also what makes a decision defensible later.