Neurodiversity and soft skills assessment: measuring without discriminating

A candidate who avoids eye contact, answers with a slight delay, or delivers a highly structured response where spontaneity was expected: in a classic interview, these signals weigh negatively. Yet none of them measures a competency. For neurodivergent candidates (autism, ADHD, dyslexia, among others), traditional assessment processes regularly confuse conformity to social codes with professional capability. How do you assess soft skills without reproducing this confusion?

Where classic assessments penalise for the wrong reasons

The interview rewards conversational ease. The format rests on implicit conventions: making eye contact, filling silences, producing a fluid narrative about oneself. These conventions vary across neurological profiles without saying anything about rigor, perseverance or analytical capacity. The halo effect does the rest: an interaction judged "odd" contaminates the evaluation of every competency, a mechanism we dissect in our article on cognitive biases in hiring.

Declarative questionnaires presuppose a social reading of the questions. Answering a personality test "strategically" requires guessing what the assessor expects, an exercise in social decoding that has nothing to do with the competency being assessed. A respondent who interprets questions literally can produce an atypical profile without it reflecting their real capabilities.

The assessment center combines both. Group exercises, observed role plays, fast social dynamics: the format structurally favours extraverted, socially fluid profiles, whatever competency is actually targeted.

The paradox is cruel: organisations seeking analytical capacity, rigor and concentration reject candidates who are particularly endowed with them, because the assessment format measures something other than what it claims to measure.

The heart of the problem: what is actually being measured?

In psychometrics, this problem has a name: construct-irrelevant variance. Part of the score does not reflect the targeted competency but something else, here conformity to the social codes of the format. The more social irrelevant variance a format carries, the more it discriminates against profiles who lack that fluency, without gaining any precision on the competency itself.

The answer is not to lower the bar, nor to assess "leniently". It is to tighten the measure onto what it claims to measure. Three principles contribute:

• Explicit criteria rather than impressions. A competency defined by observable behaviours leaves less room for the assessor's "feeling".
• Written, asynchronous situations rather than live social performances. The candidate reasons about professional situations, without speed of repartee or body language entering the score.
• The same standardised exercise for everyone, documented. Standardisation protects all candidates from variations of mood and format, and atypical candidates more than others.

How the Rising Up approach fits this logic

The Soft Skills Scan rests on design choices that mechanically reduce construct-irrelevant social variance.

It assesses neither face, nor voice, nor body language. The assessment relies on written, contextualised professional scenarios. Eye contact, prosody or response latency simply do not exist in the measure.

It ties every score to a public framework. The 18 behavioural competencies of the Rising Up framework are defined by published observable behaviours. What the score measures is documented and auditable; you can verify that no criterion rewards social conformity as such.

It produces a decision aid, not a verdict. The result informs the exchange between recruiter and candidate, it does not replace it. For a neurodivergent candidate, that changes a lot: the discussion is about competencies measured in situation, not about the impression left in an interview.

This measurement requirement joins the European regulatory framework: the AI regulation imposes transparency and human control on candidate assessment systems, as we detail in our article on the AI Act and recruitment tests.

A matter of accuracy as much as fairness

Objective assessment of behavioural competencies is not a concession made to atypical candidates: it is an improvement of the measure for everyone. A process that confuses social ease with competence is also wrong about neurotypical candidates; it is simply wrong less visibly.

Behavioural skills predict professional success, contribute to it causally, and can be developed (research by James Heckman, Nobel laureate in economics). Measuring them precisely, across all profiles, is therefore both a performance issue and an equity issue. The market is moving in this direction: according to APEC, 41% of companies recruiting executives now formally assess soft skills, up 9 points in one year. What remains is to do it with tools that measure accurately.

Recognised as Deeptech by the European Innovation Council (EIC) in 2023, Rising Up designs its assessments on this principle: measure the competency exercised in situation, nothing else. To see what this changes in a hiring process, discover the recruitment test grounded in cognitive science.