Research-informed. Still being tested.

What is established—and what is new

We reviewed NIOSH WellBQ, ergonomics guidance, ProQOL, COPSOQ III, the CFPB Financial Well-Being Scale, the Recovery Experience Questionnaire, WHO burnout definitions and HSE work-design guidance.

All 25 questions are newly authored for InterpretWell. We adapted the application of broad constructs to interpreting; we did not copy, translate or adapt individual instrument items. None of the original instruments' reliability estimates, norms or cutoffs apply to this assessment.

The emotional area includes exhaustion, distancing, emotional carryover and support access. It is not a burnout, PTSD, secondary traumatic stress or vicarious trauma scale. The five areas are practical groupings of mixed indicators, not proven psychological factors.

Exactly how scoring works

The recall period is four weeks. Responses range from 0 (Never) to 4 (Almost always). Every scored item points in the same direction: more frequent reported pressure. There are no reversed items, hidden weights or demographic adjustments.

An area summary is the arithmetic mean of valid answers, with at least four of five items answered. Not applicable, skipped and invalid answers are excluded, never treated as zero. Partial coverage is shown. Individual answers of 3 or 4 are highlighted even when the area cannot be summarized.

Bands use the unrounded mean: below 1 = Infrequent; 1 to below 2 = Occasional; 2 to below 3 = Recurring; 3 to 4 = Frequent. Display values use two decimals. These boundaries are provisional author choices tied to response anchors, not empirically established risk thresholds.

There is no total sustainability score, risk probability, percentile, diagnostic label or automated fitness-to-work decision. Areas are not ranked as statistically different. Results from older InterpretWell versions cannot be compared numerically with v3.

How we can earn stronger reliability claims

The next stage is expert review and cognitive interviews with a diverse group of interpreters, including Deaf interpreters, students, staff and freelance workers across settings and modalities. We need to test what each question means to people, what they recall, and whether answer choices fit.

A separate, explicitly consented pilot should examine missing answers, burden, floor and ceiling effects, short-term stability, domain structure and relationships with suitable established measures. Reliability must be evaluated for defensible subscales, not assumed from a single alpha for all 25 mixed indicators. Sample size and analyses should be planned with a measurement specialist.

Independent replication, accessibility review and review of fairness across groups are needed before stronger claims or empirical thresholds. Product feedback is not a validation study; self-selected app users are not a representative interpreter sample. No pilot findings are claimed here.