What “emotion skills” means

In this proposal, “emotion skills” means the set of abilities, habits, and modifiable traits that help someone work with emotions effectively: noticing emotions, putting them into words, tolerating them without avoidance, and using regulation strategies when appropriate. It also includes beliefs about emotions (e.g., whether they are useful and controllable), because beliefs shape whether people engage with emotions or shut them down. It also includes empathy (accurately recognizing others’ emotions and resonating appropriately), because it matters for real-world interpersonal functioning.

Importantly, this is not a claim about symptom severity, diagnosis, or whether a person is “more emotional.” It’s a construct / process description grounded in what shows up in language during therapy.

In practice, this includes emotion clarity and emotional granularity: moving beyond vague feeling states to more specific, differentiated, and integrated descriptions.

We use established questionnaires as our reference point for these constructs (see Perth Emotion & Psychopathology Lab resources).

What ELI is

ELI is a transcript-based system that looks for structured, verbatim evidence of emotion skills in what a client actually says during therapy (not just what they endorse on a questionnaire). Outputs should be auditable: each ELI claim is backed by specific excerpts.

ELI outputs are best treated as measurement / process markers. Clinical outcomes (symptoms and functioning) should be evaluated separately using validated outcome measures appropriate to the population.

What we want to learn

• Does ELI track self-report? Do ELI-derived emotion skill signals align with digital versions of PERS, PES, EBQ, and PERCI in the ways we’d expect?
• Is it stable? If someone’s context hasn’t meaningfully changed, do ELI signals look reasonably consistent across nearby sessions?
• Does it move with learning? If a skills-based intervention should improve emotion skills, does ELI move in the expected direction?
• Do humans agree with the evidence? Can trained coders review excerpts and generally agree with ELI’s tagging / interpretation?

What data we need

• Participants: therapy clients with multiple recorded sessions (so we can look at within-person patterns over time).
• Measures: PERS / PES / EBQ / PERCI administered digitally on a schedule aligned to sessions (baseline + follow-ups).
• Optional benchmark (performance-based): LEAS (Levels of Emotional Awareness Scale) as a complementary measure of emotional awareness (not self-report). See Lane & Smith (2021).
• Transcripts: speaker turns + timestamps, with clear consent, governance, and redaction workflows.

How we’ll check it

• Mapping: define how each battery’s construct maps onto transcript evidence (and what “counts” as evidence).
• Alignment tests: compare ELI signals to each battery (overall + subscales), and test expected patterns (not just a single correlation).
• Generalization checks: check whether results hold across sites, therapists, and populations without changing marker definitions.
• Human coding audit: define a coding guide + adjudication rules for a subset of excerpts, then compare to ELI outputs.

What ELI could be used for

• Personalized psychoeducation: highlight the specific emotion skills a client seems to be using well vs struggling with, and generate targeted practice.
• Treatment selection / monitoring: track skill change over time as a process marker.
• Research measurement: use transcript-based skill estimates as a complement to self-report batteries.

We’d scope any “diagnostic” language very carefully. The immediate goal is measurement and monitoring, not replacing clinical judgment.

Who we’re looking for

• US-based academic researchers who can run IRB-approved studies with therapy-session transcripts.
• Emotion skills / emotion regulation labs using PERS, PES, EBQ, and/or PERCI (or adjacent constructs).
• Graduate students / postdocs who want a clear, publishable validation project with a defined coding protocol.

We prefer US-based partners because HIPAA compliance and data export logistics are usually simpler, but we’re open to conversations and collaboration with researchers outside the US as well.

Ethics & privacy

This work involves therapy-session transcripts and sensitive text. Any publishable study should use clear participant consent and an appropriate IRB determination. The design should prioritize data minimization, clinician control, and the ability to trace outputs back to specific excerpts (so errors can be audited and corrected).

For governance details, see our Privacy & Compliance guide.