IRI: Interpersonal Reactivity Index

What this assessment is

The Interpersonal Reactivity Index, or IRI, is a 28-item self-report measure that reports four separate empathy-related tendencies. Each subscale produces a score from 0 to 28. The original instrument has no standard overall total, diagnostic threshold, or universal low-to-high interpretation bands.

The IRI reports separate scores for perspective taking, fantasy, empathic concern, and personal distress. Each subscale covers a different response to other people's experiences.

The IRI is a research measure of how people describe their own empathy-related tendencies, not a test of their performance in real interactions.

Who it was designed for

Mark Davis developed the IRI for research on individual differences in empathy. The original work used university-student samples.

Later studies have used translated and adapted versions with adults, adolescents, clinical groups, and professional trainees, but those versions do not automatically share the same norms or factor structure.

This page covers the original 28-item, four-subscale IRI. Short forms need their own scoring and validation evidence.

How it is administered

The person rates how well each statement describes them using five response levels. Seven items contribute to each subscale.

Some responses are reverse-scored. Use the exact published key for the version being administered.

The standard scoring material does not provide a general method for replacing missing responses. Do not prorate a subscale unless the chosen protocol has validated that method.

What it measures

The four IRI subscales are:

• perspective taking: the tendency to consider another person's point of view
• fantasy: imaginative involvement with characters and fictional situations
• empathic concern: other-focused compassion and concern
• personal distress: self-focused discomfort in tense interpersonal situations

High personal distress does not mean high empathic concern. The two can move differently.

The TEQ uses one total focused mainly on emotional empathy. The IRI keeps several cognitive and emotional tendencies separate.

How scoring works

The standard IRI scoring sheet codes each response from 0 to 4 after reversing the keyed items.

Add the seven values within each subscale. Each subscale ranges from 0 to 28.

Do not add the four subscales into a standard overall total. The instrument was designed to preserve their distinct meanings, so one combined number would discard information that the IRI is meant to retain.

Some papers report subscale means instead of sums or shift the response values to 1 through 5. Confirm the coding before comparing numbers across studies.

How to interpret IRI scores

A higher score means the respondent endorsed more of the tendency named by that subscale.

No universal clinical bands exist. The original research compared subscale patterns and group distributions rather than defining thresholds for low, normal, or high empathy.

Interpret all four scores by name. A high fantasy score cannot substitute for perspective taking, and a high personal-distress score is not evidence of greater helping behavior.

Use norms from the same language, age group, and setting when a comparison is needed.

What the score cannot tell you

The IRI cannot show how accurately someone reads another person's thoughts or emotions. It measures self-perception.

Responses can be affected by insight, current mood, culture, and the wish to appear caring.

The IRI cannot diagnose autism, a personality disorder, psychopathy, or another condition. A low score on one subscale has many possible explanations.

It also cannot predict compassionate behavior in every situation. Beliefs about empathy and observed behavior are related but not identical.

Evidence and limitations

The 1983 validation paper found different and expected relationships for each of the four subscales, supporting a multidimensional model.

A systematic review of empathy measures found that the IRI was among the most frequently studied instruments. It also found limited standardization and continuing uncertainty across adaptations.

The four-part structure is a strength when the research question is specific. It also makes careless use more likely if the subscales are collapsed or relabeled.

Record the version, language, response coding, and reporting convention with every result.