{{ summaryTitle }}
{{ summaryValue }}
{{ summaryLine }}
{{ badge.label }}

Before you begin

{{ assessmentDefinition.instrument.name }} is the official public-domain 80-item English full form. Keep the same robot and observation context fixed for every rating.

Privacy: {{ assessmentMemoryDisclosure }} {{ assessmentDefinition.privacy.disclosure }}
This changes only the presentation order of the answer ledger; it never changes scoring.
Score contribution is optional scoring detail for informed review, not a separate result or recommendation.
{{ progressPercent }}%
{{ progressLabel }}
{{ currentQuestion.text }}

Rate your impression of this robot in the same observation context. Choose the agreement response that fits best.

Safety and support
  • {{ action.label }}

This shared result is read-only. Start a new session to record different answers.

What this means

{{ interpretationArtifact ? interpretationArtifact.lead : 'No interpretation is available for this result.' }}

What to do next

{{ Array.isArray(interpretationArtifact && interpretationArtifact.points) && interpretationArtifact.points.length ? interpretationArtifact.points[0] : 'Review this result in context and use the stated limits before acting.' }}

  • {{ point }}

{{ interpretationArtifact ? interpretationArtifact.uncertainty : 'This result is informational and should be interpreted within the assessment scope.' }}

Understand your scoreSee the canonical score in its declared range

{{ scorePositionArtifact.summary }}

Loading score-position chart…

{{ assessmentChartStatus }}
Review your answersCheck the responses behind this result
ItemPromptAnswerCopy
{{ row.label }}{{ row.prompt }}{{ row.answer }}

No answer rows are available for this result.

Save or shareOptional handoffs for this completed result
Share result

The link contains your responses in its URL. It is portable, not encrypted; share it only with someone you trust.

This link includes: {{ assessmentSharedFieldLabels.join(', ') }}.

Include this completed result in a copied link?

{{ assessmentDefinition.privacy.disclosure }}

{{ shareResultStatus }}
Score chart
Answer review
IPIP Potential Emotional Intelligence Components Complete 68 IPIP items and compare seven emotional-component totals and averages without inventing an overall EQ score, cutoff, or ability rank.
Schutte Self-Report Emotional Intelligence Test (SSEIT) Score all 33 SSEIT responses, verify the three reverse-keyed items, and read the 33 to 165 total without invented cutoff bands.
IPIP Personality Inventory (IPIP-NEO-60) Assessment Build a 60-item Big Five profile with reverse-keyed domain means and two-item facet clues, while keeping non-normative limits clear.
Mini-IPIP Big Five Assessment Explore a 20-item Mini-IPIP Big Five profile with reverse-keyed trait means, response consistency cues, and clear non-normative limits.

A robot can appear socially aware even when the observer cannot inspect how it reached a response. Timing, motion, speech, gaze, and the surrounding situation all influence that impression. Perceived social intelligence measures the observer's judgment of those visible behaviors; it does not test the robot's internal reasoning or prove that the behavior will repeat safely.

The 80-item Perceived Social Intelligence (PSI) full form separates that impression into twenty four-item scales. Thirteen scales concern social information processing: a three-by-three set of recognizing, adapting to, and predicting human emotions, behaviors, and cognitions; three forms of identifying people; and overall social competence.

Observed capacityEmotionsBehaviorsCognitions
RecognizesDetects feelings or preferencesNotices actions or movementInfers thoughts or beliefs
Adapts toChanges behavior for feelingsChanges behavior for actionsChanges behavior for thoughts or beliefs
PredictsAnticipates feelingsAnticipates actions or intentionsAnticipates thoughts or beliefs

The other seven scales describe social presentation: Friendly, Helpful, Caring, Trustworthy, Rude, Conceited, and Hostile. A robot may look friendly in a scripted demonstration yet fail to recognize a person's intentions, or may appear capable in one interaction while behaving differently in noise, with a new user, or under another task. Keeping these impressions separate is more informative than treating social intelligence as one vague quality.

Perceived behavior
What the observer inferred from one defined interaction, recording, scenario, or period of prior experience.
Actual capability
Performance established through independent technical testing across relevant people, tasks, environments, and failure conditions.

PSI profiles support human-robot interaction research, structured observation, and design reflection. They are not norms, safety ratings, compliance checks, purchase recommendations, or certifications of intelligence. Comparisons are meaningful only when the rated robot, observer instructions, exposure, and context are sufficiently consistent.

How to Use This Tool:

Define one robot and one observation context before answering, then keep both fixed for the full form.

  1. Select Start assessment after watching or interacting with the robot long enough to form an impression, then rate all 80 statements on the five-point agreement scale. Judge what appeared to happen in that context; do not combine different robots or unrelated scenarios.
  2. Read the twenty scale means before the three composites. The means show the detailed perceived profile; the composites summarize broader groups and do not add new observations.
  3. Use Review your answers to inspect keyed contributions when a scale is surprising. Sorting by contribution changes the ledger presentation only.

Interpreting Results:

Each scale mean runs from 1 to 5. A larger value means the observer endorsed more of the named perception, including on Rude, Conceited, and Hostile. Those three scale names retain their ordinary direction in the detailed profile.

The composite totals have a different purpose. Social Information Processing adds thirteen scale means. Social Presentation combines four positive-presentation means with reversed versions of the three negative-presentation means. Social Intelligence adds those two composites. None has a published cutoff or population label.

  • Use differences between scales to identify behaviors worth examining in recordings, interviews, usability notes, or controlled tests.
  • Do not treat a high score as proof of technical competence, trustworthiness, safety, or readiness for deployment.
  • For repeat observations, keep the robot, task, exposure duration, environment, and rating instructions fixed enough to support a fair comparison.

Technical Details:

The official full form has four items per scale. Raw agreement responses use values 1 through 5. Positively keyed items keep that value, while reverse-keyed items contribute 6 minus the response. Each scale is the arithmetic mean of its four keyed item scores.

Formula Core:

The first equation constructs one scale mean. The remaining equations combine those means into the three published composites.

Mk= ∑i=14si4 ,si=ri or 6−ri SIP=∑k=113Mk SP=Friendly+Helpful+Caring+Trustworthy +(6−Rude)+(6−Conceited)+(6−Hostile) SI=SIP+SP

Here, r is a raw response, s is its keyed score, and M is a four-item scale mean. The negative presentation scales are reversed only when they enter the Social Presentation composite; their displayed scale means still indicate more perceived rudeness, conceit, or hostility as they rise.

ResultConstructionRange
Each PSI scaleMean of four keyed item scores1 to 5
Social Information ProcessingSum of 13 information-processing scale means13 to 65
Social PresentationFour positive scale means plus three reversed negative scale means7 to 35
Social IntelligenceSocial Information Processing plus Social Presentation20 to 100

All 80 responses are required. Scale means and composites are descriptive values without weighting, imputation, norm conversion, bands, or cutoff rules.

Limitations and Privacy:

Observer ratings can change with expectations, prior familiarity, novelty, task difficulty, camera angle, interaction length, and environmental conditions. A profile from one context cannot establish performance in another, and agreement among observers does not replace independent capability or safety testing.

Responses are held in browser memory. Choosing to copy a result link places all 80 raw ratings in the URL fragment after a confirmation step. The fragment is encoded but not encrypted or access-controlled, so anyone with the link can recover the response pattern. Starting over clears the current responses and removes the result fragment.

References: