Seven steps from concept to instrument

1. Literature review. We started from current research on emotional intelligence as an ability, within the Salovey, Mayer and Caruso model, which is the best empirically grounded approach in this field.
2. A bank of 69 scenarios. Each scenario describes a situation in which the respondent decides how best to read or handle an emotion. Every item was reviewed internally against three criteria: a wide range of emotions (not only sadness and anger, but also gratitude, pride, contempt, frustration), a range of motives (power, care, achievement), and a 50/50 balance between working life and private life.
3. Content validation by competent judges. Psychologists from outside the team rated how well each item captures emotional intelligence. On the basis of their agreement we dropped the 13 weakest items.
4. The expert key. Thirty-three people with established emotional intelligence competence, mostly psychologists, EI trainers and diagnosticians, answered the items themselves. Their responses form the key: points are awarded for agreement with the answers given by people whose ability has been verified.
5. Validation study. More than 500 people completed a 56-item version of the test together with three other instruments: a reasoning test, a trait emotional intelligence questionnaire and a personality inventory. The point was to check whether EQ-TIZE measures what it should, and whether it confuses emotional ability with general intelligence or with personality traits.
6. Factor analyses. Exploratory and confirmatory factor analysis (EFA and CFA) confirmed that the data follow the theoretical model: perceiving, understanding, using and managing emotions. The 44 items with the best psychometric properties were kept; the remaining 12 did not reach satisfactory reliability and were dropped.
7. External validation. We examined how EQ-TIZE relates to other measures. The results support its ability-based character: a low correlation with a reasoning test, a moderate correlation with a self-report EI questionnaire, no meaningful relationship with extraversion or emotional stability, and low to moderate correlations with agreeableness, openness and conscientiousness.

Reliability

Reliability of the total score is omega = .94, which places the instrument among tests meeting the most demanding psychometric criteria, including those applied to individual diagnosis (the customary threshold is omega of at least .90; Nunnally and Bernstein, 1994). Scale reliabilities range from good to very good and exceed the threshold of .70 adopted for scales used in research and group comparisons (Cicchetti, 1994; DeVellis, 2017).

Scale	McDonald's omega	Interpretation
EQ-TIZE total score	.94	Very high, suitable for individual diagnosis
Perceiving emotions	.76	Good
Understanding emotions	.71	Good
Using emotions	.80	Good
Managing emotions	.87	Very good

Differences between Cronbach's alpha and McDonald's omega are small (.001 to .018), which indicates that the assumption of tau-equivalence within the scales is empirically justified and that the reliability estimates are stable regardless of the measurement model adopted.

Relationship with personality

In the validation study, EQ-TIZE scores were correlated with the five scales of the IPIP-BFM-50 inventory. The pattern follows theoretical expectations for an ability test of emotional intelligence and replicates findings known from the Polish literature.

This gives two pieces of validation evidence. Convergent: the relationships with agreeableness and openness replicate the pattern known for ability tests of emotional intelligence and confirm that EQ-TIZE measures a construct related to theoretically adjacent traits without reducing to them. Discriminant: the absence of relationships with extraversion and neuroticism separates EQ-TIZE from trait EI questionnaires, in which neuroticism typically absorbs a large share of the variance. The pattern holds at every level of aggregation, both for the four scales and for the two-factor models (experiential and strategic emotional intelligence).

Validation sample (N = 516)

Characteristic	Composition of the sample
Age	Mean 36.2 years (SD = 9.6, median 35, range 16 to 83). Most respondents were aged 26 to 45 (72.5 per cent); 3.7 per cent were 56 or older.
Gender	409 women (79.3 per cent), 103 men (20.0 per cent), 4 respondents indicated another answer.
Education	85.7 per cent held a higher education degree, 9.3 per cent secondary education, 4.3 per cent were students, 0.8 per cent primary or vocational.
Place of residence	60.9 per cent in cities of more than 200,000 inhabitants, 14 per cent in medium-sized towns, 25.2 per cent in small towns and villages.
Occupational status	81 per cent were in work. The largest sectors were HR and recruitment (21.5 per cent) and IT (20.8 per cent).

The sample is therefore made up mainly of working adults rather than students, which is unusual for studies of this kind, but it is dominated by women, by people with higher education and by residents of large cities. The norms derived from it are weighted to the structure of the Polish adult population by sex, age and education, using Statistics Poland data. The NORM26 study, running on a quota sample of about 4,000 adults until March 2027, is designed to remove this limitation.

What this means for a research team

The English version contains all 44 scenarios and the instructions, and the test runs online. Response times are recorded because they are needed to calculate the Emotional Calibration Index. Reliability and validity data come from the Polish version; there are no English-language norms yet, which is exactly why studies with English-speaking samples are of interest to us. Terms of use, including which data are shared, are agreed individually for each project.

Main English page: EQ-TIZE in English · Polish version of this page: Metodologia EQ-TIZE