Background: Our aim was to study the psychometrics and factor structure replicability of the Temperament Evaluation of Memphis, Pisa, Paris and San Diego Autoquestionnaire (TEMPS-A) in its Italian (Rome) Version. The questionnaire is a self-report 110-item measure that postulates five affective temperaments-the depressive, cyclothymic, irritable, hyperthymic, and anxious-which embody both strengths and liabilities along affective reactivity. In Italian, the TEMPS has previously been validated in its original 32-item version, the TEMPS-I (Pisa), one which did not yet include an anxious subscale.
Methods: The present sample consisted of 948 nonclinical subjects (27.39 years+/-8.22 S.D.). There were 476 men (50.2%: 28.56 years+/-8.63 S.D.) and 472 women (49.8%: 26.21 years+/-7.61 S.D.). Reliability and validity were assessed by standard psychometric tests.
Results: Principal Components Analysis with Varimax rotation resulted in a 3-factor solution: the first with highest explained variance (8.84%) represents Dysthymic, Cyclothymic and Anxious (Dys-Cyc-Anx) temperaments combined; the second identifies Irritable temperament (5.65% of variance); and the third Hyperthymic temperament (5.16% of variance). Cronbach Alpha coefficients for the three subscales were respectively .89, .77 and .74. The rates for the Dys-Cyc-Anx were 2.7%, and for the Irritable 3.1%. Despite the low rate of the Hyperthymic temperament (.2%), nonetheless 16% were between 1st and 2nd SD. Exploratory factor analysis revealed a positive loading combining Dys-Cyc-Anx with the Irritable; the Hyperthymic loaded negatively on this factor. In terms of dominant temperaments, based on z-scores, 2.7% were dysthymic, 1.7% cyclothymic, .7% hyperthymic, 3.5% irritable and 3% anxious.
Limitation: Although developed for self-rated use, the Italian authors nonetheless administered the TEMPS-A in an interview format. It is uncertain in what ways this procedure could have influenced our results, if any. Another limitation is that we did not assess test-retest reliability.
Conclusions: These data identify at least 3-factors, Dys-Cyc-Anx and Irritable (which are correlated), and Hyperthymic, which is uncorrelated with the others. Though our data are reminiscent of the neuroticism-extraversion distinction, importantly traits are operationalized in affective terms. Beyond the well-known relationship between the Dysthymic and Cyclothymic subscales and that between the Dysthymic and Anxious, the present data reveal a strong relationship between the Cyclothymic and Anxious as well, which is of great relevance for bipolar II. It is also provocative that much of hyperthymia (16%) in the +SD is between the 1st and 2nd SD, thereby "normalizing" this temperament in Italy, as previously reported by TEMPS-I (Pisa) from Northern Italy (and TEMPS-A from Lebanon and Argentina).