Development and validation of the Combined Emotional and Affective Temperament Scale (CEATS): Towards a brief self-rated instrument

Abstract

Background

Temperament relates to both emotional dimensions and prevailing mood, but these different views are rarely integrated. Based on a model where temperament works as a system with activation, inhibition and control (inspired in Cloninger's and Rothbart's models), which produce the affective tone (inspired by Akiskal's and Kraepelin's model), we developed and validated the Combined Emotional and Affective Temperament Scale (CEATS).

Methods

1007 subjects (28% males) from the general population and university students filled in the instrument either in the Internet or in a pen and paper version. The CEATS has an emotional section (dimensional only), an Affective section (both dimensional and categorical) and an evaluation of problems and benefits related to temperament. The data was analyzed with standard psychometric batteries and different sections were compared.

Results

In the emotional section, 4 factors with Eingenvalue > 1 explained 46% of the variation. These factors were interpreted as drive, control, disinhibition-fear and anger, had a normal distribution and had satisfactory Chronbach's alphas (0.70–0.82). Anger was particularly associated with problems and drive with benefits. In the Affective section, all 10 categorical affective temperaments were selected, being euthymic and hyperthymic the most prevalent (18–23%), followed by cyclothymic and irritable (11–13%), anxious and depressive (8–9%) and dysphoric, disinhibited, labile and apathetic temperaments (3–7%). The dimensional evaluation of affective temperaments showed 95% of the sample was able to ascribe to at least one affective temperament. Only the euthymic and hyperthymic temperaments were clearly associated with a favorable problem/benefit profile. The comparison between the emotional and affective sections revealed that each affective temperament had a particular emotional configuration.

Limitations

Both computerized and pen and paper versions were used. The sample was not evaluated for psychiatric symptoms. Quantification of the dimensional assessment of affective temperament is limited.

Conclusions

The CEATS is a brief and adequate instrument to evaluate emotional and affective aspects of temperament simultaneously.

Introduction

Temperament relates to the emotional nature and the quality of the prevailing mood, being mostly inherited and relatively stable over time (Allport, 1961, Cloninger et al., 1993). Since the four humours of Hippocrates and Galen, the concept of temperament has had new interpretations by Eysenck (1987), Gray (Pickering and Gray, 1999), Cloninger (Cloninger et al., 1993), Akiskal (Akiskal et al., 1989) and others. Two of the most intensively studied temperament constructs in psychiatry are the psychobiological model by Cloninger, with a focus on behaviors and basic emotions, and the model of affective temperaments by Akiskal, based on Kraepelin's fundamental states.

For Cloninger, temperament dimensions are named as behaviors routed on basic emotions (in brackets): novelty seeking (anger), harm avoidance (fear), reward dependence (attachment) and persistence (ambition) (Cloninger et al., 1993). These are normally distributed dimensions that can occur in any combination, as they are independently inherited. The Temperament and Character Inventory (TCI) is the self-report instrument used to measure this construct, which was conceived more specifically for personality and its disorders. However, several studies have shown that in virtually all groups of patients with psychiatric disorders at least one of these temperament dimensions is altered compared to healthy controls (see Lara et al., 2006 for review). Most commonly novelty seeking (NS) and/or harm avoidance (HA) are either high or low, and self-directedness from the character section is often reduced in psychiatric disorders.

Akiskal has conceived temperament as the affective predisposition or reactivity, based on the original descriptions by Kraepelin (1921) of fundamental states (manic or hyperthymic, irritable, cyclothymic, anxious and depressive (Akiskal et al., 1989, Akiskal, 1998). These five affective temperaments would be the predisposing ground for the development of mood disorders. The TEMPS is the self-report scale to assess this construct (Akiskal et al., 2005a, Akiskal et al., 2005b, Akiskal et al., 2005c). These affective predispositions are often present in individuals who develop mood disorders, as well as in their relatives, with different distributions according to the type of mood disorder (e.g. more hyperthymic traits in bipolar I disorder, cyclothymic traits in bipolar II disorder and depressive traits in unipolar depression) (Evans et al., 2005, Kesebir et al., 2005, Akiskal et al., 2005a, Akiskal et al., 2005b, Akiskal et al., 2005c).

As Rothbart et al. (2000), our view coincides with Allport's (1961) definition of temperament as ‘an individual's emotional nature, including his susceptibility to emotional stimulation, his customary strength and speed of response, and the quality of prevailing mood, these phenomena being regarded as dependent upon constitutional make-up'. However, current temperament models assess either the emotional nature (e.g. TCI) or the prevailing mood (e.g. TEMPS). Rothbart et al. (2000) thoughtfully incorporated and studied the role of attention and self-regulation to the concept of temperament, but have not directly addressed the prevailing mood or affective predisposition.

Three studies (Maremmani et al., 2005, Akiskal et al., 2005a, Akiskal et al., 2005b, Akiskal et al., 2005c, Rózsa et al., 2008) have been conducted with simultaneous assessment with the TCI and the TEMPS, showing essentially that: hyperthymic temperament is associated with high NS and low HA; irritable with high NS and moderate HA; cyclothymic with both high; anxious with moderate NS and high HA and finally that depressive temperament is associated with low NS and high HA. Reward dependence and persistence are weakly correlated with these five affective temperaments.

Recently we have proposed an integration of emotional and affective temperament constructs, named as the fear and anger model, with clinical (Lara et al., 2006), neurobiological and treatment implications (Lara and Akiskal, 2006). This model is based on the principle that activation (related to anger and drive/pleasure) and inhibition (related fear and caution) are the two main emotional forces or ‘vectors of the mind’, as coined by Thurstone (1934) and inspired by Gray (Pickering and Gray, 1999). Their interaction would produce a resulting affective trend or prevailing mood. For practical reasons, in our model the dimensions of activation and inhibition can be conceived as high, moderate and low, as they are normally distributed in the population. Therefore, nine combinations can arise from their 2 × 2 interactions. In order to complete the other putative combinations of activation and inhibition, we have proposed four new affective temperaments: disinhibited (originally called hyperactive) with moderate NS and low HA; labile with both low; apathetic or passive with low NS and moderate HA, and euthymic with both moderate (Lara et al., 2006). We also added the dysphoric temperament, with high activation and inhibition, similarly to cyclothymic temperament. Based on other models and our preliminary versions of the scale, we included a factor that is related to self-regulation of activation and inhibition, which we called control. This concept is related to frontal lobe functions of attention, emotion self-regulation and behavioral adaptation and is similar to consciousness in the big five model (McCrae and Costa, 1985), some aspects of persistence and self-directedness in Cloninger's model (Cloninger et al., 1993) and executive functions and effortful control in Rothbart's concept (Rothbart et al., 2000).

Here we present the validation data of a new scale developed to evaluate emotional and affective temperaments simultaneously, trying to capture the original concepts and the complementary advantages of the instruments by Cloninger and Akiskal. The necessary adaptations and changes were conducted to allow for their integration in a short format. Named as the Combined Emotional and Affective Temperament Scale (CEATS), the emotional section consists of twenty-seven 5-item multiple choice questions on disinhibition-fear, drive, control and anger, whereas the affective section includes 10 descriptions of affective temperaments with a dimensional 5-point scale and a categorical choice of the best description. We also included two questions that subjects could rate in a 4-point scale according to the degree of problems and benefits related to their temperament.