The assessment of affective reactivity using films: Validity, reliability and sex differences

Abstract

Researchers interested in individual differences of brain asymmetry and affective reactivity have recently used film clips to elicit emotions and provided some standardisation data of emotion ratings for English-speaking samples. The aim of the present study was to examine whether (1) a set of films which already proved to elicit different emotions in English-speaking samples would demonstrate a similar capacity to induce emotions in a German sample and (2) to explore the validity, reliability and sex differences of aggregated indices of affective reactivity. Based on two film sets which have been examined by other investigators in the field, we formed a set of 13 emotion-eliciting films and presented the films along with emotion rating scales to university students. Our findings indicate that most films had the capacity to elicit specific discrete target emotions and the emotions elicited by the negative clips were more distinct than the emotions due to the positive clips. The factor structure of the emotion rating scales indicated that a distinct positive and negative dimension was extracted. The indices of positive and negative affective reactivity demonstrated good reliability and internal consistency. Women reported on greater levels of affective reactivity than men, both for positive and negative affective reactivity.

Introduction

In recent years, some investigators have proposed that individual differences of baseline asymmetrical activation levels in anterior brain regions are associated with trait-like affective behaviours (e.g. Davidson and Tomarken, 1989). Some studies reported on a positive relation between relative right-sided anterior cortical activation and generalised negative affect, negative affective reactivity to negative film clips, or trait anxiety; others reported on relative right-sided anterior cortical activation in depression (for a review, see Davidson, 1995). However, some data is at variance with these findings (for a review, see Hagemann et al., in press).

In these studies, all affective traits were measured with published and widespread standard self-report instruments, with the important exception of affective reactivity. The latter was assessed by self-made procedures, either by presenting brief emotion eliciting film clips (Tomarken et al., 1990; Wheeler et al., 1993), or by presenting affective slides (Hagemann et al., in press). Immediately after each stimulus presentation, the S indicated on rating scales for different emotions the intensity of each felt emotion during the stimulus presentation. An index of positive affective reactivity (PAR) was computed as the mean across all positive emotion ratings across all positive stimuli and an index for negative affective reactivity (NAR) was formed in a respective manner. While Hagemann et al. (in press)used slides of the well standardised International Affective Picture System (Lang et al., 1988), both Tomarken et al. (1990)and Wheeler et al. (1993)used films which were culled especially for each study. For the latter two studies, the films were selected on the basis of initial emotional ratings provided by undergraduate students. Only those clips were used that elicited the highest ratings on each film's target emotion (e.g. happiness, disgust). Tomarken et al. (1990)provided a brief description of their standardisation data.

Compared to many other emotion eliciting procedures like interaction with trained confederates, stressful interviews, hypnosis, repeating phrases, facial muscle movements, imagery, music, sleep deprivation and drugs, the presentation of slides and films has the benefit of an easy standardisation and does not require deception or specific S populations (for a comparison of different procedures, see McHugo et al., 1982; Philippot, 1993; Gross and Levenson, 1995). The capacity of slides to elicit affective reactions has been successfully validated by the use of self-report affect measures (e.g. Lang et al., 1988; Wagner, 1990) and by the use of physiological measures like facial EMG, heart rate, skin conductance and startle response measures (for a review, see Bradley et al., 1993). However, some authors believe that slides can only induce weak positive or negative affective states without much further distinction (e.g. Philippot, 1993).

In contrast, there are many studies that used a variety of self-report measures and demonstrated that films have the capacity to induce distinct target emotions like anxiety (Pillard et al., 1974), anger (Brown et al., 1977), depressive mood (van Rooijen and Vlaander, 1984) or sadness (Engel et al., 1984; Marston et al., 1984). In one intriguing study, Philippot (1993)presented a selection of segments from commercial colour movies, targeting the emotions of anger, disgust, sadness, happiness and fear. Ss emotional states were assessed after each movie presentation with the Differential Emotions Scale (DES; Izard et al., 1974), the Semantic Differential (Osgood, 1964) and with free labelling. Results suggested that film segments can elicit a diversity of predictable and distinct emotions that are similar in a majority of individuals. These findings are further corroborated by evidence that films have also the capacity to elicit affective reaction on the physiological level as measured with facial EMG, heart rate, skin conductance and startle response measures (Mewborn and Rogers, 1979; Hubert, 1988; Hubert and de Jong-Meyer, 1990; Jansen and Frijda, 1994; Kumari et al., 1996). In addition, film clips have been found to elicit emotions of higher intensity than slides (Heiman, 1980; Julien and Over, 1988).

Recently, Gross and Levenson (1995)developed a film set for the induction of different target emotions and provided extensive standardisation data. Inter alia, they reported on more intense emotion reports by women than men (but see McHugo et al., 1982> and Philippot, 1993, who did not observe such between-sex differences). Again, this finding is supported by physiological data that indicates a greater affective reactivity in women than men. Kumari et al. (1996)investigated the startle response during the presentation of positive or negative films and reported on longer latencies in males than females for both positive and negative films. Hubert and de Jong-Meyer (1990)recorded facial EMG while Ss watched different films and reported that changes in facial muscle activity between films of different target emotions was more pronounced in females than males.

The present study is part of a larger investigation project on the frontal resting asymmetry and affective traits, including the affective reactivity to films. The aim of the present study was to form a set of film clips which proved to elicit different emotions in English-speaking samples and to investigate each film's capacity to induce the intended target emotion in a German sample. In addition, we examined the reliability and validity of the affective reactivity indices and explored sex differences of affective reactivity.