Emotional engagement during literary reception: Do men and women differ?

Abstract

This contribution examines the emotional engagement of men and women when reading narrative texts, aiming to see under which textual conditions men and women turn out to be different from or similar to each other in what they think and feel during reception. As part of an experimental mixed-methods study, male and female readers are asked to read either experience-type texts (focusing on inner experiences of characters) or action-type texts (focusing on actions as part of a suspenseful plot) and report their engagement on questionnaire scales and in written protocols. Results show that men and women differ in their engagement when reading action texts and in their emotional affinity to plots. They are highly similar when reading the experience texts, however, and in their affinity to characters. This study underlines that the emotional responses of males and females during reading are highly dependent on (con)textual cues.

Keywords:

• Narrative engagement
• Reading engagement
• Sex/gender and reading
• Sex/gender and emotions
• Emotional response

Acknowledgments

This article is a reanalysis of data previously published in Odağ (2008). The previous publication is cited wherever the present paper draws on it.

The author would like to thank Prof. Dr Margrit Schreier for her invaluable comments on an earlier version of this paper.

Notes

¹In order to facilitate the flow of reading, the terms gender and sex will be used interchangeably throughout the entire paper, despite conceptual differences between them (elaborated in, for instance, Archer & Lloyd, 2002, or Smith, 2011).

²This English title was suggested by the author of this paper. The original German title is Die Frau an der Tankstelle.

³This English title was suggested by the author of this paper. The original German title is 365 Tage im Eis.

⁴Because previous research (e.g., Oatley, 1999; Oliver, 2000) has shown that women are able to become equally involved with both female and male characters, and that men are less able to become involved with characters of the opposite sex, all main protagonists of the selected texts in this study were male (see discussion for an empirical argument supporting this choice).

⁵The optimal sample size for this design amounted to 88 participants (i.e., 11 participants per condition) at a maximally tolerable α-error of 5%, a β-error of 20%, and an estimated medium effect size (f=0.30; see Cohen, 1988, for effect sizes; see Hussy & Jain, 2002, pp. 165 & 283, for the calculation of the optimal sample size).

⁶While not the focus of this paper, it should be mentioned that this design also comprised three covariates assessing reader characteristics: empathy, gender role orientation, and reality perception. A more detailed account of covariate effects can be found in Odağ (2011).

⁷The amount of flexibility of this procedure over timing and context on the one hand allowed less experimental control and put at risk the internal validity of the study. On the other hand, however, it was precisely this flexibility that increased the external validity of the present reading study—a quality criterion that was considered at least as important as the criterion of internal validity in the context of a study on reading and emotional engagement.

⁸In German: Fragebogen zum Leseerleben.

⁹It should be noted that the Reading Engagement Questionnaire (Appel et al., 2002) comprises many more scales than the ones used for the reanalysis in the current paper. An example study comparing the emotional reading engagement of men and women using all scales of the questionnaire can be found in Odağ (2008).

¹⁰The scale pleasure in reading was subsumed under plot-driven engagement because its items denote an engagement with the text and its storyline as a whole (rather than its characters).

¹¹The main effect of text category, the two-way interaction between text category and gender, and the three-way interaction between thematic focus, text category and gender were non-significant (see Odağ, 2008).

¹²For calculating and comparing the effect sizes of text and gender, a regression analysis was preferred over an ANOVA procedure because it represents the more conservative and careful technique: in a regression analysis the effect of a given predictor on a criterion variable is exempted from possible effects of other predictors, as well as their interactions, on the criterion variable (based on semi-partial correlations; see Cohen, Cohen, West, & Aiken, 2003).

¹³Women: M=4.04 (SD=0.74); Men: M=3.77 (SD=0.72).

¹⁴It would thus have been more appropriate in this study to operate with notions of femininity and masculinity, as these gender role orientations are more flexible than the crude categorisation of readers into male and female (see Odağ, 2011, for an empirical assessment of the relevance of these concepts for emotional response during reading), but current instruments measuring these concepts have been proven to be largely insufficient to capture any relationships between gender-role orientation and emotional reading response (Odağ, 2011).

¹⁵As participants were unsupervised in this study and given a three-week time window to read and report, the author cannot be sure about the length of such a delay in this study. At the same time, however, the personal comments of readers were of such detail and density, that most doubts about the immediate succession between reading and reporting can be erased—a larger time lapse would not have allowed such detail in reporting.