A Critical Evaluation of the International Affective Picture System in the Study in the Study of Emotion

Mgr. et Mgr. Kristína Czekóová, Ph.D.

Doctoral thesis

Abstract:

The main focus of this thesis is on one of the most extensively used methods for eliciting emotions – International Affective Picture System (IAPS). IAPS is a large set of photographs that depict diverse semantic content including natural sceneries, food, animals, violence, objects of everyday use, accidents, pollution, erotic pictures, and so forth. The theory behind the IAPS method relies on a dimensional view that defines all emotions as a coincidence of values on three major dimensions: valence, arousal, and dominance. Each photograph within the system is defined by normative ratings for all three dimensions. The purpose of developing the IAPS was to facilitate comparisons between and replications of the results from studies on emotion from different laboratories. To this end, normative ratings (mean values of evaluations per picture) on all three dimensions have been acquired and many researchers have been selecting their stimuli from the IAPS database based on these norms. Recent studies, however, challenge the idea of a single set of norms for participants of different age, gender and from different culture background. Furthermore, it seems that Osgood’s proposed dimensions (i.e. pleasant – unpleasant, calm – excited and controlled – dominant) do not differentiate between the IAPS pictures sufficiently, and alternative perspectives – such as discrete categorisation of this material – have been suggested. The aim of the studies that comprise this thesis was to look into possibilities and limitations of IAPS as an experimental instrument and to offer alternative approaches to the analysis of data acquired by its application. A General Introduction reviews existing studies employing the IAPS, its theoretical background, theoretical and methodological limitations of its application, and suggestions for future use. Three data-driven chapters (Chapters 3, 4, and 5) discuss methodological and statistical considerations of using the IAPS, and offer alternative approaches to the analysis of data collected through its application. Although various perspectives are considered, a semantic approach is a key feature throughout the thesis. Chapter 3 looks into common statistical and methodological procedures routinely employed in emotion research using the IAPS. It is demonstrated that typical approaches to data analysis do not take into consideration the nature of the data or individual differences among participants. In addition, general aspects of emotionality such as relationship of valence and arousal, negativity bias and nature of affective space are addressed. Differences in these respects found through detailed analyses comparing the data comprising this thesis to those published elsewhere stress the need for employing distinct techniques such as meaning-focused analysis to obtain more precise results. Additionally, many significant differences with large effects provide reliable evidence for cultural differences. Gender differences were also confirmed. Furthermore, it is revealed that the sequence of presentation has a considerable effect on the evaluation of the slides. In Chapter 4, IAPS data acquired traditionally are analysed in a novel way – by employing hierarchical cluster analysis (HCA) in order to identify whether and how the pictures are clustered together. Results of HCA reveal spontaneous categorisation of the slides with respect to their content. Chapter 5 offers yet another way of investigating meaning within the IAPS method – text analysis of verbal report. Meaning dimensions that are shared and others that are specific for each picture suggest that two or three dimensions are not sufficient for capturing all the information about photographs within the IAPS set. In conclusion, the presented results illustrate substantial cultural and gender differences in the processing of the IAPS stimuli. In addition, evaluations from our participants demonstrated reliably that people tend to categorise the

Abstract:

IAPS pictures into meaningful semantic clusters spontaneously. Finally, verbal report on the meaning of the pictures shows that there are many different meaning variables to be taken into account in addition to valence and arousal. All these findings have significant implications for future research into the perception of emotion, and the use of IAPS in particular. Mean ratings per picture and/or the use of a few broad basic dimensions offer only limited information on emotion. Moreover, the sequence effect identified here presents a confound in studies using the IAPS that should be considered in future study designs. Evidence of photograph categorisation into small and meaningful groups was present even when applying a standard evaluation process that considers ratings on scales of valence and arousal only. These data demonstrate the feasibility of a semantic approach to the study of emotion.

Abstract:

The main focus of this thesis is on one of the most extensively used methods for eliciting emotions – International Affective Picture System (IAPS). IAPS is a large set of photographs that depict diverse semantic content including natural sceneries, food, animals, violence, objects of everyday use, accidents, pollution, erotic pictures, and so forth. The theory behind the IAPS method relies on a dimensional view that defines all emotions as a coincidence of values on three major dimensions: valence, arousal, and dominance. Each photograph within the system is defined by normative ratings for all three dimensions. The purpose of developing the IAPS was to facilitate comparisons between and replications of the results from studies on emotion from different laboratories. To this end, normative ratings (mean values of evaluations per picture) on all three dimensions have been acquired and many researchers have been selecting their stimuli from the IAPS database based on these norms. Recent studies, however, challenge the idea of a single set of norms for participants of different age, gender and from different culture background. Furthermore, it seems that Osgood’s proposed dimensions (i.e. pleasant – unpleasant, calm – excited and controlled – dominant) do not differentiate between the IAPS pictures sufficiently, and alternative perspectives – such as discrete categorisation of this material – have been suggested. The aim of the studies that comprise this thesis was to look into possibilities and limitations of IAPS as an experimental instrument and to offer alternative approaches to the analysis of data acquired by its application. A General Introduction reviews existing studies employing the IAPS, its theoretical background, theoretical and methodological limitations of its application, and suggestions for future use. Three data-driven chapters (Chapters 3, 4, and 5) discuss methodological and statistical considerations of using the IAPS, and offer alternative approaches to the analysis of data collected through its application. Although various perspectives are considered, a semantic approach is a key feature throughout the thesis. Chapter 3 looks into common statistical and methodological procedures routinely employed in emotion research using the IAPS. It is demonstrated that typical approaches to data analysis do not take into consideration the nature of the data or individual differences among participants. In addition, general aspects of emotionality such as relationship of valence and arousal, negativity bias and nature of affective space are addressed. Differences in these respects found through detailed analyses comparing the data comprising this thesis to those published elsewhere stress the need for employing distinct techniques such as meaning-focused analysis to obtain more precise results. Additionally, many significant differences with large effects provide reliable evidence for cultural differences. Gender differences were also confirmed. Furthermore, it is revealed that the sequence of presentation has a considerable effect on the evaluation of the slides. In Chapter 4, IAPS data acquired traditionally are analysed in a novel way – by employing hierarchical cluster analysis (HCA) in order to identify whether and how the pictures are clustered together. Results of HCA reveal spontaneous categorisation of the slides with respect to their content. Chapter 5 offers yet another way of investigating meaning within the IAPS method – text analysis of verbal report. Meaning dimensions that are shared and others that are specific for each picture suggest that two or three dimensions are not sufficient for capturing all the information about photographs within the IAPS set. In conclusion, the presented results illustrate substantial cultural and gender differences in the processing of the IAPS stimuli. In addition, evaluations from our participants demonstrated reliably that people tend to categorise the

Abstract:

IAPS pictures into meaningful semantic clusters spontaneously. Finally, verbal report on the meaning of the pictures shows that there are many different meaning variables to be taken into account in addition to valence and arousal. All these findings have significant implications for future research into the perception of emotion, and the use of IAPS in particular. Mean ratings per picture and/or the use of a few broad basic dimensions offer only limited information on emotion. Moreover, the sequence effect identified here presents a confound in studies using the IAPS that should be considered in future study designs. Evidence of photograph categorisation into small and meaningful groups was present even when applying a standard evaluation process that considers ratings on scales of valence and arousal only. These data demonstrate the feasibility of a semantic approach to the study of emotion.

Keywords

emotion, emotion perception, International Affective Picture System, Self-Assessment Manikin, semantic approach

Language used: English

Date on which the thesis was submitted / produced: 2. 12. 2011

Thesis defence

• Date of defence: 22. 12. 2011
• Supervisor: prof. PhDr. Tomáš Urbánek, Ph.D.
• Reader: prof. MUDr. Miloslav Kukleta, CSc., doc. PhDr. Marek Preiss, Ph.D.

Masaryk University

Faculty of Social Studies

Doctoral programme / field:
Psychology (4-years) / General Psychology