Methods Comparison for Collecting Subjective Valence and Arousal Judgements

Human emotion and affect are critical for regulating behavior, guiding decision-making and facilitating social communications. In everyday language, emotion and affect are interchangeable. However, in psychology, they are conceptualized as two distinct constructs. Emotion refers to discrete, prototypical subjective experiences elicited by specific stimuli. Affect, in contrast, refers to more elementary feelings that can be described along two continuous dimensions: valence (positive-negative) and arousal (low-high). According to the circumplex model (Russell, 1980), emotion can be represented within a two-dimensional core affective space defined by valence and arousal. Although valence and arousal are conceptually interrelated, most studies measure them separately using discrete Likert scales or pictorial tools. Some studies have adopted a valence-by-arousal grid to capture simultaneous measurements, yet systematic validation is needed to ensure that the grid’s axes accurately reflect the core affective dimensions underlying discrete emotions. This study compared the use of a pictorial 9x9 valence-by-arousal grid and multidimensional scaling methods to collect affective data. All 32 participants were presented with 50 mental imagery prompts and rated “how they feel about their imagery” (affective-focused) and “what they know about the prompt” (semantic-focused) using the grid. Participants also rated their experienced emotions based on 12 emotional words on a 5-point Likert scale. Data from participants showed a similarity in the two approaches. Multidimensional scaling offers flexibility to study multiple dimensions of affect, making it ideal for exploratory research, but demands more time and data. The valence-by-arousal grid allows for quicker data collection with similar results but limits to the fixed dimensions of valence and arousal.