Introducing the trier univalence neutrality ambivalence (TUNA) database: A picture database differentiating complex attitudes

Abstract

Using validated stimulus material is crucial for ensuring research comparability and replicability. However, many databases rely solely on bidimensional valence ratings, ranging from negative to positive. While this material might be appropriate for certain studies, it does not reflect the complexity of attitudes and therefore might hamper the unambiguous interpretation of some study results. In fact, most databases cannot differentiate between neutral (i.e., neither positive nor negative) and ambivalent (i.e., simultaneously positive and negative) attitudes. Consequently, even presumably univalent (only positive or negative) stimuli cannot be clearly distinguished from ambivalent ones when selected via bipolar rating scales. In the present research, we introduce the Trier Univalence Neutrality Ambivalence (TUNA) database, a database containing 304,262 validation ratings from heterogeneous samples of 3,232 participants and at least 20 (M = 27.3, SD = 4.84) ratings per self-report scale per picture for a variety of attitude objects on split semantic differential scales. As these scales measure positive and negative evaluations independently, the TUNA database allows to distinguish univalence, neutrality, and ambivalence (i.e., potential ambivalence). TUNA also goes beyond previous databases by validating the stimulus materials on affective outcomes such as experiences of conflict (i.e., felt ambivalence), arousal, anger, disgust, and empathy. The TUNA database consists of 796 pictures and is compatible with other popular databases. It sets a focus on food pictures in various forms (e.g., raw vs. cooked, non-processed vs. highly processed), but includes pictures of other objects that are typically used in research to study univalent (e.g., flowers) and ambivalent (e.g., money, cars) attitudes for comparison. Furthermore, to facilitate the stimulus selection the TUNA database has an accompanying desktop app that allows easy stimulus selection via a multitude of filter options.

Fig 1. Depiction of the procedure of the validation studies.

Blue borders indicate the self-reported picture ratings. Within the randomized presentation the sequence of the attitude question block was fixed (i.e., valence was followed by positivity which was followed by negativity which in turn was followed by felt ambivalence). Items with gray backgrounds were only presented for food pictures.

Fig 2. Screenshot of the TUNA app.

In the sidebar on the left side are the filter options displayed. In this example, all pictures with an anger rating between 0 and 54.5 were selected. At the top of the page the tabs to change to the information side, the full data table, the self-reported variables tables, and the image characteristics are located.