How Do We Trust?

Affect and Sweet Taste on Interpersonal Ratings of Trust in a Specific Other

The project that follows was completed fall semester 2013. The design of the project and research was conducted myself, with consultation from lead researcher and professor for the class, Dr. Linda Lockwood. The IRB for Metropolitan State University approved this project.

Why is trust important?When do we choose to trust?Trust Chart

In individuals, trust fosters a broadened amount of interpersonal relationships6 and allows people to organize into groups 17. In larger networks of individuals, such as corporations and nations, trust becomes a measure of success. Economic processes and attempted management are dependent upon trust3, and managers who appear trustworthy improve profits 21. National economic success has been correlated with measures of trust7,8.

Trust is defined by Rousseau et al. (1998) as a psychological state with two related facets:

1. the intention to accept vulnerability
2. positive expectations of the intentions or behavior of another

In addition, trust is positively influenced by feeling respected at a deep level 9 and mediated by personality and the type of relationships involved 14. Dunn and Schweitzer (2005) performed emotional manipulations on participants, and compared valence (positive or negative) of emotion on decisions to trust another coworker. Their findings suggested emotions with positive valence – such as happiness – increase trust while emotions that have negative valence – such as sadness – decrease trust.

Partitioning of Trust

+….Congruent sensory-motor states?

In my fourth semester at Metropolitan State University, I became interested in embodiment as a philosophy and the psychology behind it. Lawrence Barsalou, in 1999, had argued that cognition is embodied: sensory-motor states give rise to perceptual symbols, these become memory, function symbolically, and can be manipulated. We can think of this as the study of how we use our senses to think.

Presenting Psychological Research

Metaphors link a sensory-motor state to a larger idea. To test this empirically, Meiers and Robinson (2005) suggested that:

1. Congruent stimuli and metaphors should be encoded and recalled faster than incongruent stimuli and metaphors

2. When metaphors are presented, the sensory-motor state grounding it should be activated

3. The link should be automatic.

Many everyday metaphors refer to sensory-motor states and are described as grounded metaphors 2,10,11,13. In a now classic experiment, Eskine, Kacinik, and Prinz (2011) gave a bitter tasting beverage to one group of participants and water to another group. The group receiving bitter tasting beverages dealt harsher moral judgments to characters in vignettes read to them. It appears that the metaphor “moral wrong is disgusting” is encoded and conceptualized through a sensory state of disgust.

“That’s disgusting!”

Meiers, Moeller, Riemer-Peltz, & Robinson (2012) found that participants that eat a sweet candy were more likely to volunteer more time to stay and help with another experiment than participants eating a control food. This suggests that we recruit metaphors to aid us in conceptualizing how a person will treat us. In other words, whether we can expect positive intentions or behavior of another may partially depend upon on recruitment of sweet taste metaphors via congruent perceptual symbols.

Expecting positive intentions is a cornerstone of trust…

So, I had my experiment. I wanted to test if sweetness increased participants’ feelings of trust in others. I decided to use Dunn and Schweitzer’s methodology to test if an emotional induction had an effect on trust (to repeat and validate the study) and to then test my own hypothesis.



Fifteen students from an introductory psychology pool at Metropolitan State University participated in exchange for class credit. The average age of the participants was 24.07 (SD = 8.59). All participants completed the survey.

Participants were given an informed consent form, a typed six page booklet, and a debriefing form. The informed consent form assured confidentiality, anonymity, and the right to refuse to participant. The title of the study, as it appeared on the informed consent, was changed to lead participants to believe they were engaged in a study involving menial task repetition. The typed six page booklet included an emotional induction exercise, a trust questionnaire, and an experimental information questionnaire. The debriefing revealed the intent of deception to the participants.

A 2×2 between subjects factorial design was used to conduct this experiment. Participants were randomly assigned in groups to two conditions of emotional induction (happiness, sadness) and two conditions of type of beverage (sweet, not sweet), and scores on the trust measure were averaged to form the dependent variable. A factorial ANOVA was performed using SPSS to analyze the main effects between conditions and their interactions.

Participants signed up to take part in the study via the SONA system, which allowed the researcher to create available time slots for voluntary participation. Each time slot was randomly assigned to one of the four conditions.

Each room was set up before or during participants entering the room with two of the 3 oz beverages at each seat, either water or Minute Maid Berry Punch.

Participants were directed to:
-“drink the beverage in front of you like you would a shot in order to establish a menial routine”
-complete the emotional induction exercise (“describe your experience in a way that a friend would completely understand how you felt”)
-“drink the second beverage like you did the first in order to re-establish menial routine”
-complete the trust questionnaire
-complete the experimental information questionnaire.


An overall trust score was obtained for each of the 15 participants by averaging all scores in the trust questionnaire per participants (shown in Table 1). Participants in the sweet conditions rated the beverages perceived the beverage to be sweet (M = 7, SD = 1.26). The participant’s ratings of preference for sweet taste in beverage was highly correlated with their perception of the sweet taste in the beverage, r(5) = .95, p=.01.

To determine the effects of emotional induction and beverage type on participants rating of trust, a 2×2 Factorial Analysis of Variance (ANOVA) was conducted. The ANOVA did not reveal significant differences on levels of trust between the type of beverage and type of emotional induction on participants F(1, 11) = .07, p = .97, ƞ p< .01 (Fig. 1). There were neither main effects of type of beverage on levels of trust, F(1, 11) = .185, p = .68, ƞ p= .02, nor were there main effects of type of emotional induction on participants levels of trust, F(1, 11) = 1.34, p = .29, ƞ p= .16.


This research had two main aims: to understand how an individual conceptualizes trust and to add to the literature demonstrating perceptual symbol systems by identifying a link between sensory perception and embodied metaphor on trust. Using sweet stimuli, we were unable to verify our hypothesis that the perceptual state of sweetness had an effect on levels of trust reported by the participant. As well, we found no effect of happy or sad emotional conditions on levels of trust reported, contrasting findings by Dunn & Schweitzer (2005).

One explanation for the result of the embodied metaphor not having an effect on trust would be that sweetness did not contribute to the participant’s perceptual system. Trust may be a higher order, amodal representation and not affected or encoded by sensory-motor states but act more like a computer would, drawing conclusions with an encoded language. This is predicted by Gabora in response to Barsalou (Basalou, 1999, p. 617). Gabora posits that amodal thought may be the tree that sprouts from the seeds of perceptual symbol systems. A similar model is proposed by Murphy (1997), in which he argues that some metaphors cannot be embodied (such as “love is a financial transaction”) and his research points to processing without a perceptual basis but rather through categorical hierarchy.

Unfortunately, all explanations for the null result must be taken lightly: the amount of participants was too low to generate sufficient power to run proper statistics in a factorial design. Future research should be conducted in order to obtain larger samples of individuals in order to generalize results. Should it be found either way that the perceptual symbol of sweetness affects trust, this could add to the cannon of research available in support of perceptual symbol symbols and embodied metaphor1,2,5,15, or could display some amodal tendency of thought18.