Combining verbal forecasts: The role of directionality and the reinforcement effect

Abstract

Recent research has shown that when people combine verbal probabilistic forecasts from two sources, they are not simply averaged but can reinforce each other; so when two advisors both said an event was "rather likely," some listeners concluded that the event was "quite likely". Conversely, when both said the event was "rather unlikely," people concluded that it was "quite unlikely." The present studies demonstrate that the direction of this effect is not evoked by high versus low probabilities, but by the directionality of verbal probability expressions. Some phrases are affirmative, directed towards occurrences ("there is a chance"), whereas others are negations, pointing to the possibility that the event might not occur ("it is not certain").Two positive phrases are perceived to reinforce each other, even when they convey low probabilities, resulting in a higher combined probability estimate, whereas two negative phrases do the opposite, regardless of the probabilities they convey. We show that this effect occurs both for equal and unequal verbal phrases, regardless of the probability equivalents of the expressions. We also found a positive, but weaker, reinforcement effect of numerical probabilities.

Combining forecasts from advisors: The impact of advice independence and verbal versus numeric format.

“…Participants in our studies attended to advice independence both when forecasts were in a verbal format (e.g., rather likely) and when both were in a numeric format (e.g., 70% probability). Replicating prior work (Mislavsky & Gaertig, 2022;Teigen et al, 2023), we also found evidence of a general advice format effect: When combining verbal as opposed to numeric forecasts, people's final likelihood estimates became more extreme (closer to certainty).…”

“…If the advisors had both said “unlikely”—suggesting the event will not happen—a final response that reflects even more certainty that the event will not happen would be considered more extreme. This counting—or getting more extreme—may have to do with people’s sensitivity to the directionality implied in verbal likelihood terms, as noted by Teigen et al (2023). Terms such as “likely” not only suggest (vaguely) a probability value, but they reveal a positive focus—a focus on the event’s possibility of happening (Teigen, 1988; Teigen & Brun, 1999).…”

People prefer to predict middle, most likely quantitative outcomes (not extreme ones), but they still over-estimate their likelihood

“…The survey also included an unrelated task presented at the onset of the survey, where participants combined two verbal probability forecasts. The task required participants to estimate the probability of an event that was described as either “not certain” or as having “a chance” to occur by two different forecasters (more details about the method and results are reported in the work of Teigen et al, 2022a ).…”

On the robustness of the brand positivity effect: Is impulsivity a moderator of overly favorable judgments and choices of focal options?

“…Unlike the warm room example, in many instances of everyday decision making the best explanation or solution is much more opaque. Indeed, most judgment and choice problems are characterized by myriad possible solutions, and evidence that is both voluminous and subject to differing interpretations (Teigen et al, 2023). Consider an individual who wishes to purchase barbeque sauce for an upcoming cookout.…”