Analysis of Judgmental Forecasting based on Communication Format of Advice

We examined how individuals behave in making judgmental forecasts when they are given advice in different communication formats: (a) face-to-face, and (b) computer-mediated. Participants were asked to make an initial forecast before receiving advice, and were told to make a final forecast after receiving advice. Participants were told the advice is made by a human expert, but one group received the advice face-to-face and the other received the advice through computer. We found that individuals in computer-mediated treatment discount advice more than individuals in face-to-face treatment. We also found that this advice discounting is serially mediated by the perceived difficulty level of the task and confidence level. Individuals in face-to-face treatment feel the task to be more difficult and feel less confident than individuals in computer-mediated treatment and therefore, take advice more. However, individuals in computer-mediated treatment are better at discerning the quality of advice than individuals in face-to-face advice. We conclude that computer-mediated advice is better than face-to-face advice in reality except when advice quality is known to be consistently reliable.