Avalle, Martino (2025) Does the aggregation of multiple forecasts (made at different time points) improve the prediction accuracy compared to the use of the most recent single forecast? Bachelor thesis, Psychology.
|
Text
Bsc-Thesis--Martino-Avalle---s5164257--.pdf Download (781kB) | Preview |
Abstract
Past research has explored the concept of aggregation in the context of human forecasting. Specifically, within-person aggregation has been found to be advantageous in cases where only singular sources are available. This process works according to the idea that averaging forecasts from the same person at different times can reduce random error. In the current paper, we try to translate this idea to the field of affective forecasting. We hypothesized that aggregating the two previous predictions for any time point would offer a predictive advantage compared to just using the single most recent forecast. Our participants repeatedly rated their current affect and predicted the emotional levels they would feel in 3 and 6 hours for 5 times a day for 14 days (n=30). These predictions were given for four emotions along the dimensions of positive and negative affect. Absolute errors were calculated, and our results showed no significant difference between using aggregation and using the single most recent forecast. The results were similar and non-significant for both positive and negative affect. Our findings are, thus, consistent with the idea that the single most recent forecast is generally more accurate than the previous ones, and contrast, within the boundaries of our experiment, with the idea that aggregation could be useful to diminish predictive error. We suggest that this could be because affective forecasting is mainly driven by systematic biases rather than random noise, which aggregation does not cancel. Keywords: aggregation, affective forecasting, time, wisdom of the inner crowd, affect, accuracy, prediction error
| Item Type: | Thesis (Bachelor) |
|---|---|
| Supervisor name: | Weigelt, O. |
| Degree programme: | Psychology |
| Differentiation route: | None [Bachelor Psychology] |
| Date Deposited: | 28 Jul 2025 11:21 |
| Last Modified: | 28 Jul 2025 11:21 |
| URI: | http://gmwpublic.studenttheses.ub.rug.nl/id/eprint/5747 |
Actions (login required)
 |
View Item |