Interactive judgemental adjustment of initial forecasts with forecasting support systems

There have been a number of empirical studies on the effectiveness of judgmental adjustment to statistical forecasts. Generally the results have been mixed. This study examined the impact of the reliability and the source of the additionally presented reference forecast upon the revision process in a longitudinal time series forecasting task with forecast support systems. A 2-between (reliability & source), 2-within (seasonality & block) factorial experiment was conducted with post-graduate students using real time series. Judgmental adjustment was found to improve the accuracy of initial eyeballing irrespective of the reliability of an additionally presented forecast. But it did not outperform the dampened reference forecast. No effect was found of the way the source of the reference forecast was framed. Overall the subjects anchored heavily on their initial forecast and relied too little on the reference forecast irrespective of its reliability. Moreover they did not improve at the task over time, despite immediate outcome feedback.