A Test of the Knew‐It‐All‐Along Effect in the 1984 Presidential and Statewide Elections¹

Abstract

The knew‐it‐all‐along effect is investigated by comparing pre‐ and post‐election predictions of subjects concerning three electoral races in the November 1984 elections. The results revealed that when the outcomes of the predictions are known, subjects recall having assigned higher probabilities and percentages to the actual winners of each election, remember having more confidence in the accuracy of these percentages, and claim to have had more knowledge of the candidates than they had before the election. The results were consistent across all three elections and in both repeated‐measures and between‐pups designs, suggesting that the knew‐it‐all‐along effect is a very robust one. Two underlying processes are hypothesized to account for these results: the availability heuristic and an increase in confidence explanation.