The Oblique Effect and the Central Tendency Bias Share an Underlying Mechanism

Orientation estimates are shaped by two biases: the sinusoidal oblique effect (Appelle, 1972) and the central tendency bias (Pandey & Cave, 2026). This implies that when orientations are drawn from a restricted range (e.g., 0–30°), estimates regress toward the mean of the distribution (the central tendency bias), whereas sampling from the full orientation range yields the oblique effect. This apparent dissociation raises the question of whether these biases reflect distinct mechanisms or a common underlying process. Here, we test a parsimonious account in which both biases emerge from serial dependency in perceptual estimates. We empirically validate the predictions of a novel computational model showing that serial dependency generates both the negatively sloped pattern characteristic of central tendency bias and the sinusoidal pattern associated with the oblique effect. We also demonstrate for the first time that increasing the stimulus–response delay systematically modulates the magnitude of the oblique effect, paralleling effects previously observed for serial dependency and central tendency bias. These findings suggest that the oblique effect is not a feature-specific bias unique to orientation, but rather a manifestation of a more general central tendency bias that has been observed across visual dimensions (e.g., size, color, spatial frequency) and beyond vision. This work provides a unified framework for understanding two biases that have historically been studied in isolation.