Individuals of many species prefer to associate with familiar conspecifics from their established social group over unfamiliar conspecifics. Such familiarity preferences are thought to be adaptive and have been documented widely in many social fishes. Zebrafish (Danio rerio) are extensively studied, highly social fish that form stable shoals in the wild, however there is only mixed evidence for familiarity preferences in this species. Here, we test how a small variation in experimental design can influence preferences for familiar conspecifics in D. rerio by varying the distance between two stimulus shoals of fish in a shoaling choice paradigm. Individual subjects tested for their preference to shoal with familiar or unfamiliar groups of conspecifics showed a preference for familiar fish when the two shoals were 30 cm apart, but not when they were 45 or 60 cm apart. Thus, choice tests can be used to detect familiarity preferences in D. rerio, but only when alternate shoals are close together, as increased intershoal distances seemingly prevent subjects from displaying their preference. Longer distances may inhibit preference behavior due to the higher risk of crossing between shoals, alternatively subjects may be unable to reliably discern distinguishing cues of familiarity when the shoals are further apart. Our results demonstrate that while familiarity preferences exist in D. rerio, experimental test design is critical for detecting and measuring these successfully.