A Comprehensive Meta-Analysis of Spatial Interference From Linguistic Cues: Beyond Petrova et al. (2018)

Estes, Verges, and Barsalou (2008) demonstrated a spatial interference effect, in which linguistic cues with implicit spatial associations (e.g., “bird”) hindered identification of an unrelated visual stimulus (e.g., “X”) at the implied location (e.g., at the top of a display).1 This effect, though counterintuitive, fits within a broader literature of visuospatial interference from linguistic cues (e.g., Bergen, Lindsay, Matlock, & Narayanan, 2007; Gozli, Chasteen, & Pratt, 2013; Richardson, Spivey, Barsalou, & McRae, 2003; Verges & Duffy, 2009). Most researchers explain spatial interference in terms of perceptual simulation (Barsalou, 1999): The linguistic cue evokes a subconscious mental image of the denoted object or event in its associated location, thereby visually masking (i.e., perceptually competing with) the target stimulus and delaying its identification. Recent alternative accounts have attributed spatial interference to more holistic event simulations (Ostarek & Vigliocco, 2017) or to conflicting semantic and spatial codes (Amer, Gozli, & Pratt, 2017; Estes, Verges, & Adelman, 2015). Estes et al. (2008) demonstrated spatial interference three times across three experiments. Those experiments shared many common features: They all included cue words with spatial associations (e.g., “hat”), visual targets that were unrelated to the cue word (“X” or “O”), brief delays between the cue word and the visual target (300 ms or less), and a task in which the target (“X” or “O”) had to be identified. The studies were conducted in English, which may be important, given that spatial interference is an effect of language on attention. Estes et al.’s first two experiments additionally included a context word to provide a spatial reference frame for the cue word (e.g., “cowboy hat”), whereas the third experiment presented the cue word alone (e.g., “hat”). Despite minor methodological differences in the delay between cue and target presentation and in the presence or absence of a context word, all three experiments demonstrated significant spatial interference. Petrova et al. (2018) question the reliability of the spatial interference effect when the cue word is presented in isolation. That is, they contest the spatial interference in the third experiment of Estes et al. (2008). Notably, they do not contest Estes et al.’s two demonstrations of spatial interference when the cue word was preceded by a single context word. Petrova et al. reported a series of experiments that varied in similarity to those of Estes et al.’s Experiment 3, successfully replicating the spatial interference effect only once. They also meta-analyzed their results together with those of a select few experiments from the prior literature. Although the overall effect in their meta-analysis was significant (N = 15, Z = 2.54, p = .011; see their Fig. 1), Petrova et al. nonetheless concluded that the spatial interference effect is unreliable. We disagree with Petrova et al.’s conclusion because (a) several of their experiments were conducted under conditions in which spatial interference is not theoretically expected, (b) their selective meta-analysis excluded many significant demonstrations of the effect, and (c) the overall effect in their meta-analysis was significant nevertheless. Petrova et al. revealed that such spatial cuing can additionally reveal a Simon effect, which occurs when the locations of the target stimulus and response key interact to affect responding. In the Supplemental Material available online for this Reply, however, we explain why the spatial interference effect cannot be understood as a Simon effect. Here, we report 794131 PSSXXX10.1177/0956797618794131Estes, BarsalouVisuospatial Interference From Linguistic Cues research-article2018