Jiří Lukavský^1,2, Natálie Šimčík^1,2, Filip Děchtěrenko¹; ¹Czech Academy of Sciences, ²Charles University

Boundary extension (BE) is the phenomenon where individuals remember a previously viewed scene as having a larger background extent than was originally presented. Over the past 30 years, numerous small-scale studies have explored this effect. This study synthesizes the existing literature to estimate the size of the BE effect and identify factors that influence its magnitude. We conducted a systematic search of the Web of Science and Scopus databases for studies citing the seminal work by Intraub & Richardson (1989). Inclusion criteria required studies to: (1) provide new experimental data, (2) focus on healthy adult participants, and (3) measure BE using a rating scale (e.g., -2 to +2). For each study, results were recorded for all close-up/wide-angle presentation combinations (CC, WW, CW, WC). When multiple experimental conditions were present, we selected the baseline condition, typically used for comparison against experimental manipulations. Data were analyzed using multi-level Bayesian meta-analysis, with results presented as credible intervals for the BE effect size and Bayes Factors for moderator analyses. The dataset included 110 outcomes extracted from 56 studies in 32 articles, encompassing data from 2,379 participants and 71,786 trials (mean: 30.2 trials per participant; 16.1 trials per outcome). On average, each study included 42.5 participants per condition (SD = 28.1). Our analysis revealed that the baseline condition produced a BE effect size of D = -0.222 (95% CI [-0.256, -0.189]). The effect was significantly influenced by presentation order of the close-up/wide-angle versions. However, there was strong evidence against the influence of other design choices (e.g., testing BE after each photograph or at the end, number of trials) or presentation factors (e.g., screen distance, viewing angle).

Acknowledgements: The research was supported by Czech Science Foundation (GF24-11506K).