Intersample variance of second-language readers should not be overlooked

Abstract Much of the literature on first (L1) second language (L2) reading agrees that there are noticeable behavioral differences between L1 and L2 readers of a given language, as well as between L2 speakers with different L1 backgrounds (Finnish vs German readers of English). Yet, this literature often overlooks potential variability between multiple samples of speakers of the same L1. This study examines this intersample variance using reading data from the ENglish Reading Online (ENRO) database of English reading behavior comprising 27 university student samples from 15 distinct L1 backgrounds. We found that the intersample variance within L2 readers of English with the same L1 background (e.g., two samples of Russian speakers) often overshadowed the difference between samples of L2 readers with different L1 backgrounds (Russian vs Chinese speakers of English). We discuss these and other problematic methodological implications of representing each L1 background with a single participant sample.