Code‐Mixing in Cultural Educational Materials: Processing and Implications for Comprehension, Perceived Vitality, and Topic Interest

ABSTRACT This study investigated how the inclusion of multilingual content (American English and ʻōlelo Hawaiʻi) in short educational texts about Hawaiian cultural history affects students’ processing fluency, text comprehension, inferences about the vitality of ʻōlelo Hawaiʻi, and interest in learning more for readers who generally were not proficient in ʻōlelo Hawaiʻi. We conducted an experiment in which participants ( n = 458) were randomly assigned to one of three language conditions: no‐, low‐, or high code‐mixing (defined by the proportion of ʻōlelo Hawaiʻi to English in the texts presented). As predicted, participants in the high code‐mixing condition reported lower processing fluency and lower interest in learning more about Hawaiian culture than participants in the English monolingual and low code‐mixing conditions (which did not differ). Text comprehension was highest in the low code‐mixing condition. These results suggest that higher levels of code‐mixing in educational materials may be disruptive to students’ learning and interest when students are not proficient in all languages used in code‐mixing, but that low levels of code‐mixing do not appear to be detrimental and may even be beneficial.