Lost in translation?: Examining couple behavioral coding across languages

Observational research is considered the gold standard for studying couple communication, and this method is increasingly being applied to diverse couples in global settings. However, observational coding systems have received little methodological scrutiny to determine if they are appropriate for use across contexts. The current study uses two samples of couples (N = 416 and N = 107) with multi-lingual data to test whether two different types of coding systems (the Iowa Family Interaction Rating Scales and the Brief Romantic Relationship Interaction Coding Scheme) exhibit measurement invariance across English- and Spanish-speaking couples. In Study 1, we observed strong support for measurement invariance across languages for all codes, whereas in Study 2, we found support for measurement invariance only for positive codes, with evidence of language-based measurement nonequivalence observed in negative codes. However, associations between observed communication and relationship satisfaction mostly did not differ by language. Together, these mixed findings suggest that observational coding systems can demonstrate measurement and functional equivalence across languages in some cases but may not be equivalent across all samples or behavioral contexts, particularly for negative behaviors. As the field prioritizes inclusion, tools for coding couple behavior should be rigorously validated across languages and contexts to ensure that we accurately capture couple communication in the populations we study.