Purpose: This paper aims to show that strength-based theories of memory provide only a partial description of how consumers retrieve brands from memory. Dual-process theories of memory such as the Source of Activation Confusion (SAC) model provide a more robust explanation of brand retrieval by accounting for the separate effects of brand familiarity and category knowledge. Design/methodology/approach: This paper examines brand image associations for 27 brands in three product categories using marketing field data. The authors apply a quasi-experimental approach to divide respondents into four groups based on their levels of brand familiarity and category knowledge. The authors compare brand retrieval for each group to test whether the SAC model, a dual-process theory of memory, or traditional strength-based theories of memory better explain brand retrieval. Findings: Familiar brands are harder to remember when consumers know more about the product category. This effect cannot be explained by strength-based theories of memory, but it is a prediction of the SAC model. This outcome is a critical test that discriminates between competing theories of brand retrieval. Research limitations/implications: Researchers may draw on the SAC model to identify new ways of analysing brand image data to better understand how consumers retrieve brands from memory. This includes, above all, developing methods to separately measure the effects of brand familiarity and category knowledge. Practical implications: To maximise the chance that consumers will remember brands, managers of highly familiar brands should avoid promoting category knowledge through their branding and communications strategies. By contrast, managers of less familiar brands should promote category knowledge by linking their brand to episodes of category consumption. Originality/value: This work illustrates that a quasi-experimental approach can be used to extend quantitative psychological models from laboratory experiments to marketing field data. It also illustrates the use of a critical empirical test to discriminate between competing theories in marketing.