ResearchPod Summary
As e-commerce adoption grows among older adults, researchers are increasingly concerned with how aging affects the ability to make optimal purchasing decisions. While product comparison tables are a standard feature in online retail, it remains unclear how individual cognitive differences, such as working memory, influence the accuracy of these decisions. This study bridges the gap between cognitive aging psychology and human-computer interaction (HCI) to evaluate how age-related cognitive decline impacts real-world e-commerce product choices.
The researchers conducted an online study with 149 participants across three age groups (younger, middle-aged, and older adults). Participants performed 64 multi-attribute product comparison tasks involving washing machines, which varied in difficulty. The tasks were designed to reflect real-world e-commerce interfaces, following standard UX guidelines. The researchers measured decision accuracy and time, while also assessing participants' visual working memory, numerical ability, need for cognitive closure, and a specific scale measuring intellectual helplessness regarding infectious diseases.
The study confirmed that decision accuracy significantly decreases with age, even when using standard, well-designed comparison tables. While all age groups spent more time on more complex tasks, this increased effort did not translate into better performance for older adults. Visual working memory was identified as a critical mediator: the age-related decline in this cognitive ability largely explains why older adults struggle more with complex product comparisons. Furthermore, the researchers found that intellectual helplessness acts as a moderator, where higher levels of helplessness were associated with lower decision accuracy in younger and middle-aged groups, but not in older adults.
These findings suggest that current UX guidelines for product comparison tables are insufficient for older consumers. Because older adults frequently make sub-optimal choices, recommendation algorithms trained on their behavioral data may develop a "self-induced bias," leading to less effective recommendations for this demographic. The results highlight a need for new, simplified, or non-numerical interface designs that do not rely on the cognitive processes that decline with age.
AI-generated third-party summary by ResearchPod. Not official content or an endorsement by the paper authors or affiliated organizations.