Inferences and the role of prior knowledge

Abstract

Information in a message can either be fully expressed (explicitly) or indirectly stated (implied) and understood by inference or association. Previous research suggested an age-related decline in performance of implicit compared to explicit information and that this relationship is moderated by prior knowledge. Whereas previous studies mainly obtained quantitative data of inferencing performance, the current study employed both quantitative and qualitative techniques to understand age-related differences in inferencing. Twenty younger and older participants evaluated whether a series of one-sentence statements were true or false based on specific two-sentence text passages. Text passages either resembled real warnings (taken from actual products) or were novel (the opposite of a warning found on an actual product). Statements either explicitly stated information from the text passage or required participants to go beyond information given in the text. Quantitative analysis showed that older adults accuracy compared to that of younger adults when evaluating real text passages, with explicit items being evaluated more accurately than implicit items. For novel text passages (generally lower accuracy scores), younger adults showed the same pattern as for real text passages, whereas older adults accuracy was low for both explicit and implicit statements. Qualitative analyses supported that participants correct answers generally reflected that the intended inference was drawn and that for incorrect answers the inference was not mentioned. The data also suggested that accuracy scores may underestimate the actual ability to infer. Both age groups mentioned most often that text-related factors (e.g., clarity) influenced their decision and brought outside information (e.g., education, experience, expectations) to the task regardless of text passage or statement type. Older adults more often referred to outside information than younger adults, particularly when evaluating novel text passages and their answer was wrong. This study substantiated that age-related differences in a task requiring inferencing may be explained by a combination of the factors of working memory (time and availability of information) and prior knowledge as well as a possible decline in inferencing ability. Prior knowledge is important for both age groups and especially so for older adults. Important implications for designers are to make information available and explicit.