Activating learned exemplars in children impairs memory for related exemplars in visual long-term memory

Abstract

A popular educational method when teaching children new information is encouraging children to activate previously learned information held in long-term memory. A potential downside of this practice is the considerable evidence that there are negative consequences of accessing long-term memory representations. Specifically, it has been shown that accessing information in long-term memory can actually impair related memories. This impairment has classically been demonstrated with verbal material (i.e., retrieval-induced forgetting) and more recently with visual material in adults (i.e., recognition-induced forgetting). The goal of the present study was to examine whether recognition-induced forgetting exists in visual long-term memory for children aged 6–10 years old, the age at which retrieval-induced forgetting appears to emerge. To this end, we presented children with an abbreviated, age-appropriate recognition-induced forgetting paradigm. If children suffer recognition-induced forgetting, practicing a subset of objects from a category of objects will impair memory for the other, non-practiced objects from the same category. The results showed that children across all ages showed impaired memory for non-practiced objects from practiced categories and no benefit from recognition practice for the younger children. These findings show that children's visual long-term memory is vulnerable to recognition-induced forgetting and does not appear to benefit from recognition practice.

Keywords:

• Recognition-induced forgetting
• Visual long-term memory
• Developmental recognition-induced forgetting
• Recognition
• Retrieval-induced forgetting

ORCID

Ashleigh M. Maxcey http://orcid.org/0000-0002-6775-6836

Jessica Bostic http://orcid.org/0000-0002-6046-2579

Notes

¹Note that for the sake of accessibility to the reader, here we revise the nomenclature used in our previous recognition-induced forgetting paradigm (Maxcey & Woodman, 2014) by using the term “practiced objects” to refer to objects that were previously denoted “Rp+ items”, “related objects” to refer to objects previously denoted “Rp- items” and “baseline objects” to refer to objects that were previously called “Nrp items”.

²One concern might be that participants were relying on verbal long-term memory to complete this task. However, previous work using this recognition-induced forgetting paradigm has shown that verbal coding of the stimuli cannot explain the forgetting effects (Maxcey & Woodman, 2014). When subjects are required to verbally recode the visual stimuli according to instructions, the pattern of the effects across the types of trials was identical.

³The JZS Bayes Factor gives an indication of the strength of the alternative relative to the null hypothesis. A JZS Bayes Factor value of 1–3 is considered anecdotal evidence for the null hypothesis, while a value greater than 3 is considered substantial evidence for the null hypothesis.

⁴ B”D (Donaldson, 1992), indicates a conservative bias with positive values and a liberal bias with negative values. B”D for practiced, baseline and related objects was –.21, –.06 and .38, respectively. There was a main effect of memory test item on the bias measure B″ _D (F(1,44) = 12.78, p < .01), due to a significant difference between B″ _d values for related and practiced objects (t(44)=5.66, p < .001, d = 1.72) and related and baseline objects (t(44)=4.67, p < .001, d = 0.67) that reflected a conservative response bias for related objects.