Improved Memory or Decreased Forgetting? Magnitude of Negative Generation Effect is Reduced in a Delayed Verbal Recall Test

There were two aims of the present study: First, to establish whether the amount of semantic encoding, beyond the mere elicitation of semantic encoding, is related to the size of the generation effect, and second, to corroborate Mulligan and Peterson’s findings that a negative generation effect (i.e., when reading yields better memory ability than generating) may be eliminated at longer delays. A total of 168 psychology students from the University of New Brunswick Saint John participated in one of two online experiments, comprising each of three reading tasks and three generation tasks across three stimulus types: nonword lists, wordlists, and passages. In experiment one, participants engaged in a free recall test of the material from the previous task after a 2-minute distraction. In experiment two, the delay between encoding and free recall was six minutes. Within-subjects factorial analyses of variance were conducted for each experiment, and cross-experiment analyses were also conducted. Although a negative generation effect emerged across both delay conditions, cross-experiment results indicated that there was no significant difference in memory ability for generated material over longer delays, suggesting that generated material may be more robust to forgetting than material that is merely read.