Directed forgetting
The directed forgetting (DF) paradigm models the type of forgetting that occurs when we are explicitly instructed that certain information is unnecessary or unwanted (Bjork et al., 1998). This can occur when old information is updated with new, competing information. Imagine a jury is presented with one set of facts about a defendant, but then promptly told by a judge to forget this information and to focus on a new set of facts instead.
In the standard list-method directed forgetting (DF) paradigm,participants study two lists of words (List 1 and List 2). After studying List 1, half the participants
FIGURE 6.2 The list-method directed forgetting procedure (Bjork, 1970)
are told to forget List 1 items, and half are told to remember List 1 items. Both groups are told to remember List 2 items, which are subsequently presented (see Figure 6.2). Participants told to forget List 1 items recall fewer items from this list than participants told to remember List 1 items: this is the DF effect (Bjork et al., 1998). Notably, competition between to-be-forgotten (List 1) material and to-be- remembered (List 2) material is necessary for DF; there is no forgetting in the absence of List 2 learning (Bjork et al., 1998). DF impairs explicit memory while leaving implicit memory intact, as demonstrated by Basden, Basden, and Gargano (1993) using a word stem completion task. Also, DF can be abolished using a recognition test rather than a recall test (Basden et al, 1993; Bjork et al., 1998). Thus, it has been argued the DF impairs memory accessibility, but not availability, since these items can still be recalled given sufficient cues (as in a recognition task; but see Sahakyan & Delaney, 2005, for an alternative, non-inhibitory account of DF).
Like the RIF paradigm, researchers have examined whether DF effects are influenced by motivation. Again, this has generally taken the form of comparing DF for emotional (positive or negative) material with DF for unemotional material (for a review, see Koutstaal & Schacter, 1997). Are DF effects selective consistent with the functional view of remembering and forgetting? To test this Payne and Corrigan (2007), for example, examined DF of emotional and neutral pictures, and found a DF effect for neutral pictures but not for emotional pictures; that is, emotional stimuli were not forgotten. In contrast, Wessel and Merckelbach (2006) found DF effects for both emotional and unemotional words. More recently, Gamboa et al. (2017) reported significant DF effects for both neutral and negative stimuli, which was not influenced by giving participants mindfulness strategies to support their attentional control. As Payne and Corrigan (2007) argued, this might be because words are unlikely to elicit emotional responses in a normal population. Laying aside questions about the stimuli, Payne and Corrigan’s (2007) findings, as well as some RIF findings, suggest that emotional material—particularly negative material—might be resistant to forgetting, although such findings are not universal. This conclusion is consistent with the functional, selective view of remembering and forgetting outlined above, although it remains controversial whether and why negative material would be particularly resistant to forgetting (Anderson & Levy, 2002; Brewin, 1998; Erdelyi, 2006; Kihlstrom, 2002, 2006; McNally, 2005).
Like RIF, much research on DF has focused on clinical populations. For example, Geraerts, Smeets, Jelicic, Merckelbach, and van Heerdan (2006b) compared DF of neutral words with DF of words associated with child sexual abuse in either participants who had reported continuous memories of abuse, participants who recovered memories of abuse, and control participants. Unexpectedly, all participants demonstrated less forgetting (no or reduced DF effects) for abuse-related words. This is similar to Payne and Corrigan’s finding (2007), which suggested that emotional material may be immune to DF. In contrast, other researchers have reported that certain populations show more forgetting (greater DF effects) of negative material. For example, Moulds and Bryant (2002) examined patients with acute stress disorder. They found that these patients forgot more trauma-related words when given a forget instruction than controls (Moulds & Bryant, 2002). Myers, Brewin, and Power (1998) examined individuals with a repressive coping style (individuals characterized by low reported anxiety and high defensiveness). They found that repressive copers forgot more negative material when given a forget instruction than non-repressors (Myers et al., 1998). Similarly Myers and Derakshan (2004) found that repressive copers forgot more negative words when given a forget instruction than non-repressors, but only when they rated the words for self-descriptiveness; when they rated them for other descriptiveness there was no difference.
Taken together, these findings suggest that DF effects are selective: some research suggests that DF operates on all kinds of material, other research suggests that DF does not operate on emotional material, and still other research suggests that DF operates particularly for emotional material, and may depend on individuals’ memory biases. Although, as suggested above for RIF, memory valence may not fully capture motivational effects on forgetting in the DF paradigm, these findings lead us to ask how DF (like RIF) might influence autobiographical memories.
Joslyn and Oakes (2005) conducted a diary study to examine this. They asked participants to record ten events from their lives each week over a two-week period. After one week, half the participants were told that the first week was for practice (Experiment 1), or that the first week memories were for a different experiment (Experiment 2). Finally, participants were asked to recall all the events they had recorded from both weeks. Joslyn and Oakes (2005) reported a significant DF effect; participants in the forget condition recalled fewer week one memories than participants in the remember condition. This effect occurred for positive and negative events, and for high intensity and low intensity events (Joslyn & Oakes, 2005). In a closer adaptation of the original DF procedure, Barnier et al. (2007) also examined directed forgetting of autobiographical memories. In our adaptation, participants elicited autobiographical memories in response to cue words such as “happy” and “sickness”. Halfway through the words, participants were either told to forget or remember the first list, before eliciting memories for a second set of cues (List 2). Barnier et al. (2007) found a DF effect for positive, negative, and neutral autobiographical memories, although neutral memories were more likely forgotten overall than emotional memories. This contrasts with Barnier, Hung et al.’s (2004) findings for RIF, where emotional memories were more likely forgotten overall than neutral memories. Again, more targeted manipulations, such as whether memories are personally significant or not and whether memories are self-defining or not might help us to better understand these different patterns for emotional and unemotional memories (as well as emotional and unemotional simple material) and better capture the goal-directed nature of remembering and forgetting. Indeed, more recent research suggests that self-referential processing reduces or abolishes DF effects (Mao et al., 2017;Yang et al., 2013).
