Wednesday, 15 July 2009

Can we deliberately forget specific parts of what we've read?

The majority of research on memory is focused, as you might expect, on the remembering side of things - how much, how accurate and so on. There is, however, a parallel, but less known, line of investigation into our ability to deliberately forget. This is no mere academic curiosity. The ability to forget selectively that to which we've been exposed would be, if we had it, an extremely useful ability - a kind of refuse collection service for the mind.

In one of the first studies of its kind, Peter Delaney and colleagues have now shown that people are indeed capable of reading a series of sentences and then selectively forgetting just some of those sentences, whilst remembering the rest.

Dozens of undergrad students were first told to memorise an initial list of 16 sentences about the imaginary men, Tom and Alex. Afterwards, half the students were unexpectedly told to forget the Tom sentences, so as to better remember the Alex sentences. The remaining students acted as controls and didn't receive this additional instruction. Finally, all the students attempted to memorise a second list of 14 random sentences about another man, Joe. A 90-second multiplication test acted as a filler task before the students were tested on their ability to recall as many sentences as possible from the two lists.

The key finding was that the students were able to follow the forget instruction so long as the sentences about Tom and Alex were of random meaning, with no discernible theme. In this case, students told to forget the Tom sentences subsequently recalled just 28 per cent of them, in a two-minute free recall test, whereas the control students recalled 39 per cent. Moreover, there was also a non-significant trend for the students who deliberately forgot the Tom sentences to remember more Alex sentences (37 per cent) relative to the controls (32 per cent) - showing that their deliberate forgetting really had been selective for the Tom sentences. Memories of the second test list (sentences about Joe) were unaffected by the forget instruction.

Another version of the experiment had the Tom and Alex sentences gradually forming discernible themes - Alex as a writer who liked snow sports and Tom as a lawyer and family man, or vice versa. Curiously, in this version, the students were not able to deliberately and selectively forget the Tom sentences. The researchers aren't sure why, but one possibility is that remembering just one sentence in a theme involuntarily cues all the other sentences, thus hampering attempts to forget.

So how do we deliberately forget a portion of previously-seen material? One possibility is that the forget instruction prompts people to cease mental rehearsal of to-be-forgotten items, in favour of extra rehearsal of the to-be-remembered items. Another possibility is that items are selectively inhibited at the retrieval stage. "Future studies should use recognition tests to determine whether forgotten items are available in memory but blocked from access, or if they are less well learned," the researchers said.

ResearchBlogging.orgDelaney, P., Nghiem, K., & Waldum, E. (2009). The selective directed forgetting effect: Can people forget only part of a text? The Quarterly Journal of Experimental Psychology, 62 (8), 1542-1550 DOI: 10.1080/17470210902770049

Post written by Christian Jarrett (@psych_writer) for the BPS Research Digest.


Neuroskeptic said...

Whoa whoa whoa. Whoa. Couldn't the students have just been deliberately not writing down sentences from the list that they were told to forget? To make it look as though they had indeed forgotten?

To get around this, you could use an implicit memory task (word-stem completion, say) so that the subjects don't know they're being tested on memory...

Unknown said...

Hi Neuroskeptic - this is a good point and something I raised with the study author before publishing my report. Here's what he told me:

"We haven't specifically checked in the current version of the task whether people intentionally withhold items, although the instructions at test tell them clearly to recall both the TBF and TBR items. Therefore, the original instruction is canceled, and you would have to assume that people are somehow still going on the original instruction (and guessing that we still secretly want them to have forgotten). It's also not clear why you'd get a difference between the two types of lists we used if people were just trying to accommodate us.

Furthermore, the possible problem you describe -- called demand characteristics of the task by psychologists -- has been ruled out in other directed forgetting paradigms. For example, using both the list-method and item-method, researchers have offered incentives to get people to recall more words. One nice paper of this type is by Colin M. MacLeod (1999), who observed no change in recall levels after introducing a monetary incentive to recall. Additionally, there are indirect effects of directed forgetting instructions in the list-method paradigm that are difficult to understand if people simply try to comply with the original instructions at test. For example, Elizabeth and Robert Bjork (2003) have shown that people who see a list of non-famous names are more likely to later erroneously call those names famous after a forget instruction than after a remember instruction. They mistake the familiarity of the name (and the absence of conscious recollection of seeing that name on the list) as evidence that it is a famous person. It would take a very wily participant to do that."

Neuroskeptic said...

Fair enough - that's a pretty good defence of what they did! I'd still be interested to see the results from an implicit memory test, though.

Post a Comment

Note: only a member of this blog may post a comment.