On the way to the grocery store, you realize you've left your shopping list at home. Before your memory fades, you try to strengthen it by connecting each item to the context you plan to use it in. You might think about the ice cream sundae you plan to make--how good the bananas will taste when you add a scoop of ice cream and some fudge--or you might consider what you'll need for afternoon coffee with a friend.

That should help, right?

Well, yes and no. Because of what State University of New York at Cortland psychologist Michael Toglia, PhD, calls the "more is less" effect, many ways of trying to improve memory--such as trying to link a word or an image to related ideas--can come at a cost.

The problem is not that semantic elaboration, as it is called, doesn't increase your chance of remembering what's on the grocery list--it does. But it also increases your chance of remembering items that might not be on the list, such as fudge and ice cream.

Does that mean you should give up trying to enhance memory, or resign yourself to bringing home more groceries than you'd planned? Not quite. Two studies published this month suggest ways of making sure you bring home most of what you need and none, or at least not much, of what you don't.

Words get in the way

It's well-known that older adults have more trouble remembering grocery lists--and other things--than younger adults. But, more recently, researchers have discovered that older adults are not just prone to forgetting; they also more often remember things that haven't happened, says Wilma Koutstaal, PhD, a psychologist y at the University of Reading in England.

Understanding why is one of Koutstaal's goals. Specifically, she's been trying to tease apart the influences of perception and language on false memory. Do older adults have perceptual deficits that keep them from distinguishing between similar objects, thus leading them to mistakenly recognize new objects as old? Or do they allow verbal labels to get in the way?

To explore this, Koutstaal and colleagues at Harvard University studied older and younger adults' ability to remember both everyday objects and abstract figures that don't fit into pre-existing linguistic categories.

For everyday objects, the results are straightforward: Older adults remember less and have more false memories than younger adults. But in several experiments using abstract figures, the researchers found that the differences between older and younger adults shrink to insignificance.

"It was intriguing," says Koutstaal. "In studies using common objects, we'd tried various manipulations to get older adults to stop showing this pattern, but we could never eliminate it. [Now] we were finding something where they looked more similar to the young adults."

To find out why, Koutstaal and her colleagues conducted an experiment reported in the July issue of the Journal of Experimental Psychology: Learning, Memory and Cognition (Vol. 29, No. 4). The researchers showed 72 older and 72 younger adults ambiguous abstract figures, either with or without labels that identified the categories to which the figures belonged. Ten minutes later they tested the participants for recognition memory, again with or without labels.

When the figures were labeled both when participants first saw them and when they were asked to remember them, older adults' false alarm rates skyrocketed. Younger adults' rates, however, went up only modestly. The results suggested that older adults use labels as a shortcut, says Koutstaal.

"Older adults look very similar to younger adults in terms of true and false recognition for things that don't have names," she explains. "But for things that do have names, older adults show much higher rates of false recognition."

That might have been the end of the story, except for one methodological glitch: Baseline false alarm rates differed between the label and no-label conditions. When the researchers corrected for the differences, the key finding was no longer significant.

So Koutstaal and her colleagues conducted another experiment, this time with 18 older and 18 younger adults and both abstract figures and drawings of common objects. The results were clear: Older adults had much higher false alarm rates for common objects than for abstract figures, compared with younger adults.

Together, the experiments suggest that the way older people use verbal labels makes it easy for them to remember the gist of what has been seen, but harder to remember the details, says Koutstaal.

The study also suggests that older adults' memory isn't equally impaired in all situations, says Koutstaal.

"It's not like, 'Oh! We can't trust older people's memory!'" she says. "It's more that older adults have a special reason to be cautious under certain conditions--in this case, where they're somehow naming [an object] quickly. If you ask them to look harder, it might make a difference."

In fact, looking harder does make a difference. In a 1999 study in Psychology and Aging (Vol. 14, No. 2), Koutstaal and her colleagues found that older adults had fewer false alarms when they paid careful attention to objects' distinguishing features. They were never as accurate as younger adults, she says, but they came close.

Generative learning

There are other ways of improving memory that simultaneously increase true memories and decrease false memories. One of them is generative learning, in which people produce words from cues instead of passively reading them. In the late 1980s, psychologist Ian Begg, PhD, suggested that generation enhanced memory in part by increasing the distinctiveness of the to-be-remembered words.

Until recently, however, no one had looked at how generation affected false memories. Would it suffer from the same "more is less" effect as semantic elaboration? Or would it boost true memories without increasing false ones?

To find out, Tufts University psychologist Sal Soraci, PhD, and his colleagues--including Toglia--conducted a study that also appears in the July Journal of Experimental Psychology: Learning, Memory and Cognition. In the study's first three experiments, the researchers gave participants lists of words to read or generate, and then tested them for either recall or recognition memorya few minutes later.

"What we found was something wonderfully different," says Soraci. Unlike other manipulations, such as semantic elaboration or "deep" processing, generation boosted both recognition and recall at no cost--that is, without increasing false memories.

"A lot of people think that the levels of processing framework and generation are virtually the same--that if you generate something it's deep processing, and if you read it it's shallow processing," adds Toglia. "This shows that that's not the case."

In the fourth experiment, Soraci and his colleagues tried to clarify how generation works. They gave subjects more complex cues that were either "congruous" or "incongruous." The congruous cues matched the target word (e.g., part of a radio: s_eaker); the incongruous cues didn't (e.g., not a tennis shoe: s_eaker).

The results of the experiment suggest that not all kinds of generation are equal. Congruous generation actually produced more false memories than passive reading, while incongruous generation produced less.

Why the difference? Probably because the incongruous condition forced subjects to consider and reject alternative words, says Soraci, effectively "immunizing" them against later mistakes.

The scientific implications of Soraci's and Koutstaal's studies go beyond false memory. Soraci's suggests that generative learning is more complex than a simple active-passive dichotomy would suggest, while Koutstaal's indicates that memory and memory judgments--in both older and younger adults--are influenced by meaning.

"Even though you can categorize this as being about false memory--and it is--in both papers I think we're using false memories as a tool to understand more about how cognition in general works," says Koutstaal.