Cognitive Psychology: Connecting Mind, Research and Everyday Experience, 3rd Edition

176 • CHAPTER 7 Long-Term Memory: Encoding and Retrieval

noting if the word is printed in capital letters. But

let’s consider a more diffi cult problem. What about

using a word in a sentence (as in the Craik and

Tulving experiment) versus deciding how useful

an object might be on a desert island (as in the

Demonstration). Which of these results in deeper

processing? Unfortunately, levels-of- processing

theory does not offer a way to answer this question.

One possibility that might be worth trying is

to pit these two procedures against each other in a

memory experiment. If, in our hypothetical experi-

ment, participants in the desert island condition

remember more than participants in the fi ll-in-the-

blanks condition, then couldn’t we conclude that

the desert island condition resulted in deeper pro-

cessing? Although this may sound logical, a little

refl ection indicates that this procedure does not

really solve the problem. We can appreciate why

by considering the reasoning behind the memory

experiment more closely.

We started by asking whether the desert island

task  causes deeper or shallower processing than

the fill-in-the-blanks task. To answer this question, we ran a memory experiment

and determined that the desert island task resulted in better memory (● Figure

7.2a). From this, we concluded that the desert island task results in deeper process-

ing (Figure 7.2b). This seems to have solved the problem, but not so fast! Once

we have determined that the desert island task results in deeper processing, we

can predict from this that memory will be better for the desert island task (Figure

7.2c). This is called circular reasoning, and it occurs because depth of processing

has not been defined independently of memory performance. We can’t use memory

performance to determine depth of processing and then turn around and use depth

of processing to predict memory performance.

Because no procedure was offered to defi ne depth of processing independently of

memory performance, levels-of-processing theory became less popular with memory

researchers. But the main conclusion of levels-of-processing theory—that memory

retrieval is affected by how items are encoded—is still widely accepted, and a great deal

of research has been done that demonstrates this relationship.

RESEARCH SHOWING THAT

ENCODING INFLUENCES RETRIEVAL

A number of different procedures have been used to show that encoding can affect

retrieval. The basic idea in all of these experiments is to vary encoding and measure

how retrieval (memory performance) is affected. We will show how this has been done

in experiments that involve (1) placing words to be remembered in complex sentences;

(2) forming visual images based on words; (3) forming links between words and per-

sonal characteristics; (4) generating information; (5) organizing information; and

(6) testing.

Placing Words in a Complex Sentence If you were given the task of remembering the

word chicken, which sentence do you think would result in better memory?

1. She cooked the chicken.


2. The great bird swooped down and carried off the struggling chicken.

Craik and Tulving (1975) found that memory for a word is much better when the

word is presented in a complex sentence. Their explanation for this result is that the

● FIGURE 7.2 The circularity of defi ning depth of processing in terms

of memory and then predicting that deeper processing will result in

better memory. See text for details.

Can therefore

conclude that

Can therefore

predict that

(c) Memory will be

better following the

desert island task.

(a) Result: Memory is

better following the

desert island task.

(b) Desert island

task causes deeper

processing.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.