Evaluation of the Craik and Tulving study

The experiments carried out by Craik and Tulving were very highly controlled. For example they used standardised procedures and the time for which the words were displayed was controlled by the tachistoscope so every word was displayed for exactly the same length of time.

Because the study was so well controlled it is easy to replicate the study. This is an advantage because it improves the reliability. The basic experiment has been replicated several times finding similar results.

The experiment was high in experimental validity due to the counterbalanced repeated measures design used. Participants are testing their recall in all of the conditions of the independent variable so this makes comparisons between their performance across conditions fair. The counterbalancing controls for order effects as participants had to answer the questions for each condition in no set order so it is unlikely that having to process information in one way would affect how they processed it in other ways.

The variable of depths of processing was also clearly operationalised in the experiment through the way the questions were asked – with each question related to a different type of processing.

It can be argued that although the sample size was small it did consist of a varied mix of participants (males, females, ages etc) suggesting that generalisation is possible.

There is also supporting evidence from Hyde and Jenkins (1973) which found that when participants carried out different orienting tasks recall was affected by the depth of processing required by the task. When participants were asked to judge something about the meaning of the words this led to 51% more recall than when they were asked to perform non semantic tasks such as having to cross out vowels.

However, studies have shown that there are other factors that can affect how well material is remembered. For example Reber et al (1994) found that emotional content of words can affect the ability to recall. Morris et al (1977) found that information that is relevant to the individual is often remembered best regardless of how they were told to process it.

Craik and Tulving’s experiment can be criticised for being low in ecological validity. Because the experiment was laboratory based and required participants to recall a word list it was not like a true to life experience of how people would normally use their memories. The task required participants to process information under test conditions which might affect effort. In reality people are not normally asked if a word is written in upper or lower case and then perform recognition tests so the test is very artificial.

Finally it is worth noting that participants did take longer to answer the semantic type questions so perhaps their memory was better due to longer processing time rather than type of processing.