Effect of the Levels of Processing on Memory:

Can subjects remember items better depending on initial exposure and level of processing of the item? We presented subjects with a list of yes-or-no questions that each corresponded to a word. The questions were of three categories, physical appearance, acoustic, and semantic. The categories represent the different levels of processing that we are studying. We found that more items for the semantic group will be correctly remembered because the questions involving semantic encoding require deeper processing. Thus, it showed strong evidence that the level of processing affected how something is remembered and recalled.
Effect of the Levels of Processing on Memory:

Does Semantic Encoding Produce Stronger Memories?

Countless studies have been done on memory and almost every aspect of memory. Craik and Tulving (1975) suggested that the different levels of processing explain why there are variations in memory. This is related to the extent that the information is actively processed in the mind. In their study, they presented participants with a list of questions and a corresponding list of words. They manipulated the level of processing from shallow to moderate to deep. Shallow processing referred to the physical characteristics; moderate referred to acoustic characteristics; finally, deep processing referred to semantic characteristics. Their study showed that the subject's memory were better for deeper processed words.

Another experiment by Rogers, Kuiper, and Kirker (1977) further analyzed the problem of levels of processing. They did a study that focused on a deeper level of processing than semantic characteristic. Their focus was to analyze the importance and effect on memory of self-referent encoding. In short, their experiment added a manipulation that focused on whether or not target words were self-descriptive. The results showed that self-descriptive words produced better recall than physical appearance or semantic characteristics. The implication of their findings is that information related to one's self receives preference in memory.

Fay, Isingrini, and Clarys (2005) performed a study to examine the effect that age and depth-of-processing during learning had on implicit memory task performance. Although they did not find evidence to support their original theory that ageing has a different effect on the word-fragment and word-stem completion tasks, they did find that younger subjects benefitted more from deep encoding situations than elder subjects. They also found evidence that supported Craik and Tulving's (1975) findings, which was that the depth-of-processing had a strong influence on how well the information is recalled.

In our study, we attempted to determine whether or not the depth-of-processing has an effect on episodic memory and how strong that effect is. Furthermore, we wished to differentiate between the three levels of encoding to display which one will yield the best results. Our work is based on Craik and Tulving's work in an effort to replicate their result. Based on the results of past studies, we expected the depth-of-processing of information will have a differential effect on how much information is recalled.

Method

Subjects

The subjects were 25 Rutgers University undergraduates who served as volunteers. 11 male and 14 female participants volunteered for the study. Thus, selection of subjects was voluntary and therefore, not random. Furthermore, all subjects live in on-campus dorms.

Materials

We used a prepared hand-out that includes a list of 18 yes or no questions about a word that corresponds to each question. The questions belong to one of three categories: physical appearance/structure; acoustic/rhyming characteristic; semantic/meaning characteristic. There were 6 questions in each category and the questions were scattered throughout the list. For example, the question would ask: Is the word in capital letters? and the corresponding word would be BOOK. The subject would be asked to circle yes or no. All subjects received the same list of 18 questions.

Procedure

During the study, all subjects were presented with the list of 18 questions. They were instructed to read each of the questions and answer yes or no with respect to the word that is associated with the question. The subjects were be given any other instructions. Once they finished answering all of the questions, they were instructed to turn the page over. Next, the subjects were instructed to write down as many of the items as they could remember (in any order). They were not permitted to look back to the previous page.

Once the subjects finished writing all of the words that they remember, we added up the total correct in each category: physical characteristic; acoustic characteristic; and semantic characteristic. (Only words that are correctly written were counted and order of the words written does not matter). Because the questions are broken up into 3 categories, we can easily add a tally for every word they correctly remember from the list to the corresponding category. We then calculated the total correct from each category for each subject. Finally, the total number correct for each subject was combined into one table. This allowed us to determine the average number of items correctly recalled per category.

Results

We used an one-way ANOVA to analyze our data because we had three categories. Once we do this, we can see if there are any groups that had significantly different number of correct recall. According to our ANOVA results, we found that there was indeed a significant difference between the number of correct responses per category (F = 10.346). While our data was significant at an alpha-level of .05 (critical F=3.32, α =.05), we wanted to note that our results were also significant at an alpha-level of .001 (critical F=7.32, α =.001). We found a significant result even at a very low level, which makes our findings important. Furthermore, we can look at the means of correct responses for each category and conclude that on average, more words semantically encoded were recalled (µ_semantic = 3.50, µ_acoustic = 2.44, µ_physical = 1.96) than words that were acoustically or physically encoded. Because the mean number of items recalled increased from the physical to acoustic to semantic category, we can further deduce that our theory of different levels of processing is supported. That is, the deeper the level of processing, the better recall. From this information that we hope to obtain, we can also propose explanations as to why deeper levels of processing results in better recall than shallow levels of processing.

To further assist with the explanation of our findings, we included tables to display our results. Table 1 will show the average number of correct responses per category and Table 2 will yield information on the ANOVA test.

Discussion

The study demonstrated several important points. We expected to find that correct recall for words in the semantic category to be the highest because semantic questions involve a deep level of encoding. Our results clearly support this. We also expected the fewest correct responses to be in the physical category because it involves a shallow level of processing, which leaves acoustic encoding in the middle. We also found this in our results. The implication of these results is that they can provide a number of practical applications. For one, it would explain why some information is remembered better than other information. It can also help improve study methods. Semantic encoding requires more neural activity and longer attention to the target word. This increase in attention can be a potential explanation for why semantic encoding results in better memory than physical appearance encoding. Furthermore, semantic questions require more thought than physical characteristic questions. Physical questions require a quick glance, while semantic questions require the subject to think about the question and figure out the answer. Because we cannot pinpoint the exact cause of our results, either the time taken to process the information or the level of processing can be used to explain the results. This is invariably due to deeper processing involves more difficult tasks, which require more time to complete.

Furthermore, we must note that our study focused specifically on written words. Thus physical appearance of the word should, in theory, have less impact on episodic memory than a vibrant picture. Our study cannot shed light on that but it does propose questions to be solved. It can potentially provide strong insight on the optimal ways to remember information. One study cannot prove everything, thus, it must be taken one step at a time.

References

Craik, F. I. M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General, 104, 268-294.

Fay, S., Isingrini, M., & Clarys, D. (2005, November). Effects of depth ‐ of ‐ processing and ageing on word ‐ stem and word ‐ fragment implicit memory tasks: Test of the lexical ‐ processing hypothesis. European Journal of Cognitive Psychology, 17(6), 785-802.

Rogers, T. B., Kuiper, N. A., & Kirker, W. S. (1977). Self-reference and the encoding of personal information. Journal of Personality and Social Psychology, 35, 677-688.