Independent measures: This involved using two groups, but all the participants took part in the same conditions. The advantages of this are, there were no order effects as each group did the experiment once. Both groups were mixed ability tutor groups.
The independent variable in the experiment was whether the questions asked were about the meaning or appearance of the word. Some words required semantic processing whereas others involved structural processing. The number of words recalled was the dependent variable.
Many variables were controlled. Participants were given standardised question sheets. (see appendix 2). This prevents any bias from the way instructions were set out. The same age group of participants was used. The target group was of mixed ability year 11 students. Participants were both male and female. The time limit given for answering the questions and, effectively, remembering the words and time allowed for recall was the same each time. A criticism of the LOP model is that it is the time and effort that affect recall and not depth of processing. Depth and effort cannot be measured, but time of attention was kept the same. The experiment was carried out at the same time of day, and during the same lesson. The noise level was controlled by asking the participants to remain silent and not confer. Participants were spaced out equally to prevent conferring or cheating. The question paper was collected immediately after the allocated time was up.
An extraneous variable in the experiment could be the way instructions were presented, some participants may have understood the task better and therefore more reliable results could have been achieved than those who did not understand so clearly. Also, students may have attended different previous lessons e.g. P.E or Maths prior to the experiment and come to the lesson with different amounts of energy and enthusiasm. The standardised instructions (see appendix 4) were given by the same experimenter.
Demand characteristics were evident in some individual cases. Participants guessed the paper to ‘lean on’ was to be used for another task. However, this is unlikely to have affected the experiment too much as it was mostly speculation and not many participants took notice of it. There were no order effects due to participants only doing the test once.
Due to the nature of the experiment, deception of participants was necessary or else participants would have displayed demand characteristics and the experiment would have been biased. Participants were given a piece of the paper to ‘lean on’, they believed this was for that purpose and not for an unexpected recall task. Consent from their parents had been acquired after a notice in the school newsletter was sent out to them, giving them the right to withdraw if they did not wish to take part, however informed consent was not gained. This raised some ethical issues but was addressed by debriefing participants as to what was the aim of the study and explained why it was conducted. Participants were also told about how confidentiality was maintained, they were asked not to write their name on the answer sheet, as their results were anonymous. In addition, they were also told they could contact the experimenter if they wished to know the results of the study, or if they wished for their data to be destroyed. As participants were under 16, parental consent had been sought. Also, it was important participants were not made to feel inadequate, or ridiculed in any way which would lower their self esteem.
Participants:
The target population was Year 11 students at Unity College, Northampton. Participants were of mixed ability and both male and female and aged between 15 and 16 years old. Participants were chosen by opportunity sampling, as this is quick, convenient and easy. Participants chosen were available at the time, and it did not disrupt their lessons. However, this type of sampling is unrepresentative of the population at large, and biased.
There were thirty-eight participants in the sample. It was not known whether English was the first language of all participants.
Apparatus & Materials:
The experiment required a set of question papers, consisting of 16 questions, of which 8 required semantic processing and 8 required structural processing. The use of capital letters, lower case letters and “yes” or “no” answers were equal, as to not influence recall in any way. Participants may recall words written in capital letters or answered YES better than those in lower case or answered NO, or vice-versa. A sample question paper can be found in appendix 1. The words selected were all nouns, and were between five letters and ten letters long, so recall would not be too difficult. The words were chosen using a random word list and could be used frequently in everyday language. All words had at least two syllables.
On the question paper, participants were asked how a word looked. For example:
“Is the word in capital letters?” The words assigned to these types of questions require use of structural processing. Other questions asked about whether or not the sentence made sense. “Does the word fit?” The words assigned to these types of questions required understanding what the word meant and using semantic processing. Two participants were used in a pilot study, they were asked to complete the question paper before the study was conducted for real in order to quash any ambiguities. Participants were timed using a stopwatch for 2 minutes to answer the paper and 2 minutes for recall of words. Standardised instructions (see appendix 4) and debrief (see appendix 5) was used.
Procedure:
Two classes of mixed ability were approached with permission of their tutors. Previously, consent from their parents had been acquired after a notice in the school newsletter was sent out to them, giving them the right to withdraw if they did not wish to take part. It was necessary to get parental consent as participants were under 16. (see appendix 7) Participants were briefed, (see appendix 3) read out a set of standardised oral instructions, and given a question paper each. Participants were told that they have all been given a piece of paper with a list of sentences and a piece of paper to lean on. They were then given two minutes to write the answer YES or NO at the end of each given question. Participants were then given 2 minutes to complete the questions after which they were asked to turn the question paper over. All question papers were collected in at this point and now participants were asked to recall as many words as they could remember from the right-hand column in a time-limit of 2 minutes.
Finally, upon collection of their papers, the participants were debriefed by telling them that their work was confidential and anonymous and would be destroyed if they did not want it to be used. They were also told what the aim of the study was. Any ethical issues such as deception were resolved and participants were thanked for taking part and were asked to contact the researcher at a later date if they wished to know the results of the experiment.
Results:
A table to show how the mean number of words recalled using semantic and structural processing:
A table of all the data recorded may be found in appendix 1
The table and bar chart showed participants recalled more words when required to use semantic processing rather than structural processing, over double the average amount of words were recalled.
The mean number of words recalled by semantic processing was 235/38 = 6.18.
The mean number of words recalled by structural processing was 111/38 = 2.92.
The standard deviation of each condition was 1.71 and 1.32. This showed there is a greater variance in results in the semantic condition. Examples of participants answers may be found in appendix 6.
Discussion:
The study found that words recalled by processing of their meaning were recalled more frequently than words recalled processed by their appearance. The mean number of words recalled semantically was 6.18, in contrast to 2.92 of structural. However there was not much variation in standard deviation 1.71 and 1.32.
The results strongly support the experimental hypothesis, which was: more words using semantic processing will be recalled than those requiring structural processing. It was suggested words processed by meaning would be recalled more frequently than those processed purely on appearance. This is because more time is spent thinking about the word if it has to be processed semantically, as the meaning has to be found to make it fit in context of the sentence whereas the appearance of the word is purely visual and takes less time to process. It may be the more time and effort put into processing the word, that makes it more frequently recalled. There were a few anomalous results where structurally processed words were recalled more frequently than semantically processed words, this was against my hypothesis so deemed as anomalous. The anomalies achieved were not as a result of methodological error, but due to variations in personal recall techniques.
Although the results appear to agree with the Levels of Processing model, the model itself has come under scrutiny. Baddeley criticised the model for being over-simplified, and merely describing rather than explaining. Eysenck also argued the lack of comment on how information is stored and retrieved from memory.
There is a circular logic in saying that more information is recalled because it has been processed deeply if the criterion for deep processing is that more information is remembered. This is a specific criticism of the LOP model. The LOP is more of a notion rather than explanation of how memory works.
Craik and Tulving used the term ‘overspill coding’ to describe the processing of a word at more than one level of the model. It is difficult to imagine participants could only process a specific word structurally without thinking at all about what the word meant. Theories of attention suggest that automatic processing of all stimuli is unavoidable without conscious effort.
Morris et al also discovered that rhyming words or phonological words were recalled better than semantic words. This contradicts the LOP model as phonological level is shallower than the semantic one therefore recall should not be as frequent.
Atkinson and Shiffrin suggested that memory is divided into structural
components including short-term memory which has a limited duration, and
long-term memory which has an unlimited duration. According to their model, information is passed from STM to LTM through the process of rehearsal or repetition. The results supports this, the more the words have been rehearsed in the mind, the better the frequency of recall.
Craik and Lockhart claimed that the idea of rehearsal did not explain
whether or not material became stored in LTM. They believed that it is the level of processing that determines whether or not something is stored in LTM. If something is processed deeply then it becomes stored in LTM; if it is processed superficially then it does not. The results strongly support this, the study has very similar ties with Craik and Lockharts study.
Craik and Tulving carried out a piece of research based on three levels of processing: structural, phonetic and semantic. They presented participants with words using a tachistoscope and asked them one of four types of question about each word.
This is a similar same study but carried out without a tachistoscope, the participants were presented with a list of preset questions involving structural and semantic words only.
One methodological limitation was that the study was a lab experiment. In general, such experiments lack ecological validity. Also in a lab experiment, participants are well aware they are taking part in an experiment and this may have lead to demand characteristics. Demand characteristics were evident in some individual cases. Participants guessed the paper to ‘lean on’ was to be used for another task. However, this is unlikely to have affected the experiment too much as it was mostly speculation and not many participants took notice of it. More limitations include that of ‘evaluation apprehension’ whereby participants’ performance may be affected by concern that their behaviour is being judged by the experimenter and ‘social desirability’ effects where behaviour may be distorted so it seems as if the participant is more impressive than they really are. Some participants cheated and tried to copy each others work. They felt a need to achieve a good score even though they were told it was not a test.
Another drawback is that of experimenter bias. The results of the study could have been subject to unintentional bias. However, an improvement for next time is to use a double blind technique, where the collection of the data is carried out by a third party, so that neither participant nor person running the study is aware of what the expected results should be.
Words which bore personal meaning to the participant and therefore made it easier for the participant to remember, would limit the findings of the study. There was no guarantee that all the words had no personal links to the participant, this may have affected results. The effect would be more frequent recall on those particular words, however this limitation could have been quashed by the use of abstract words.
As an improvement, participants could have been tested in smaller groups so it was less likely that they would cheat by looking at each others paper. The sample size was rather small, a much larger sample taking in all age groups is needed to gain conclusive results. Socio-economic groups should be also assessed in order to have ecological validity.
The findings can be related to everyday life as the results show the deeper the information is processed the more frequently it is recalled. This would be extremely helpful in revision techniques, drawing spider-diagrams to summarise notes, asking yourself questions about the topic, and explaining the topic to someone who knows nothing about it. It is important the material being explained is well understood in order for it to be conveyed to another in a way that they also understand it. Revisiting the same topic in a number of ways reinforces learning.
Further research could be carried out to investigate whether participants recall of words increases more frequently after answering YES to a semantically processed answer rather than NO. Depth of processing or recall may be affected by positive responses.
Conclusion:
The study shows that words that required understanding the meaning in given questions (semantic processing) were more frequently recalled by participants than those associated with the appearance of the word (structural processing).
References:
Cardwell M, Clark L, Meldrum C (2000) Psychology for AS Level, Collins
Cox E (2002) AS Level Psychology for Specification B, Oxford
Pennington D (2002) Introducing Psychology Approaches, Topics and Methods, Hodder Arnold
l
Accessed on 21st March, 2007
Accessed on 15th March, 2007
Accessed on 26th March, 2007
Accessed on 21st March, 2007
Appendix:
PRACTICAL ACTIVITY BASED ON A STUDY BY CRAIK & TULVING (1975)
Read each of the following questions and answer YES or NO with respect to the word written on the right hand column.
- Does the word have seven letters? RHINO
- Does the word fit? I saw a _______. knowledge
- Does this make sense? My son likes ________. computers
- Is the word in capital letters? KEYBOARD
- Is the word in lower case letters? CURTAIN
- Is this a six-letter word? poster
- Is this sentence meaningful? Today I shall ________. PIANO
- Is this a five-letter word? RAZOR
- Does this make sense? My father is a ________. policeman
- Does the word fit? Brian enjoys _________. SWIMMING
- Is the word in capital letters? dishwasher
- Is this sentence meaningful? Yesterday I saw a ________. nonsense
- Is the word in capital letters? RUCKSACK
- Does the word fit? Gemma had a _________. shower
- Is the word in lower case letters? CABINET
- Does this make sense? Sometimes I feel like ________. doorbell
BRIEF:
I am a 6th Form student currently doing my AS Psychology coursework. I would very much appreciate it if you could help me by participating in this short experiment. Please remember this is NOT A TEST. This experiment is totally anonymous. Please do not write your name on the paper. All data will be held in the strictest of confidence and accumulated data may be destroyed at any given time at your say so. Please refrain form looking at each others papers. You have the right to withdraw if you feel uncomfortable participating. Thank you for your time.
INSTRUCTIONS:
I have given you all a piece of paper with a list of sentences and a piece of paper to lean on. Write the answer YES or NO at the end of the question. When you are finished, please turn your paper over. You have 2 minutes to do this experiment.
On your blank piece of paper, please write down all the words that you can recall that were written on the right hand column of the sheet I had just given you. You have 2 minutes.
DEBRIEF:
This was an experiment on the levels of processing within memory. The aim was to see which words were recalled most frequently. Your answers are confidential and anonymous and will be destroyed if you do not wish for them to be used. If you wish to know the results of the experiment, please contact me at a later date. Thank you again for your participation.
