Null Hypothesis: ‘There will be no significant difference in the amount of words recalled from structural, phonological and semantic questions’
Method
A questionnaire has been developed to assess participants recall ability when different levels of processing are required. This questionnaire will consist of eighteen questions each associated to a word positioned next to the question. The participant will have to answer ‘yes’ or ‘no’ to each question. They will have two minutes to answer these questions. After completion of the questionnaire it will be taken away and the participant will have to write down all the words that they can remember. The results should provide valid data that will be analysed and evaluated.
The research method that will be used is an experimental method. This was selected and considered most appropriate as this task involves altering factors and measuring the effect on another. I have chosen a repeated measures design as all the participants will be using all of the independent variables throughout the one task. This is the most appropriate choice of design as it is carried out all at once and the questions do not follow an order. This eliminates order effects from the experiment.
As this is an experiment, the independent variable will be manipulated and the dependant variable will be measured. The independent variable in the experiment will be the type of question processing: structural, phonological or semantic. This was defined as:
Is the word written in red? Rug _____
This is a structural question and the participant will answer in the space provided next to the word ‘yes’.
Does the word rhyme with look? Book _____
This is a phonological question and the participant should answer ‘yes’ in the space provided.
Is this something that can be worn as clothes? Tree _____
This is an example of a semantic question that the participant will answer ‘no’ in the space provided.
The dependant variable will be the number of words recalled by the participant after the completion of the questionnaire. The participant will write down all the words they can remember from the right hand side of the questionnaire from memory. The number of words for each process that is recalled will be tallied and then be statistically analysed.
An extraneous variable within this experiment could be noise and distraction. This would be a problem as it may cause the participant to not pay full attention to the task and may therefore give poorer results. If levels of noise or distraction fluctuate while carrying out the experiment then one participant may be affected more than another which could cause anomalies. This would make it much more difficult to produce a conclusion form the results as the amount of valid data will be reduced. In a noisy or distracting situation the amount of words that a participant can recall would most likely fall as they may not be fully focused on the task. To control this extraneous variable the task should take place in a quiet room with only one participant present at a time.
When carrying out the experiment there may be certain ethical issues that could arise. The main ethical issue is confidentiality. As each participant produces individual results they should be ensured that their score will remain completely confidential. Keeping the results of each individual participant separate will control this factor. When drawing the results table and conducting the evaluation participants should be referred to as ‘participant 1’ ensuring that no names are exposed in the experiment write up or to anyone else. Also, participants will have the right to withdraw throughout the whole experiment as it requires a high level of concentration. If the participant feels that they are not ‘doing well’ in the recall task they may become under considerable stress. Informing the participant in the instructions that they can withdraw from the experiment at any time by requesting to stop will control this. They also can completely withdraw their results from the experiment. Participants will be notified about the aim of the experiment before carrying out the task. This eliminates any problems concerning deception.
Participants
The target population of the study is students in Tapton sixth form in Sheffield, South Yorkshire. This population is built up of about two hundred and fifty students, male and female, ranging between the ages sixteen to nineteen. The reason for this choice is that it is easily accessible to get a wide range of pupils willing to do the study. Therefore the sampling method that will be used is opportunity sampling. This method simply involves selecting those participants that are around and available at the time. This is a quick and convenient method which is suitable for this experiment due to the time restraints. This method is more appropriate than a systematic sampling method. In a systematic sample, names of all students in the sixth form would have to be collected and then for example, every twentieth person would have to be selected. This is a lot more time consuming than opportunity sampling. There will be a set of twenty results and therefore twenty participants will be required. Due to some questions in the investigation the participants must not be colour blind. Apart from this factor there are no other requirements from the participant.
Materials
In the experiment there will be eighteen questions that each participant will have to answer. There will be six questions for each level of processing: structural, phonological and semantic. Having eighteen questions ensures that the participant will have minimal order effects of boredom and fatigue while still being a sufficient amount to be able to draw a valid conclusion. Having the same number of questions for each participant for each level of processing will provide reliable accurate results. The words associated with each question will be monosyllabic and commonly used words to allow equal chance of the word being recalled and to make sure no other factor apart from the level of processing affects the recall. Having more than one syllable and complicated words may affect the consistency of the results as the recollection may be affected by the knowledge or familiarity of the participant with the particular word. A stopwatch is needed to make sure each participant has the same amount of time to answer the questionnaire (two minutes) and has the same length of time for recall (two minutes). A pen will be needed for the participant to write down what they recall. Therefore paper will be required for participants to write down their answers.
Procedure
Potential participants were approached and given a brief. The brief allowed the participants to decide whether they wanted to participate in the experiment. This demonstrates ethical treatment of the participants because participants were not pressurised into taking part in the investigation as they had the freedom to refuse to do so. Within the brief the participants were made aware that they had the right to withdraw from the experiment at any time and without enlightening the experimenter of the reason(s). Participants were also told that they would remain anonymous so they would not be pursued again, by the experimenter or any other researcher to take part in further investigations.
Brief: ‘Hello, I am carrying out an experimental study for my AS level psychology coursework and was wondering whether you would like to participate in it. The study involves a series of simple questions that only require a ‘yes’ or ‘no’ answer to them. The results of the study are completely confidential and no one will access to your results apart from me. If you choose to take part then you will have the right to withdraw at any time during the experiment as well as withdrawing your results.’
If the participant opted not to take part in the study then I replied ‘Thank you for your time’ and moved on to the next potential participant. If the participant agreed, a quiet place would be found for the experiment to be carried out. The participant was then given verbal standardised instructions:
‘I am going to present you with eighteen questions which you are required to answer with either ‘yes’ or ‘no’. You will have two minutes to complete the questionnaire. You can withdraw from the experiment at any time. Time starts now. (After the questions have been completed and the time has commenced I will take away the question sheet an provide the participant with a blank piece of paper) Now I will restart the stopwatch for another two minutes and I would like you to write down as many words as you can remember from the list running down the side of the question sheet. Time will start now.’
Each participant would do their study on their own with no distractions. The same instructions were given to each participant. Communication from me to the participant was minimised to allow the participant to have full attention on the task unless they needed to ask a question or request to stop. The participants do not have to use the whole two minutes if they do not wish to.
After the experiment the participants were debriefed about the aims of the investigation and the intentions the experimenter had for the results. For example, who would have access to the results and where they would be published. Participants were reminded that their anonymity would be kept. If there were any objections then participants had the right to withdraw their results from the experiment and this would be done immediately. The debrief also thanked participants for taking part. Furthermore, participants may be under the impression that their intelligence was being observed the test. The participant therefore needed to be told the exact aim of the experiment as they were not told about the recall section of the questionnaire. This is also a reason why all participants had to be debriefed. The experimenter needed to clarify this because participants may have been reluctant to have their results published and this may put them at ease.
Debrief: ‘Thank you for taking part in this experiment. The study was carried out to investigate the effect that different levels of processing have on the recall of information from memory. For this reason, you were not informed that you would have to recall the list of words. However if you have been told the effort put in to learning the word would eliminate the level of processing required as you may only concentrate on learning the particular word. My hypothesis states that words that require a semantic (meaning) level of processing will be more likely to be recalled rather than words that require a structural (appearance) or phonological (sound) level of processing. You may still withdraw your results from the experiment if you wish to. If not your result, along with the other participants, will be analysed and evaluated. All the participants’ results will be kept completely confidential, including yours. I would like to thank you again for taking part in the experiment. If you wish to know more information about my investigation please feel free to ask any questions.’
This procedure was repeated with each participant.
Results
A table to show the results of 20 participants, results in a study investigating the levels of processing theory
Analysis of results
The results from this experiment showed that participants were able to recall more words from semantic questions than phonological or structural questions. The mean shows this. The mean number for semantic recall, 4.05, was significantly higher than that of the phonological mean number, 1.95, and the structural mean number, 1.00. A limitation of this measure of central tendency is that one anomalous result can alter the mean greatly. However as the possible outcomes for this experiment only range from 0-6 it would be difficult to consider a result to be an anomaly. Therefore the mean value is an accurate measure for this experiment. The median is found by placing all the numbers in order and locating the middle value. Since the middle value has an even number of observations in this experiment the mean of the two middle values is taken. The modal value is found by counting the most common value. The median and modal values are also higher for words associated with semantic processing rather than phonological and structural processing. This further implies that semantic words are more likely to be recalled. The median and mode also suggests that words associated with phonological questions are more likely to be recalled than words associated with structural questions. The standard deviation as calculated to measure the spread of the outcomes for each level of processing. The calculation shows that there was a wider range of results for semantic followed by phonological and with structural having the smallest range of values. This suggests that some participants excelled at recalling the semantic questions while others were average or slightly below average. However almost all participants performed poorly on structurally based questions. This implies that the results for shallow processing were more consistent than deeper processing. As the level of processing became shallower, the more consistent the results became.
Discussion
The results show that the experimental hypothesis: ‘More words from questions that required a semantic process will be recalled rather than words that needed a structural or phonological level of processing’ was correct and is accepted. Therefore the null hypothesis is rejected. Almost every participant could recall a greater amount of words that were connected to a question that required a semantic level of processing rather than a structural or phonological level of processing. The results clearly showed that more words could be recalled when processed on a semantic level rather than a phonological or structural level. The results also showed that phonological processing is deeper than structural processing. This is evident from the results as the mean, median and mode is highest for the semantic level, followed by phonological level and finally the structural level. However from the experiment there does appear to be more consistent as processing becomes shallower, therefore the outcomes for structural processing are the most consistent followed by phonological and then semantic processing. This is shown by the standard deviation. The standard deviation for the structural level produced the smallest value, and therefore has the narrowest spread. The standard deviation for the semantic level produced the largest value and therefore has the widest spread.
The results of this experiment evidently confirm what Craik and Lockhart’s experiment had suggested. They believed that semantic processing is the deepest process and therefore would be more likely to be recalled in a task like this experiment. Craik and Lockhart found that the overall percentage of words recalled from structural questions was 15%. The overall percentage recall of phonological questions was 35% and the semantic overall percentage was much higher at 70%. These figures can be directly compared to the overall percentages from the experiment I carried out:
The results of my study clearly show evidence for processing and support Craik and Lockhart’s theory as the percentage figures are so similar with only a slight difference in each level of processing.
The results of the study that was conducted shows that a significantly greater amount of words can be recalled when processed semantically as opposed to being processed phonologically or structurally.
The study produced reliable and accurate results that highlighted Craik and Lockhart’s idea of the levels of processing. However the study I conducted does have limitations. One major limitation is that the amount of effort the participant puts into the initial questionnaire can have an affect on their ability to recall the words. The questionnaire is built up of 18 simple questions which means that most participants assume that the task will finish after the time to complete the questionnaire has commenced. Therefore participants go through the questions with minimal effort and sometimes hardly read the question which will effect their ability to recall the words when asked. This also caused some participants to rush through the questionnaire as they attempted to predict that the study was aimed at who could complete it the fastest. This affects their result on the recall task as the process has no time to be stored. This is a difficult limitation to overcome as deception has to be part of the experiment. If the participant was informed that they would have to perform a recall task after the questionnaire they may just learn the list of words down the side and not read the question. Therefore the recall will not be affected by the level of processing. Another limitation is the familiarity with the word that the participant has. For this study commonly used, monosyllabic words were chosen to try and limit this factor however participants will still be more familiar with some words than others. This creates a problem with flashbulb memories, which suggests that distinctiveness is a factor in recall. As the participant was not informed that a recall task was to be carried out the length of time they spent processing the information would influence hove effective their recall would be. This is another limitation. This limitation would make it impossible to see if the length of time of processing or the depth affects the participants’ recall. This again could only be eliminated by informing the participant that a recall task will be carried out after the questionnaire. The sampling method used was opportunity sampling which may not give reliable and accurate results as the population was generalised to a specific group (6th form Tapton students). Most students in the 6th form are in general, of high and similar intelligence, at the same brain developmental stage and would therefore be expected to produce similar results. The experiment was carried out during school when most students are distracted by school work and other thoughts at that particular time despite the study being carried out in an isolated, quiet area.
If the study were to be repeated, there would be several ways to improve the reliability and the success of the investigation. Firstly, the experiment could be carried out in a different environment to remove the distractions of school work and other studies. For example, testing students while at home at the weekend. This would hopefully reduce the limitation of participants’ minds being on other things. In order to increase reliability and to be able to generalise the results, a different sampling method could be used in and in a different population. In order to decrease the limitation of the length of time of processing the standardised instructions could be changed to stress the importance of the questions.
There is several different ways that further research for this study could be carried out. Participants could be grouped, for example, in ages to see if age affects the level of processing efficiency. The level of intelligence could be tested to see if high intelligence levels increases recall efficiency. People of different personalities, jobs and ethnic backgrounds could also be placed in groups to see if processing is affected. The same study could be carried out orally instead of in written form to see if this aids or restricts the amount of recall. The independent variable could be changed so that there are 3 separate groups that each complete only one level of processing. From this it would be clear to see if structural processing produces the least recall and whether semantic processing produces the greatest amount of recall.
References
-
Craik, F.I.M., and Lockhart, R.S. (1972), cited in Taylor, I. Active Psychology. Longman. pp. 646 – 647.
-
Craik, F.I.M., and Lockhart, R.S. (1972), cited in Pennington, D. Introducing Psychology Approaches, Topics And Methods. Hodder Arnold. pp. 270-271.
-
Craik, F.I.M., and Lockhart, R.S. (1972), cited in Carstensen, L., and Ornstein, R. Psychology, The Study Of Human Experience Third Edition. Harcourt Brace Jovanovich. pp. 332-333
Appendices
Mean recall number of each level of processing
Semantic = 81/20 = 4.05
Mean number of semantic words recalled = 4.05
Phonological = 39/20 = 1.95
Mean number of phonological words recalled = 1.95
Structural = 20/20 = 1.00
Mean number of structural words recalled = 1.00
Median number of each level of processing
Semantic = 1, 2, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 6, 6, 6
4 + 4 = 4
2
Semantic median = 4
Phonological = 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3
2 + 2 = 2
2
Phonological median = 2
Structural = 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2
1 + 1 = 1
2
Structural median = 1
Modal value of each level of processing
Semantic = 3, 4, and 5
Phonological = 2 and 3
Structural = 1
Standard deviation of each level of processing
√ ∑ (number of words recalled – mean)2
Participants
Semantic standard deviation
S.D = √34.95 = 0.2956 (4 decimal places)
20
Phonological standard deviation
S.D = √18.95 = 0.2177 (4 decimal places)
20
Structural standard deviation
S.D = √10 = 0.1581 (4 decimal places)
20
Total percentage recall of each level of processing
Semantic = 81 x 100 = 67.50%
120
Phonological = 39 x 100 = 32.50%
120
Structural = 20 x 100 = 16.67%
120