Hypotheses
Experimental Hypotheses-
There will be a significant increase in the number of words recalled when semantically (deeply) processed compared to when visually (shallowly) processed.
Null Hypotheses-
There will not be a significant increase in the number of words recalled when semantically (deeply) processed compared to when visually (shallowly) processed.
The hypothesis is a one tailed (directional) hypothesis that shows the direction that the results are expected to go. A one tailed hypothesis is specific; it has been used on the basis of previous research which also suggests the results will go in a specific direction. Craik and Lockhart’s theory tells us that semantic is the deepest level of processing, this can be applied to this experiment in which case it is expected that there will be increased recall in the semantic condition compared to the visual condition, this is stated in the hypothesis.
Level of Significance and Justification
The level of significance used was 0.05 as this was a good balancing point between making a type 1 and type 2 error.
Methodology
Design
The method was an experiment and the design was independent groups. There were two conditions, condition one was the visual condition and condition two was the semantic condition. One group did condition one and the other group did condition two.
The independent variable were the questionnaires (word lists) that the experimenter designed and controlled, the dependent variable was the results (number of words recalled).
Before the experiment general consent was gained and the participants were given the right to withdraw at any time. A script was used to introduce the experiment, see appendix 1. At the end of the experiment the participants were given a full debrief and informed that the experiment was a memory test investigating levels of processing, see appendix 2.
Participants
The researcher was a 17 year old, female, second year A Level student.
The target population were female and male students between the ages of 16-17 studying level 3 qualifications.
It was an opportunity sample, I used a class that became available to me.
The actual sample experimented on were 20 16-17 year old students, 9 females and 11 males.
Participants were assigned to the conditions randomly; in this case every other participant was given a different word list (visual or semantic).
Apparatus/materials
The materials consisted of two questionnaires including word lists with 10 words on each list, one list with questions leading to visual processing (see appendix 3) and the other with questions leading to semantic processing (see appendix 4). When designing the word lists I avoided rhyming words and repetition to minimise confounding variables.
Standardised procedure
Two word lists were designed with 10 words on each. A question was assigned to each word. For word list one questions relating to the appearance of the word were assigned, this leads to visual processing, see appendix 3. For word list two questions relating to the meaning of the word were assigned, this leads to semantic processing, see appendix 4. The words in the list did not relate to anything in a classroom as this was where the experiment took place and could have provided memory cues. This would have affected the results. At the bottom of the page was a distracter task to minimise demand characteristics.
AS first year students were experimented on to ensure that the participants had no idea what was being studied and observed and so therefore decreased demand characteristics. The researcher introduced herself using a script, see appendix 1, general consent was gained. People were given the chance to withdraw at any time and made aware that if they did not wish to participate they could leave the room. As a control measure the participants were asked to remain silent during the experiment and to keep their work to themselves. The word lists were given out face down and alternatively so that people sitting next to each other had a different one so they could not copy each other. The participants were given 3 minutes to answer the questions and complete the distracter task and asked to put the sheet face down again when they had finished. The word lists were then collected in. Blank pieces of paper were given out to each participant and they were asked to recall and write down as many words as they could in 2 minutes and to turn the paper over once they had finished. These were then collected in. The participants were then debriefed see appendix 2.
Controls
There were two conditions involved in the experiment (visual and semantic). The questionnaires were alternatively assigned to the group.
Everybody was given the same amount of information about the experiment and treated in exactly the same way i.e. participants were given the same amount of time to do the task and the same amount of time to recall the words.
The results for the experiment were all obtained in the same hour so the time factor could not affect the results (memory may be better at different times of the day).
The participants were asked to remain silent during the experiment so they could not discuss or confer.
Results
Descriptive Statistics
Summary Table showing the Mean, median, Mode and Range
The table shows that the mean number of words recalled in each condition out of a possible 10 words. There is a higher mean number of words recalled in the semantic condition (5.6) compared to the visual condition (2.6), the mean for condition two is over two times higher than condition 2 suggesting that the semantic condition led to the deepest level of processing.
The median is much higher in condition two compared to condition one confirming that recall is much better for semantic rather than visual learning.
The mode shows the most common number of words recalled in the conditions, this is much higher in the semantic condition showing a higher most common recall (of 5 words) compared to the visual condition where the most common number of words recalled was 1.
The range shows the spread/distribution of the scores. The range is lower in the semantic condition, showing that the scores were closer together (more dense) whereas the visual scores are more spread out and less consistent.
This graph shows the mean number of words recalled in each condition out of a possible ten words. The graph shows that the mean number of words recalled in the semantic condition is over two times higher (5.6) compared to the visual condition (2.6). This suggests that the semantic condition led to the deepest level of processing and therefore increased people’s recall.
Inferential Statistics
The research method was independent groups and the data (words recalled) was interval but was turned into ordinal data for the purpose of statistical testing. This meant the Mann Whitney ‘U’ test was the correct statistical test to use.
Independent groups are where one group does one condition and the other group does condition two. In this case one group did the visual condition (condition 1) and the other group did the semantic condition (condition 2). There was a 0.05 level of significance, N = 10, N = 10 and a one tailed hypotheses was used.
The observed value of U was 8 and the critical value was 27. The observed value was less than the critical value therefore the null hypotheses could be rejected as the results supported the experimental hypothesis.
The results were statistically significant at the 0.05 level.
For mathematical calculations see appendix 7.
Discussion
Explanation of Findings
The aim of the investigation was to find out if semantic processing leads to better recall than visual processing.
The findings showed that significantly more words were remembered in the semantic condition compared to the visual condition suggesting that semantic processing leads to increased recall and so is the deepest level of processing. The mean score for condition 1 (visual condition) was 2.6, the mean score for condition 2 (semantic condition) was 5.6, this supports the fact that the average number of words recalled is higher in the semantic condition than in the visual condition therefore the null hypotheses can be rejected as there is an increased number of words recalled when semantically processed compared to when visually processed. The Mann Whitney stats test showed the observed value of U (8) was less than the critical value (27) therefore the null hypothesis was rejected.
An additional finding was also noted from the participant’s results, a primacy recency effect was observed. Participants remembered more words from the beginning and the end of the word list. This means that they remember the most recently processed words (at the end of the list) because they are fresher in the mind but also remember the first words they read (primacy)- first impressions. The participants from the visual condition particularly remembered one word, in fact 8/10 recalled the word ‘shell’. This word was the only word on the list in colour. It was in a shiny bright colour which stuck out on the page which made people remember it. Words in bold and capital letters were also recalled by many participants compared to plainly typed words in a standard black Arial font.
Relationship to background research
The findings of this experiment showed that thinking about the actual meaning of a word (semantically processing a word) leads to increased recall of the word suggesting that semantic processing is the deepest level of processing as over double the amount of words were recalled in this condition compared to the visual condition where only the physical appearance of the word were looked at.
These findings are backed up by Mandler (1967). He found that the degree of organisation according to modality, language, or category lead to increased recall. In order to categorise and organise, thought must go into the meaning of the word and so this leads to semantic processing. This explains why more words were recalled in the semantic condition of this experiment; more thought was involved in the processing of the semantic questions compared to the visual condition which led to deeper processing. In Mandler’s experiment those who had sorted into the most categories had the best recall suggesting that the act of organising makes it memorable without conscious effort or rehearsal as deeper processing is automatically taking place when the meaning is thought about in order to categorise.
Another piece of research that supports the findings of this experiment is Craik. Craik and Lockhart’s Levels of Processing Model (1972) identifies three levels of processing including visual and semantic processing. Craik and Lockhart proposed that information could be processed in a number of different ways and the durability/strength of the memory trace was a direct function of the depth of processing involved. They suggested that the depth of processing falls on a shallow to deep continuum. They concluded that deeper analysis produces more elaborate, longer lasting and stronger memory traces which explains how participants recalled more words from the semantic condition, they had made stronger traces enabling the memory to be remembered. Craik backed up his own theory in his elaboration investigation, Craik and Tulving 1975. They argued that elaboration (the amount of processing of a particular kind) is important, as shown in this experiment, semantic processing was important in increasing recall.
The results also showed that the primacy recency effect occurred when participants recalled the words from the list. Participants tended to remember more words from the beginning and the end of the list. This is supported by Glanzer and Cunitz 1966 research. They explain it using the familiar U-shaped serial position curve which shows that more words are remembered from the beginning (primacy) and end (recency) of a list rather than the words in the middle.
Limitations and Modifications
In this experiment one of the limitations was the sample. The sample size was fairly small (only 20 people took part, 10 for each condition). This meant that quite a small amount of results were gathered. The sample itself was a limitation in terms of the type of people. The participants were all AS first year student and between the ages of 16-17 so the group was limited. To modify this, a bigger sample could be gathered and from the general population so the participants would be from a wider radius rather than just college students. For example you could go into town and randomly select a sample.
In the procedure independent groups were used. This made it very difficult to rule out and control individual differences for example one group may have just had better memory than the other group affecting the reliability of the results, if the test was to be repeated with the same participants but by mixing the groups it may not obtain the same results. To modify this, a different experimental design could be used such as repeated measures. When using repeated measures all the groups do all the conditions, this prevents individual differences from affecting the results.
The use of two word lists was problematic as one word list may have been more difficult than the other therefore the difference in difficulty could act as a confounding variable affecting the results of the participants recall between the two groups. In future, one word list incorporating both levels of processing would be much easier as all the participants would be doing the same thing, ruling out any individual differences between the two groups.
The participants had to recall a list of words they had answered questions on. This is fairly unrealistic because in everyday life we are not required to remember random words, the experiment lacks ecological validity. To modify this the experimenter could have used a shopping list (including food/drink etc) as this happens in real life and has some significance.
The environment that the experiment took place in was problematic as the classroom had lots of posters and displays on the wall as well as other objects in the room, this could have provided the participants with cues to some of the words on the list, jogging their memory rather than them recalling it freely. For example the word ‘clock’ appeared in the list and there was a clock on the wall in the classroom, looking at the clock could have provided a memory cue to this word. To modify this the experiment could be carried out in a plain room (experiment room) with just tables and chairs, with no posters or other objects that could provide any cues. Instead of this the experimenter could ensure that none of the words on the list related to anything in the classroom, but this would be more difficult.
Implications and suggestions for further research
The information gained from this experiment could be used in everyday life. The fact that semantic processing leads to increased recall could be used as an advantage. In schools teachers could incorporate this level of processing in children’s learning to increase children’s recall and memory. By getting children to think about the meaning of words rather than just their appearance should help them to understand and remember the word. Using this level of processing in teaching will increase the likelihood of what’s being taught to enter the children’s long term memory rather than their short term memory and then being displaced. It could also be used for spelling test expansion, using the meaning to help with spelling.
For further research an investigation into the comparison of how semantic processing increases young people’s recall in comparison to older people’s recall. Different levels of processing may suit different ages. The results may reveal that different levels are more appropriate for different ages. Older people are likely to already have a wider range and depth of knowledge compared to younger people, in which case a lower level of processing may be more sufficient and less time consuming. Whereas with younger people who are at school or college are still in the process of learning and need to remember a great deal of facts to pass exams, so in this case a deeper, more intense level of processing may help the information to sink in and stay in the long term memory ready to be brought back into the STM to be recalled in an exam.
It would also be interesting to add in the third level of processing which is acoustic. It would be interesting to see how the results would alter. Some people’s preferred learning style involves saying the word or having it played on a tape recorder and so this may increase their recall.
Conclusion
Semantic processing appeared to significantly increase the recall of words compared to the visual condition. The findings of this investigation have supported the hypothesis and already existing research. Therefore the null hypothesis was rejected. It has also shown that more time was needed to carry out the investigation in order to gain more results from a larger sample and this has proved to be a limitation and something that could be rectified if the experiment was to be repeated.
References
Craik, F.I.M. and Lockhart, R.S. (1972) Levels of processing: A framework for memory research.
Craik, F.I.M. and Tulving, E. (1975) Depth of processing and the retention of words in episodic memory.
Eysenck, M.W. Eysenck (1980) Depth, elaboration and distinctiveness. Levels of processing in human body.
Eysenck, M.W. and Eysenck M.C (1980) effects of processing depth, distinctiveness and word frequency of retention
Mandler, G. (1967) Organisation and Memory, Levels of Processing
Palmere, M. Benton, SL., Glover, J.A. and Ronning, R. (1983) Elaboration and recall of main ideas in prose, Journal of educational psychology
Tyler, S.W., Hertel, P.T., McCallum, M.C, Ellis, H.C. (1979) Cognitive effort and memory journal of experimental psychology.
Bibliography
Cardwell. M, Clark. L, Meldum. C (1996)
Psychology For A Level
Collins Educational, Hammersmith
Cardwell. M and Flanagan. C (2003)
Psychology AS, The Complete Companion
Cheltenham
Nelson Thornes Ltd
Rice. D. et al (2000)
Psychology in Focus AS Level
Causway Press, Ormskirt
Website-
Appendix 1
Before the experiment took place I introduced the experiment. I used the following script:
Hello, my name is Laura Bishop. I would like to ask you if you would mind taking part in my experiment to help me with my coursework. If you do not wish to take part you have the right to leave the room now. If you do wish to take part you can withdraw at any time or even withdraw your results at the end of the experiment.
I am going to hand out a sheet of paper to each of you, I would like you to answer all the questions and turn your sheet over when you have finished. You will have 3 minutes to answer the questions. During the experiment it is compulsory that you remain silent and keep your work to yourself. I will instruct you further when you have completed the questions.
From now on can you please remain silent, I will answer any questions you have after the experiment. If you do not wish to participate please leave now.
(When the time is up I continue with my script):
Time is up now, can you please turn your papers over and I will collect them, while also handing out another piece of paper.
Can you now please write down as many words as you can remember from the words list. You will have 3 minutes to do this from now.
Appendix 2
Debrief
At the end of the experiment the participants were debriefed. The following script was used:
You have just taken part in a memory test.
The experiment was investigating levels of processing and how they affect your memory/recall. There are three levels of processing including visual processing which is the shallowest level of processing and involves looking at the appearance of the word, acoustic processing which involves listening to the sound of the word and the final and deepest level is semantic processing which analyses the meaning of the word. The deepest level also incorporates the other two levels which is what makes it the deepest level.
There were two different word lists handed around, one with questions leading to visual processing (the shallowest level) and one with questions leading to semantic processing (the deepest level), you had one or the other. People that took part in the semantic condition are very likely to have remembered more words than the visual condition.
If you would like to know the overall results for condition one and condition two you can email me. If you have any other questions please feel free to ask me.
Thank you
Appendix 3
Psychology Experiment
- sheep Is the word in upper case?
-
muscle Is the word in italics?
- hair Is the word in red?
-
radio Is the word in bold?
-
leaf Is the word underlined?
-
Is the word in black?
- CLOCK Is the word in lower case?
-
road Is the word in italics?
- pillow Is the word in lower case?
- Time Is the word in yellow?
Now please complete the task below:
49+24+8-11+162x3-122 =
I will now collect this sheet from you.
Thank you for taking part in my experiment.
Appendix 4
Psychology Experiment
1)GREEN Is the word a type of colour?
2)apple Is the word a type of car?
3)boat Does this ever go in water?
4)flour Can you use this to make bread?
5)drawer Is this a type of animal?
6)Pencil Can you draw with this?
7)Camera Is this a type of bridge?
8)Music Can you eat this?
9)CANDLE Can you burn this?
10) Brush Is this a type of food?
Now please answer the task below:
43+29+4-30+182x2-122 =
I will now collect this sheet from you.
Thank you for taking part in my experiment.
Appendix 7
Results
Descriptive statistics; Number of words recalled in each condition
This table shows the results gained from the experiment. It shows how many words each individual recalled and shows the total words recalled for each condition. Over double the amount of words were recalled in the semantic condition compared to the visual condition suggesting that semantic is a deeper level of processing and leads to increased recall. The most common number (mode) of words recalled in the visual condition was 1, the mode for the semantic condition was 5, again this shows that the semantic condition is more successful in increasing recall.
Calculations of mean, median, mode and range;
Mean for visual condition =
No: words recalled by each participant in visual condition ÷ No: participants in visual condition
1+1+1+1+2+2+4+4+5+5 = 26
26÷10 = 2.6
Mean for semantic condition =
No: words recalled in semantic condition ÷ No: participants in semantic condition
4+5+5+5+5+5+6+7+7+7 = 56
56÷10 = 5.6
Median
Number of words recalled ranked in order, find the middle.
Median for visual condition = 1 1 1 1 2 2 4 4 5 5 = 2
Median for semantic condition = 4 5 5 5 5 5 6 7 7 7 = 5
Mode
= The most common score
Mode for visual condition, 1 1 1 1 2 2 4 4 5 5 = 1
4x1
2x2
2x4
2x5
Mode for semantic condition, 4 5 5 5 5 5 6 7 7 7 = 5
1x4
5x5
1x6
3x7
Range
= The highest value/score – the lowest value/score
Range for visual condition = 5-1 =4
Range for semantic condition = 7-4 =3
Appendix 8
Man Whitney test including mathematical calculations.
A one tailed hypothesis and a significant level of 0.05. The critical value= 27
As the observed value of U (8) is less than the critical value (27) the null hypothesis can be rejected.
The results are statistically significant at the 0.05 level (even at the 0.005 level).