Rationale
I am going to try and replicate Craik and Tulving’s study in order to see if semantically processed words are still most deeply processed and remembered best in today’s society, hence my aim is similar to theirs, however I will be making a few minor adjustments to it. Firstly, I will conduct a laboratory experiment which was carried out by them; it might not have very high ecological validity however it will make sure all the external variables are controlled. I also intend to keep the same principal of a word list, questions and an unrelated task that Craik and Tulving used but of course I will use different words, questions and the task from Craik and Tulving that the participants have to perform.
Aim
I am going to try and see if the words that are processed semantically really are remembered better than those that are processed phonetically or structurally in today’s society or if this is not in fact true.
Hypothesis
Semantically processed words will be remembered better than structurally processed words.
Null Hypothesis
Semantically processed words will not be remembered better than structurally processed. Any difference will be due to chance.
Method
Method/Design
I have chosen to conduct a Laboratory experiment in order to test my hypothesis. Because Laboratory experiments are conducted in controlled environments, it will be much easier for me to conduct the experiment without too many interruptions, which should help the participants concentrate more on the task and reflect the true abilities of their memory. The experiment will however have low ecological validity.
Variables
I will have the opportunity to control the external variables such as weather and noise levels. Laboratory experiments do not have very high ecological validity (as field experiments or observations do), the results will have higher validity. If for example I was to conduct a field experiment, noise levels might disrupt the participants from remembering some things faster and true results won’t be reflected in their answers
Participants
I am going to get 20 participants between the ages of 16 and 20 From Harrow Weald College. I will have to keep the ratio of males to females of 1:1 in order for my experiment not to be biased. As I do not have a lot of time to do this experiment I will have to use the opportunity sampling method, this will involve only choosing people who are willing to take part in my experiment however I will still be able to be sure all the participants fall into the age and gender gap if I perform the task in college, where there's a wide choice of participants suited to my age and gender group. If I chose a random sample from the list of available people, I would spend a long time chasing them up as some of the people might have had a lesson or work at particular times on the other hand if I chose a self-selecting method some of the participants might not have been willing to take part. Opportunity sample will insure that anyone that wants to take part will do it when they have time and are not distracted from the task by pressures of wanting to quit.
Apparatus
- Standard Instructions – in order to save time briefing people
- Paper for sums (Exercise 2) – for participants to be enabled to do sums on
- Pen – to record their answers
- Quiet Room – to annihilate any chances of disruptions
- Stopwatch – so I can make sure everyone one gets an equal amount of time not to be biased
- Word list – which i prepared from the random words I thought up which matched the questions. Word list is used to enable participants to recall the words they remember. I tried to pick the words which are relatively common and everyone would know their meaning. I also kept them between 3 and 7 letters long (the average number of items a person can hold in the memory is 5)
Procedure
Firstly I asked the person if he was able to participate in my experiment and then I made sure he has not performed this task or any similar to it in the past (in order for my experiment to be valid). I then took each participant individually to a quiet room nearby where they could concentrate on the task, then I briefed them, answered any questions they had and gave them the word list with instruction on the first page. The task involved giving participant a list of questions with related words next to them. A person then performed an unrelated task (Calculating Maths sums) after which the participant was asked which words he could remember from the first task. I allocated and timed 15 minutes for each respondent to perform the tasks. After they have completed the task I debriefed the participant and thanked him for taking part.
Controls
Noise levels could have been a huge influence on how the words are remembered hence I have used a quiet room to make sure this situational variable did not intervene with my experiment and make the participant forget some of the words which might otherwise have been remembered.
Results
Summary Table
The table below shows how many structurally and semantically processed words were recalled by each participant along with the mean, mode and median number of recalled words.
Table Commentary
From the table, the mode number suggests that most people remember around 7 words if they are processed structurally and around 15 if they are processed semantically. There were anomalies in my research, which were when only 1, 3 and 2 shallowly and 7, 6 and 7 deeply processed words were respectively recalled by participants (see Reliability section for more information).
Graph
A Graph to show the amount of deeply processed VS shallowly processed words by each participant.
Graph/Descriptive Statistics Commentary
The graph clearly shows that many more deeply processed words can be remember than those that are shallowly processed. The graph shows that nearly twice as many deeply processed words can be remembered when compared to those shallowly processed. If some of the anomalous results aren’t taken into the account and the line of best fit is drawn this can effectively be seen.
Relationship of Results to Hypothesis
The results support my hypothesis to a vast extent. I stated that I thought semantically processed words will be remembered better than structurally processed words. As stated earlier and shown on the graph, almost in all cases, twice as many deeply processed words are remembered compared to shallowly processed words the hypothesis should therefore be accepted.
Discussion
Validity
From an experimental perspective, my test had quite high validity. I managed to perform all the tasks in the same quiet, noise free environment which insured that all the participants would have an equal chance and opportunity to concentrate on the task at hand and answer the questions without any interruptions. One thing I was missing was the ecological validity. Because the experiment was not performed in the natural environment it had low ecological validity which could possibly affect how things are remembered. Validity could also have been affected by how much time people had. The pressure of someone being in the same room with them with the stop watch might have affected their memory in a way; a person might have being trying to concentrate on how fast he can do the exercise instead of how much he could remember and hence forget things faster resulting in anomalous results. Overall the experiment had good face validity as it tested what it was supposed to and came out with appropriate results
Improving Validity
In order to improve the validity I could conduct a field experiment instead. This way the participant would be caught in their natural environment and are more likely to reflect the natural ability of their memory.
Reliability
Some participants only remembered 1, 2 or 3 words which were processed structurally which brought the average down quite significantly, making it unreliable as a measure. This was also the case for the deeply processed words; as a matter of fact it was the same participants that had low
Improving Reliability
Implications of the Study
When Craik and Tulving conducted their experiment their results showed that the semantically (deeply) processed words were remembered the best and structurally (shallowly) processed words were remembered much less. This is similar to my results
Generalisation
Application
References
Angles On psychology – Page 46
http://www.psypress.co.uk/ek5/resources/demo_ch06-sc-03.asp
Appendices
List of words from task one coded (S) or (D)
NEAR (d)
grim (S)
canvas (s)
BOOT (D)
picture (D)
ring (S)
engine (D)
free (S)
pizza (D)
inside (d)
FOOD (s)
dress (s)
crease (s)
PAN (s)
RUSSIA (D)
frost (S)
GLASSES (D)
FIT (S)
bandage (S)
FREAK (S)
pen (s)
pig (D)
BULB (d)
ROPE (s)
CUTLERY (d)
spicy (D)
sugar (s)
chalk (d)
BROOM (d)
SALMON (d)
14 14