Practical Investigation (Edexcel A-Level Psychology): Revision Notes

Guide to the Practical Investigation

Overview of the practical investigation

The practical investigation requires designing and conducting a laboratory experiment to gather quantitative data on a topic within cognitive psychology. Students must demonstrate competence across multiple areas of research methodology.

Learning objectives

To successfully complete the practical investigation, you must:

• Design and conduct a laboratory experiment to gather quantitative data on a cognitive psychology topic
• Make informed design decisions during planning
• Collect and present data using appropriate tables, graphs, descriptive statistics and non-parametric tests of difference
• Consider strengths and weaknesses of the experiment and suggest possible improvements
• Use typical reporting conventions to document procedure, results and discussion

Aim

All research begins with an aim, typically based on current theory or research in an area. The aim establishes the purpose and focus of the investigation.

For this experiment, the aim is to investigate whether acoustic similarity affects short-term memory. This is based on theory suggesting the short-term store uses acoustic encoding, meaning similar sounding words and letters are more difficult to sub-vocalise and encode, resulting in poorer recall performance. Reading around the topic before planning helps establish a rationale for your own aim.

Hypotheses

Experimental hypothesis

Your practical investigation requires both an experimental hypothesis and a null hypothesis. The experimental hypothesis may be directional or non-directional.

Before writing a hypothesis, clearly define and operationalise the independent variable (IV) and dependent variable (DV). This practical investigation examines whether acoustically similar or dissimilar sounding words (IV) affect recall (DV).

For this practical investigation, the experimental hypothesis is directional because prior theory and research indicates more acoustically dissimilar sounding words will be recalled than acoustically similar sounding words.

Null hypothesis

The null hypothesis states there will be no difference between conditions, with any difference found being due to chance. For this investigation: "There will be no difference in the number of acoustically similar and dissimilar sounding words recalled, and any difference found will be due to chance."

Experimental design

When selecting an experimental design, consider the strengths and weaknesses of each approach.

Independent groups design

An independent groups design allocates different participants to each condition. This design is appropriate when you wish to avoid order effects and demand characteristics.

However, independent groups designs mean individual differences between participants in each group may affect results. A repeated measures design avoids individual differences but introduces order effects and demand characteristics.

Repeated measures design

In this experiment, the aim is to examine encoding in short-term memory, which should be reasonably similar between participants unless affected by age, illness or learning impairments. If the sample is fairly homogenous (similar), individual differences should not pose problems. However, order effects are likely if participants repeat the memory test immediately after one another, as they may guess the study aim.

On balance, an independent groups design should work well to avoid order effects and demand characteristics, whilst ensuring individual differences in memory between participants do not impact results significantly. To evenly distribute any individual differences in short-term memory span, participants can be randomly allocated to each experimental condition.

Sampling

Selecting an appropriate sampling method involves considering the target population and using a sampling technique that draws a representative sample. This allows confident generalisation of findings back to the target population. For this experiment, the target population is very large as it can involve any individual with reasonably intact and unimpaired short-term memory. Using a random or stratified sampling technique would be difficult and time-consuming, so either a volunteer or opportunity sample would be more efficient.

The volunteer sampling technique encourages participants with a particularly compliant nature. This would not pose problems for social psychological research, but as cognitive processes are relatively unaffected by personality type and an independent groups design is being used, volunteer sampling is probably more ethical than an opportunity sample, as no one is directly asked and pressured to participate.

Operationalisation

The independent variable in this experiment is whether words presented to participants are acoustically similar or dissimilar in sound. This will be operationalised by presenting one group with ten monosyllabic words that rhyme (acoustically similar) and ten monosyllabic words that do not rhyme (acoustically dissimilar). The dependent variable will be the total number of accurately recalled words from the original list in a free recall memory test. It is essential to operationally define both independent and dependent variables clearly.

Psychology as a science

Operationalisation is essential to establish objectivity and reliability in psychological research. In this experiment, operationalisation has been achieved by defining the type of words each group will receive and exactly what will be measured as an outcome. With proper control over extraneous variables, it becomes possible to establish a cause and effect relationship between the IV and DV in this experiment.

Controls

Situational variables

Because you are conducting a laboratory experiment, the environment will be reasonably well controlled already. However, consider any situational variables likely to affect participants. In this experiment, it would be important to control for any noise or interruptions that might affect learning and recall, so participants will be tested individually with a sign placed on the door to prevent interruptions. To prevent conferring and therefore the potential for demand characteristics, all participants will be placed in a room, called out individually to do the memory test, and told not to return. The procedure will also be standardised to prevent experimenter effects occurring.

Participant variables

Although cognition is reasonably similar in the majority of people, individual differences are not likely to have a huge impact on your investigation. However, consider whether there are any individual differences such as age and memory impairments. Once you have identified potential individual differences between participants, you will need to either control them or eliminate them.

There will be participant variables not controlled for, such as motivation. However, the experiment will be conducted in the morning to prevent any possible effects of fatigue that might occur later in the day.

Developing a procedure

A procedure should be a schedule of what happens, where, when and how. This ties into how you will control for extraneous variables and it is important that the procedure stays the same for all participants.

In this experiment, participants will be read a set of standardised instructions by the researcher and asked to sit at a desk directly in front of a whiteboard and projector. To prevent demand characteristics, participants will be told what they will be required to do but not why they are doing it or what other participants will be doing. The words will be presented one at a time for a duration of 3 seconds per word. Participants will be given a pen and paper to recall in any order the words they have learned immediately after the words have been presented. They will be given 1 minute to recall the word list. This standardised procedure will mean that all participants are treated in exactly the same way and will minimise any experimenter effects.

Apparatus

Any research into memory typically involves participants learning something, whether it is a list of words, letters or digits, a set of images, or a simulated event. Your apparatus will depend on your aim, but it is worth considering the nature of the apparatus you ask participants to remember. A word list may seem fairly straightforward, but you need to remember that some words are easier to remember or more memorable than other words. It may be worth considering using a list of high frequency words, words of similar syllable length (particularly for short-term memory research) or using letters or digits instead.

Table of apparatus: word lists

Acoustically similar sounding words	Acoustically dissimilar sounding words
Cat	Pan
Mat	Ten
Hat	Mill
Lot	Hit
Hot	Gun
Dot	Dog
Cot	Get
Den	Hot
Pen	Bun
Hen	Pit

Ethical issues

Before undertaking any psychological research, it is essential to consider the ethical implications. Both the British Psychological Society (BPS) and British Education Research Association (BERA) guidelines should be consulted and adhered to as closely as possible, even for a small A-level practical investigation. Any research can make participants feel pressured, intimidated, embarrassed or concerned. Ethical issues require careful thought before proceeding with your experiment. If you are unsure whether your experiment will present ethical issues, you may wish to conduct a pilot study on family and friends first and ask them how they felt during the experiment.

Valid consent

In this experiment, participants have been asked to volunteer for an experiment into short-term memory. This is clearly stated on the recruitment advertisement. However, participants will not be entirely aware of the full aim of the experiment until after they have completed the memory test. This means that, although consent to take part has been given, fully informed consent has not been achieved because the true nature of the experiment has been partially withheld.

Because participants being recruited for this experiment are under the age of 18 years, it is necessary to gain consent from a parent or guardian of the child. In this experiment, details of parents/carers were gained from the volunteers and a consent form was sent out to parents with information about the experiment. You will also need to consider whom consent needs to be gained from if your participants are considered vulnerable children. You should provide an information sheet for both parents/guardians and participants setting out the nature of your experiment.

Information sheet contents

The information sheet should include:

• The aim of the study
• The type of data to be collected
• The method of data collection
• Confidentiality and anonymity conditions
• Compliance with the Data Protection Act and Freedom of Information Act, how the results will be made available and details on destruction of data
• The time commitment participants should expect
• A right to decline or withdraw from the study without consequence
• The possible risk to the participants
• The nature of debriefing
• How the data will be used and the benefits of the study
• The name and contact details of the researcher and supervisor

(Adapted from the BPS Code of Conduct 2010)

Right to withdraw

In this experiment, participants were offered a right to withdraw in the recruitment advertisement, the standardised instructions and debrief.

Risk

It is important to consider whether your participants will be protected from harm. Harm can be physical or psychological. Under no circumstances should you physically harm your participants, and you will have to think very carefully about whether they will suffer any psychological harm, even modest harm such as embarrassment or stress. In this experiment participants are reminded of their right to withdraw, the results were anonymised, and they were told that the test is not a measure of intelligence. It is also important that participants are given an opportunity to ask any questions they may have arising from the research. This can help alleviate any anxiety before the test and any embarrassment caused by the test.

Analysing the results

Gather together the results from the participants and present them in a raw data table, like the one in Table 2.14.

Table 2.14: Raw data table showing recall of acoustically similar and dissimilar sounding words

Participant number	Total number of similar sounding words recalled	Participant number	Total number of dissimilar sounding words recalled
1	4	11	9
2	5	12	7
3	3	13	8
4	7	14	6
5	9	15	8
6	3	16	10
7	7	17	9
8	8	18	5
9	5	19	10
10	4	20	9

This raw data indicates that the highest number of words was recalled when the words presented were acoustically dissimilar and the lowest number recalled was from the acoustically similar list. The raw data seems to support the experimental hypothesis. However, raw data can be difficult to interpret and represents individual scores. To help interpret the findings, descriptive statistics can be used to generate a summary of the average score achieved in a data set. Measures of central tendency, such as mean, median and mode, and measures of dispersion, such as range and standard deviation, should be presented in a summary table, as shown in Table 2.15.

Table 2.15: Summary table showing typical recall score and distribution of scores between acoustically similar and dissimilar sounding words

	Acoustically similar sounding words	Acoustically dissimilar sounding words
Median of words recalled	5	8.5
Mode of words recalled	-	9
Mean number of words recalled	5.5	8.1
Range of words recalled	6	5
Standard deviation	2	1.6

The mode score for acoustically similar sounding words were 3, 4, 5 and 7, so is not a useful modal score to present. However, overall, the measures of central tendency suggest that more acoustically dissimilar words were recalled than similar sounding words. The measures of dispersion suggest that there was greater spread of results for the acoustically similar sounding words than the dissimilar sounding words, but this was only slightly greater. It perhaps suggests that some participants found it slightly easier or more difficult to recall the words than others within the acoustically similar word group.

Graphical representation

The mean of words recalled seems to reflect the typical score achieved by participants in both conditions of the experiment, so this statistic can be graphically represented in a bar chart.

Drawing conclusions

From the data gathered and presented, it is important that you can draw conclusions from your findings. This can involve going beyond the findings and relating your data to the concepts under investigation. This practical investigation found that participants recalled fewer acoustically similar sounding words compared to acoustically dissimilar sounding words. The typical recall score achieved by participants given acoustically similar sounding words was, on average, three words fewer than for dissimilar sounding words. However, the distribution of scores suggests that there was a large degree of individual variation in recall for both groups, which was marginally greater for the acoustically similar words group. This demonstrates that some individuals found word recall easier and some more difficult than others.

Inferential test of significance

To determine whether the findings of the practical investigation are statistically significant, or just due to chance, you will have to run your data through an inferential test. For your practical investigation you will need to gather quantitative data that is at ordinal level or above, and therefore conduct either a Mann-Whitney U or Wilcoxon non-parametric test of difference. If you have used an independent groups design you will use a Mann-Whitney U test, or if you have used a repeated measures design you will use a Wilcoxon test.

This practical investigation used an independent groups design, so a Mann-Whitney U test was run on the data.

Mann-Whitney U test formulae

$U_A = n_A n_B + \frac{n_A(n_A + 1)}{2} - \sum R_A$

$U_B = n_A n_B + \frac{n_B(n_B + 1)}{2} - \sum R_B$

(U is the smaller of $U_A$ and $U_B$)

Table 2.16: Scores for each group: words recalled

Total number of similar sounding words recalled Group A	Rank	Total number of dissimilar sounding words recalled Group B	Rank
4	3.5	9	16.5
5	6	7	10
3	1.5	8	13
7	10	6	8
9	16.5	8	13
3	1.5	10	19.5
7	10	9	16.5
8	13	5	6
5	6	10	19.5
4	3.5	9	16.5
Sum total of points	89.5	Sum total of points	138.5

This should be compared to a table of critical values for a Mann-Whitney U test.

Table 2.17: Critical values of U for a one-tailed test at 0.05; two-tailed test at 0.1 for a Mann-Whitney U test

[The full critical values table is shown in the image with values for n ranging from 1-10]

The calculated (observed) Mann-Whitney U-value is 16.5. This is less than the table (critical) value at $p < 0.05$ of 27, with N = 10, 10. Therefore, the result is significant at $p < 0.05$ for a one-tailed test. Therefore the experimental hypothesis can be supported. In fact the calculated (observed) value is less than the critical value for a one-tailed test at $p < 0.01$, meaning that the result is highly significant and therefore the difference between the recall of the groups is unlikely to be due to chance. This means that the likelihood of making a Type 1 error is reduced and the experimental hypothesis can be supported with confidence.

Making a statistical statement

Following your statistical test, it is important to make a statistical statement and support or reject your hypotheses. Your statistical statement should include the following information:

• The statistical test used
• The observed/calculated value
• Whether a one- or two-tailed test was used
• The level of probability
• The number of participants
• The critical value used
• Whether the calculated value was greater than/equal to/less than the critical value
• Whether the result is significant or not
• Which hypothesis it supported

Discussion

A discussion section of a report will include a summary of the findings and how these relate to the wider concepts, theory and prior research related to the study. It will also include the strengths and weaknesses of the practical investigation and suggestions for possible improvements and new directions for the research.

In this practical investigation it can be concluded that there was a significant difference between recall of acoustically similar sounding and dissimilar sounding words; acoustically dissimilar sounding words were significantly less well recalled. This finding is consistent with theory which suggests that encoding in the short-term store is primarily acoustic, making similar sounding words difficult to sub-vocalise (rehearse) and maintain compared to dissimilar sounding material. This is also consistent with research conducted by Sperling (1963) who found that participants had difficulty remembering acoustically similar letters (B, D, T) compared to acoustically dissimilar sounding letters (F, L, M, X).

Critical evaluation

It is important to consider both the strengths and weaknesses of the practical investigation. For this, you will be better placed to judge the successes and failures with regard to your own procedure and outcomes. However, a number of general questions should be asked:

• Is the study ethical?
• Can you generalise the findings to others/different cultures/different eras?
• Is the study reliable?
• Is the study carried out in a natural or artificial environment?
• Is the task ordinary?
• Are the findings useful in real life?
• Is the research valid?
• Is there any conflicting evidence from other research?

Is the task ordinary?

This is a common question asked of much cognitive research conducted in laboratories using series of stimulus lists of words, letters and digits, seemingly unrelated to memory tasks that occur in ordinary everyday life. It is true that we are often required to draw on our memory for shopping lists, telephone numbers or random sequences. However, this is not a typical activity for memory to undertake and often takes a degree of conscious processing and effort. For this reason, the task used in this practical investigation may be criticised for not reflecting an ordinary use of memory.

Suggestions for improvement

In addition to considering the strengths and weaknesses of your practical investigation, you should also refer to how your study could be improved. To do this effectively you will need to consider the weaknesses of your research and how these weaknesses could be overcome. Suggestions for improvement can be ambitious, but should not be impractical or impossible to achieve.

Writing up the report

Psychological investigations are written following a set of conventions for report writing as shown below.

Conventions of report writing

Abstract: A summary of the background theory/research, aims, hypotheses, method, results and discussion. This is a short paragraph overview of the entire report.

Introduction: An overview of related theories and research in the topic area. The introduction provides a rationale for the current investigation that links prior research to the study aims and hypotheses.

Method: A detailed account of the participants, sampling method, apparatus, procedure, controls and ethical issues.

Results: A detailed account of the data gathered and its analysis using descriptive and inferential statistics.

Discussion: Conclusions drawn from the results analysis, reference to prior research, strengths, weaknesses and possible improvements for future research.

Remember!