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

Guide to the Practical Investigation

Learning objectives

In this section of biological psychology, you will develop skills to:

• Plan and carry out a correlational study linked to aggression or attitudes towards drug taking
• Examine correlational data by performing an inferential test (Spearman's rho)
• Interpret significance levels and their meaning
• Generate descriptive statistics from research data
• Compose an abstract summarising the investigation
• Construct a discussion section analysing the investigation's conclusions

Aim of correlational research

Correlational research examines the relationship between variables. It determines both the strength of the relationship and whether it is positive or negative. In biological psychology, you can explore connections between biological variables (such as brain structure or hormone levels) and behavioural variables (such as aggression or drug-taking attitudes).

All research begins with a research question. For example, you might ask: "Is there a link between the masculinity of a person and aggressive behaviour?" This question guides the development of your hypotheses and methodology.

Hypotheses and variables

Alternative hypothesis

The alternative hypothesis predicts that there will be a relationship between the variables. This is called a directional hypothesis when it specifies the nature of the relationship. For instance: "There will be a positive correlation between masculinity and aggression." This predicts not just that a relationship exists, but specifically that as one variable increases, so does the other.

The null hypothesis states that there will be no relationship between the variables. Any observed relationship would be attributed to chance rather than a genuine connection. For example: "There will be no relationship between masculinity and aggression, and any relationship found will be due to chance."

Operationalisation of variables

Once you have formulated your hypotheses, you must operationalise both co-variables. This means defining precisely how each variable will be measured. For example, masculinity could be measured using a brain sex quiz that generates a masculinity score, whilst aggression could be measured by observing participants playing a moderately violent video game and counting the number of characters they knock down (with each receiving 1 point).

The refined hypothesis then becomes: "There will be a positive correlation between a participant's masculinity score on a brain sex quiz and their scores on a moderately violent video game."

Design

Correlational designs involve measuring two variables from the same participants and then calculating the precise nature of the relationship between them. In the example study, you would first measure the masculinity of participants, and then for the same individuals, measure their aggressiveness. This paired data allows you to examine whether changes in one variable correspond with changes in the other.

Controls

Controlling extraneous variables is essential to ensure that your measures are not influenced by factors other than the variables of interest. When testing both brain sex score and aggressive acts in a video game, several controls must be implemented:

Standardisation of testing conditions: All participants should complete tests in the same environment. For the brain sex measure, everyone should be tested identically with scores calculated using the same objective measurement system. A computerised test that generates a masculinity score would provide consistent, objective measurements.

Equalising participant experience: When measuring aggression through video game performance, all participants must be equally proficient at the game. This might require creating a bespoke game to ensure no participant has prior exposure. Everyone should receive the same amount of practice time to control for skill differences.

Controlling environmental variables: Participants should engage with the game under identical circumstances. This eliminates extraneous variables such as distractions, time of day effects, or variations in supervision that could affect performance.

Finding a sample

Once you have prepared the materials for your study, you need to recruit participants. The sample should be representative of the target population to which you wish to apply your findings. In studies of male aggression, the target population would be the male general public.

At recruitment, participants must be informed about the study's aims and nature so they can make an informed decision about participation. Several sampling methods are available:

Opportunity sampling involves recruiting people who are readily available. This method is quick and convenient but may introduce bias as participants might share characteristics (such as being from the same location or social group).

Volunteer sampling involves advertising the study in locations such as workplace noticeboards, college noticeboards, or online platforms. Interested individuals then come forward. Whilst this can attract motivated participants willing to commit time, it may also compromise representativeness as volunteers might have particular personality traits or interests.

Ethics

Addressing ethical issues before participants begin the study is paramount.

Informed consent

Participants must be clearly informed about:

• The nature of the study and what you are investigating
• Exactly what they will be required to do, in sufficient detail for them to decide whether to participate
• Any possible implications of participation (in this example, the implications of playing an aggressive video game)

Right to withdraw

Participants must be notified that they are free to withdraw at any point without consequence. They should also be informed that they can take their data with them if they choose to leave.

Confidentiality and anonymity

Data must be kept confidential so that information about individual participants is not shared with others. Data should be made anonymous by replacing names with participant numbers. All data should be destroyed within a reasonable timeframe after the study concludes.

Age restrictions

All participants must be able to give informed consent. No one under the age of 16 should be included in the study. For this investigation, it is particularly important that all participants are over 16 because they will be asked to play a video game involving aggression. The game must be age appropriate, and all participants must be informed in advance that they will play this type of game.

Consent form components

A properly designed consent form should include:

• The aim(s) of the study
• Detailed description of what participants will be expected to do
• Any possible implications for the participant
• A clear statement that the participant can withdraw at any point without consequence
• Reassurance that data will be anonymised and destroyed after a certain date
• Information about who will see the findings and how the information will be distributed
• Contact details of the researcher and supervising teacher/school/college
• Signatures of both the researcher and participant

Developing a procedure

Now you can plan how to conduct the study. Consider what materials you need to measure your variables and how you will obtain them. In the example, you would need to arrange for all participants to complete the brain sex test under supervision to ensure proper concentration and attention. Think through practical details such as whether participants work on individual computers or in groups, and how much time researchers need to allocate.

As a researcher, you must develop a standardised procedure for gathering data. This procedure must be detailed enough to allow others to replicate your study and test whether they obtain the same results. Your written procedure should include:

• All instructions given to participants
• A step-by-step account of the study's execution

For instance, once participants have been recruited, you might schedule a time for them to attend the laboratory to establish their brain sex score. This could occur simultaneously for all participants or separately. Next, you need to determine how to measure aggression using the video game.

Pilot study

Before conducting the main study, it is advisable to run through the procedure using a pilot study. This preliminary run-through helps identify any procedural or ethical issues you may have overlooked in planning. Pilot testing allows you to:

• Test the video game and brain sex quiz
• Verify that instructions are understood by all participants
• Confirm that timings are appropriate
• Ensure the venue is suitable for the study
• Check whether your variables actually measure what you intend (whether the study is reliable and valid)

Example procedure from the investigation

The study took place during one afternoon between 2 pm and 4 pm. Following the initial briefing and after participants had given informed consent, ten volunteer participants (aged between 19 and 52 years old) were gathered in a lounge area at the university. Each participant was allocated a number.

Participants were shown into an IT room where each person was seated at a computer preloaded with a brain sex test. They were given as much time as needed to complete the test. Completion times ranged from 25 minutes (fastest) to 42 minutes (slowest). The test was completed in silence and supervised by the researcher, who recorded each participant's score against their allocated number. Scores were calculated automatically by the computer program.

After completing the test, each participant returned to the lounge area and waited until called. Refreshments were available during the wait. Participants were then shown individually or in pairs into one of two small laboratory rooms. Each participant sat alone at a laptop set up with a bespoke computer game requiring them to perform moderate acts of aggression to score points. Supervision was provided by a research assistant who had minimal interaction with participants other than to set up the game and input participant numbers.

The game included a 3-minute practice session followed by a 5-minute data-gathering session. All instructions appeared on screen. Participant scores were logged by their numbers. Once the game was complete, participants met with the main researcher in another small room and were debriefed individually or in pairs. They were reminded of their right to withdraw, any questions were answered, and they were offered the opportunity to see the results once available. Participants were thanked for their participation, asked whether they wished to offer any insights into their experience, and then released.

Analysing results

Analysing correlational data requires two scores from related sources. In this case, the source is each participant, and the two scores are their brain sex measures and their aggression score from the video game. This data must be at least ordinal (ranked data).

Raw data

The video game data provided an aggression score for participants ranging from 0 to 20, where a low score represented low aggression levels.

The brain sex test data was calculated by the computer program. A high score indicated high masculinity. The maximum possible score was 100 and the minimum was 25.

The data should initially be tabulated as follows:

Participant number	Gender	Brain sex score	Aggression score
1	M	56	17
2	M	69	18
3	M	30	10
4	M	88	10
5	M	80	15
6	M	87	16
7	M	95	17
8	M	67	15
9	M	77	14
10	M	45	6

Scatter diagrams

Because this study examines whether a relationship exists between two co-variables, there is no need to analyse overall group performance. Measures of central tendency (mean, median, mode) are therefore irrelevant when conducting a correlation. Bar charts and similar displays would serve no purpose.

However, it is necessary to create a scatter diagram (also called a scatterplot or scattergraph). This provides an easy visual display of any relationship and indicates whether a relationship exists and what type of relationship has been found.

The scatter diagram for this data shows a slight trend from bottom left to top right, suggesting a potential weak positive correlation. However, to accurately determine the strength of any relationship, you must conduct a statistical test. For correlational analysis, you need to perform Spearman's rho test.

Conducting Spearman's rho

If you have gathered interval or ratio data, the data will be reduced to ordinal level when ranked during the test procedure. The steps for calculating Spearman's rho are:

1. Rank the scores on one variable, giving 1 to the smallest score and continuing upward. When ranking scores, you may find it easier to keep track if you write out the ranks first (1-10) and then allocate them. This is especially helpful when you have tied data.

2. Do the same for the other group of variables.

3. Calculate the difference (D) between the ranks for each score.

4. Square each difference (D²).

5. Find the sum of the squared differences (ΣD²).

6. Count the number of participants (N).

7. Apply the Spearman's rho formula.

The Spearman's rho formula is:

$r_s = 1 - \frac{6\Sigma D^2}{N(N^2-1)}$

Where:

• $r_s$ = Spearman's rank order correlation coefficient
• $\Sigma D^2$ = sum of squared differences between ranks
• $N$ = number of participants

Finding the critical value

The final stage in statistical analysis is to find the critical value appropriate to the data. This requires using a critical values table for Spearman's rho.

The study had a directional (one-tailed) hypothesis predicting a positive relationship between masculinity of brain sex and aggression scores. Therefore, a one-tailed test will be used. The minimum level of probability acceptable in psychological research is the 0.05 level (5% level). There were ten participants ($N = 10$).

Using this information, the observed (calculated) value of Spearman's rho of 0.318 is compared with the critical value from the table. At $N = 10$ and $p = 0.05$ for a one-tailed test, the critical value is 0.564.

Interpreting critical values

For the result to be considered statistically significant, the observed value must be equal to or greater than the critical value from the table. In this case, 0.318 is not equal to or greater than 0.564, so the result is not significant.

Understanding the relationship between variables

If variables are positively related, participants who score low on one measure should also score low on the other, and similarly for high scores. If variables are negatively related, those who score low on one variable should score high on the other.

When variables are ranked separately, high ranks for one variable should correspond with high ranks on the other if there is a positive correlation. If there is no correlation, then rank distributions are unrelated to one another – rank of score on one condition is unrelated to rank on the other.

Presenting findings

You must state whether the null hypothesis should be accepted or rejected based on the data analysis. Researchers must provide sufficient data to allow readers to understand the decision.

Raw data should be available, though typically only as an appendix to the final report. However, descriptive statistics including tables and graphs must be included at the point of presenting findings in the results section.

The conclusion of the inferential statistical analysis must be stated in the following format:

Drawing conclusions

Once data has been gathered and analysed, you will be able to state whether you reject or accept the null hypothesis.

However, you cannot be reasonably certain that a genuine relationship existed in this study because of the inferential test results. This does not mean that other factors did not influence the data and therefore challenge the validity of any conclusions drawn.

When considering the findings of any investigation, the researcher must examine the reliability and validity of the methods and procedures used to gather data.

Validity

Validity refers to whether the variables truly measured the concepts being tested. In this case, was the brain sex test a true measure of masculinity and the video game a true measure of aggression?

The brain sex test could challenge this measure as a valid reflection of masculinity because it is very superficial and ignores many other factors that influence masculinity, such as social roles and norms of behaviour.

The computer game could be challenged as lacking ecological validity in the way it measured aggression. Sitting at a screen and manipulating virtual characters for points does not necessarily translate into real-world aggressive tendencies, which are likely to be constrained by social regulations.

It could be argued that the number of times a character was hit in the game was merely a reflection of the video game's rules and not an indication of aggression.

Reliability

Reliability concerns whether the procedures provided a consistent test of aggression and masculinity. The use of standardised procedures and objective measures in the example practical investigation increases the reliability of the data. It is realistic to expect that the procedure could be replicated and that another researcher could consistently record and interpret the data on both measures in the same way as in this study.

However, challenges to reliability remain. For some participants, completing the brain sex test first might have alerted them to the study's goals, which may have affected their performance on the video game. Some students also had longer gaps between the brain sex test and the video game. This might have relaxed them, meaning they did not operate at the same level of focused attention as others who completed tests consecutively.

Although the study ensured no participant had prior knowledge of the game and all participants had the same practice time, some participants might have been experienced gamers with higher generic skills. This could have led them to score more points, meaning we measured gaming performance rather than consistently measuring aggression.

Generalisability

In the example study, the sample size was small at only ten people. Although a range of ages was tested, it might be that the sample is not representative. This could be especially true because it was a volunteer sample, so those who came forward might represent only a certain type of person. Volunteers tend to have more compliant personalities, so perhaps could have altered their behaviour to meet the researcher's expectations.

The amount and type of aggression displayed in different cultures and subcultures vary considerably. Some cultures nurture aggression whilst others actively prohibit it. Because this example practical investigation is based on a sample of participants from an industrialised Western culture, the study can be regarded as ethnocentric and the findings will not apply to other cultures. The study is also limited to explaining male behaviour, so cannot be applied to explain female aggression.

Writing the abstract

Once the research is complete and has been written up in the appropriate format, the final task is to write the abstract. This is a brief summary of the aims, procedure, result and conclusions drawn from the study. It is designed to allow others to quickly assess whether the study is appropriate for their needs and whether to continue reading or purchase the entire study.

Writing the discussion

The discussion section of a report should be where conclusions about the data are drawn from the data analysis. This is where findings are explained within a wider context of background theories and previous research in the area. It is a section where researchers suggest limitations and strengths of the investigation and possible improvements that could be made to the methodology used.

A discussion section also offers practical uses of the findings and potential implications for knowledge within a wider context such as organisations, education, clinical practice or society as a whole.

When writing the discussion section, you should link your findings to theories and research in biological psychology, evaluate the strengths and weaknesses of your methodology, suggest improvements for future research, and consider the practical applications of your findings.