Psychology Research Methods
Key Terms
Aims
The purpose of the study.
Independent Variable
The variable that is manipulated.
Operationalisation
Defining variables
Control Variable
Something other than the independent variable influencing the dependent variable.
Hypothesis
A prediction
Dependent Variable
Null
Directional
Non-directional (Two-tailed)
The variable you measure.
Extraneous Variable
Anything other than the independent variable which can have an affect on the dependent variable.
Experimental Design
How participants are organised in the study.
Types
Independent Measures
Seperate group of participants per condition of the study.
Strengths
Weaknesses
No order effects
No fatigue
2x as many people
Individual differences
Repeated Measures
Same group of participants for every condition of the study.
Strengths
Weaknesses
No individual differences
Fewer people
Can be counterbalanced
Order effects
Demand characteristics
Fatigue
Matched Pairs
Seperate groups per condition of the study, whilst being matched by certain characteristics (e.g age, sex, IQ.)
Strengths
Weaknesses
Controls order effects
Reduces individual difference
Takes more time and effort.
Still a chance of individual differences.
Differences in personality.
Tested participants against themselves.
Sampling
How participants are selected into your study.
Types
Random
All members of a target population have a equal chance of being selected.
Strengths
Weaknesses
Most representative
Not fully representative
Time consuming
Miss people
People may refuse to take part.
Opportunity
Using participants that are most readily available.
Strengths
Weaknesses
Easy
Less expensive
Less time-consuming
Very unrepresentative
There is clear sample bias.
Volunteer
Study looks for participants through advertisements, whatever participants reply will be included in the study.
Strengths
Weaknesses
Target populations
Bias of volunteers means a bias of of the sample.
Systematic
Selecting every nth member of a target population.
Strengths
No chance of researcher bias.
Weaknesses
Still could be unrepresentative.
Stratified
Strengths
Weaknesses
Using groups in representation of the target population, if the target population has 70% of women then the sample will also have 70% women.
Avoids researcher bias
Very representative
Sample doesn't reflect the way people can differ.
Could end up with a fluke, accidentally selecting all male participants.
Putting advertisements to access to those you wish to target.
Experiments
Types
Lab
The independent variable is manipulated by the experimenter.
Strengths
Weaknesses
Control
Internal validity
Artificial
External validity
Field
In a natural environment, independent variable is manipulated by the experimenter.
Strengths
Weaknesses
Improved external validity
Loss of control
Poorer internal validity
Natural
Independent variable is naturally occurring.
Strengths
Weaknesses
High external validity
Allows studies to be made on things hard to control.
No random allocation
Poor reliability
Quasi
Independent variable is fixed.
Strengths
Weaknesses
Can take place in controlled conditions.
No random allocation
Guarantee that the IV is the only thing affecting the DV.
Isn't likely to occur the same way in the real-world.
This means there could be confounding variables.
Cannot be replicated
This means there could be confounding variables.
Data
Information collected by the end of an experiment.
Types
Primary
Retrieved first-hand, collected personally.
Strength
Weakness
Specific to what you are intending to study (your aim.)
Harder to obtain
Costly
Time-consuming
Secondary
Retrieved from pre-existing sources.
Strength
Weakness
Pre-existing
Easier to gather
Not specific to your aim.
Qualitative
Non-numerical data.
Strengths
Rich and detailed
Weaknesses
Harder to compare
Quantitative
Numerical data
Strength
Weakness
Easy to compare
Lacks detail
Mean
Add and divide by the total.
Strength
Weakness
Looks at all scores
Affected by extreme scores.
Median
Middle number
Strength
Unaffected by outliers
Weakness
Does not take in all scores to account.
Mode
Most common number
Strength
Weakness
Unaffected by outliers
Does not take all scores into account.
Harder to replicate
Easier to replicate
Range
Largest number subtract the smallest number.
Strength
Weakness
Easy to calculate
Influenced by extreme scores
Standard deviation
The average distance from the mean.
Strength
Weakness
Takes all scores into account.
If it is a larger standard deviation = the scores are more spread out.
Harder to calculate
Peer review
Research checked by other experts.
Strength
Weakness
Checks the quality of research.
Allows for amendments to be made (changes to study to improve it.)
Allocation of funding
More likely to get research funding in the future.
Anonymity
Publication bias
New research
Experts could say that your theory goes against theirs, pushing their own agenda on the quality of your research.
Significant research is more likely to get published.
Measures of Central Tendency and Dispersion
Correlations
Single-blind
Double-blind
When details are kept away from participants in an effort to reduce demand characteristics.
Participants may not be told the aim of the study, the condition they are in or even if there is another condition.
When the aim of the study is not revealed to the researcher or the participants.
Often an independent party will be used to observe participant's behaviours.
A demonstration of the strength and direction of association between two co-variables.
Co-variables = variables investigated in correlations.
Positive correlations
Negative correlations
Strength
As one co-variable increases, the other increases.
As one co-variable increases, the other decreases.
No correlation
Co-variables do not share any relation with each other, they are random.
If a correlation is strong, it can be said that the relationship between the co-variables is more obvious.
A strong correlation plotted on a scattergram would display plots as being closer together in a line format.
Descriptive statistics - the use of graphs, tables and summary statistics to identify trends and analyse sets of data.
Measures of central tendency - the general term for any measure of an average in a set of data.
Measures of central tendency
Measures of dispersion
Median - the central value in a set of data when values are arranged from lowest to highest.
Mean - the arithmetic average calculated by adding up all the values in a set of data and dividing them by the number of values.
Mode - the most frequently occurring value in a set of data.
If there isn't an exact middle value, the two values that are closest to the middle are split to make a new value.
Standard deviation
Value that tells us how far a score deviates from the mean.
A larger deviation means the data is much more spread out and a smaller deviation means the data is more clustered.
Range - a value that calculates the spread of scores in a data set, the highest value is subtracted by the lowest value.
Presentation of quantitative data
Tables
Bar Charts
Histograms
Scattergrams
Tables are used to summarise the descriptive statistics found within a study, they are no longer raw scores.
Bar charts are used when data can be placed into categories, known as discrete data. This can be helpful when presenting data from seperate conditions of an experiment.
Shows the x-axis as continuous, rather than discrete. The x-axis shows equal-sized intervals (gaps between numbers) of a single category.
e.g a maths test being broken down into intervals of 0-9, 9-18, 18-27.
Used to show associations between co-variables,
Distributions
Normal distributions
Skewed distributions
Follows a bell-shaped curve.
Do not form a balanced, symmetrical pattern (no bell-shaped curve.)
Most scores will land in the middle of the data set, being relatively symmetrical towards the edges.
Scientific method
Central tendency - the middle of a data's distribution.
Aims
Probability
Inferential statistics - statistics that determine whether a finding in a study is significant.
Nominal
Simplest level of measurement.
Data is arranged into categories, e.g people placed into 'tall' or 'short' categories.
Ordinal
Slightly more sophisticated level of measurement.
Data is arranged in a certain order, e.g people placed in ascending order in terms of height.
Interval
Most sophisticated level of measurement.
Data is measured according to a scale, with use of fixed intervals, e.g people's height being measured by feet.
Absolute certainty - is indefinitely going to happen.
Allow us to determine whether our results occurred due to chance or the independent variable we manipulated.
Represented as 1
Not an absolute certainty - is indefinitely not going to happen.
Represented as 1
The probability level - the probability that we are in error.
Represented as p
In psychology we take our results to have a 95% chance of being significant, showing a significant difference.
In medical fields a 99.99% chance is taken.
The p level would be classed as: p ≤ 0.05 (with 0.05 being the decimal equivalent to 5%.)
Ethics in Psychology
British Psychological Society (BPS)
Responsible for the promotion of excellence and ethical practice within psychological studies (practices.)
Ethical guidelines
Confidentiality - researchers must guarantee anonymity of participants.
Informed consent - participants must give informed consent, where participants are aware of the true aims of the study.
Deception - sometimes researchers may choose to lie to participants about the true aims of the study.
The Right to Withdraw - participants have the freedom to leave a study at any point.
Protection from physical and psychological ahrm
Privacy - participant's must have their privacy protected.
Determine the purpose of your study, does not state a prediction.
Hypothesis
A statement that predicts whether a significant difference will be made and if so, what it will be.
Directional (One-tailed)
Suggests which direction the results will go in.
Non-directional (Two-tailed)
Suggests that there will be a significant difference, but with no clear direction on where the results will go.
Alternative
States that there will be a significant difference, one variable will effect the other.
Null
States that there wasn't a significant difference, the variables do not affect each other.
This is accepted if your results show no significant difference.
Empirical evidence
Falsifiability
Replicability
Control
Objectivity
click to edit