Research Methods Flashcards
Alternate/ Experimental Hypothesis
A clear, precise, testable statement that includes named variables which are operationalised and includes and prediction.
Null Hypothesis
A statement of coincidence chance.
Direction/ One-Tailed Hypothesis
Predicts the direction in white change is expected to occur, using previous research to suggest what direction the change may be while using words such as faster/ slower, bigger/ smaller and more/ less.
Non-directional/ Two-Tailed Hypothesis
Predicts change but does not specifiy direction. This is used when there is no previous research, is non-specific and uses words like effect, change and difference.
Laboratory Experiment
An artificial experiment where all conditions are controlled.
Field Experiment
A more realistic experiment where the variables and conditions cannot be controlled.
Natural Experiment
An experiment into the effects of things already happened.
Quasi Experiment
Investigation into a naturally occurring independent variable e.g. age or autism.
Laboratory Experiment Evaluation
+ Complete control of variables.
+ Forces pace of research.
+ Reliable
+ Quantitative Data
+ Able to use technical equipment
- Lacks ecological validity.
- Demand characteristics
- Sampling bias
- Ethics
Field Experiment Evaluation
+ Ecologically valid
+ Reduce demand characteristics
- Cannot control variables as easily
- Unable to control research pace
- Harder to replicated
- Unable to use technical equipment
Natural Experiment Evaluation
+ High ecological validity
- No control over variable
- Ethically wrong- might help more people
Quasi Experiment Evaluation
+ Reduction of demand characteristics
+ Lack of direct intervention
- Loss of control over variables
Demand Characteristics
A subtle queue that makes participants aware what the experimenter expects, influencing their behaviour.
Social Desirability
The participant changes their behaviour or answers to look good and be deemed social acceptable.
Investigator Effect
The researcher influences the outcome of the research, possibly due to their body language, or their own expectations.
Situational Variables
Outside influences on the experiment affecting the results.
Order Effect
Influences like practice, boredom and fatigue that often occur when a task is undertaken more than once.
Counterbalancing- order effect
Used in repeated measures design- changes the order of the tasks to reduce order effect.
Randomisation- order effect
Randomising the order of tasks/ presentation of data.
Individual Differences
Differences between participants such as intelligence, age, gender, social class, fitness, etc.
Repeated Measures Design
Involves the same participant in each condition.
Repeated Measures Design Evaluation
+ Control for individual differences
+ Need fewer participants
- Generates order effect
- Vulnerability to low validity
Independent Measures Design
Uses different participants in each condition
Independent Measures Design Evaluation
+ High validity
+ Control for order effect
- Reduced control for individual differences
- More participants
Matched Pairs Design
Different participants with similar characteristics in each condition.
Matched Pairs Design Evaluation
+ Controls for individual differences if identical twins.
+ Order effect
- Harder to find participants
- Expensive
Naturalistic Observation
People are studied in their natural environment. Researcher doesn’t interfere in any way and are often looking for a certain behaviour.
Controlled Observation
The environment is artificial and completely controlled - like an experiment.
Structured Observation
Researcher decides in advance what sort of behaviour they are looking for and devises a checklist. This generates quantitative data (numbers).
Unstructured Observation
Researcher records everything that happens collecting qualitative or quantitative data.
Event Sampling
Keeping count of when behaviours occur.
Time Sampling
The observer decides on a time interval and notes any behaviours that are displayed during that time.
Inter-rater Reliability
Using multiple observers to increase the reliability- must be trained in the same way and achieve an agreement level of 80%.
Covert Observation
Participants are unaware they are being observed.
Overt Observation
Participants are aware they are being observed.
Participant Observation
The researcher is part of the action they are observing. Can lead to observer bias.
Questionnaire
Self-reported paper questions.
Interview
Oral questions.
Open question
All participants to elaborate on their answers and give more detail.
Closed Questions
Force participants to choose an option.
Contingent Questions
Questions depend on answer to preceding question.
Likert Scaling
Generates quantitative data by assessing strength of opinion. At risk of invalid data if people choose middle answer.
Overuse of Jargon
Using technical/ complicated language.
Emotive Language
Influencing a viewpoint with emotion.
Leading Questions
Prompting a desired answer.
Double-barrelled questions
Confusing two part questions.
Double Negatives
Confusing/ complicated
Interviews Evaluation
+ Reliable
+ Ethical
+ Quantitative and qualitative
+ Quick and easy
+ Lots of questions
- Researcher/ volunteer bias- lack validity
- DC & SD
- Order effect
- Leading questions = bias
- Difficult to control sample
Correlations
Look for relationships between two variables.
Positive Correlation
Both variables move in same direction.
Negative Correlation
Variables move in opposite directions.
Correlation Co-efficient
Strength of a relationship from -1 to +1.
Correlation Evaluation
+ Good starting point
+ Quick, easy and cheap.
- No cause of the relationship between variables.
- Validity can be affected by how variables measured.
Random Sampling
Equal chance of being selected by a random selection tool.
Random Sampling Evaluation
+ Unbiased
+ Fair
- Representation not guaranteed
- People may not consent
- Can be time consuming and expensive
- May be difficult to obtain list.
Stratified Sampling
Target population split into sub sets e.g. age, gender, class, etc and a random sample is then taken.
Stratified Sampling Evaluation
+ Representative
+ More generalisable
- May be biased.
- Time consuming/ expensive
- Doesn’t always reflect the way people are different.
Opportunity Sampling
Selecting anyone available at the time.
Opportunity Sampling Evaluation
+ Convenient
+ Saves time and money
- Not representative
- Vulnerable to bias
Self-Selected Sampling
Participants volunteer to be apart of the investigation
Self-Selected Sampling Evaluation
+ Consensual
+ Quick and easy
+ Participants engage
- Volunteer bias
- Less representative
- Participants often similar
Systematic Sampling
Selected from a list at fixed intervals.
Systematic Sampling Evaluation
+ Can be representative
- Somewhat bias
- Time consuming
- Expensive
- Random sampling more efficient
- May not consent
Snowball Sample
Access a hidden population by asking participants to recruit others.
Snowball Sample Evaluation
+ Access hidden populations
+ Representative
+ No researcher bias
- Consent
- May be unethical- exposing them to harm
- No information guaranteed
Reliability
Consistency of a test/ procedure
Validity
Accuracy of a test/ procedure
Face Validity
Does it seem okay? Starting point.
Predictive Validity
Predicts future performance.
Concurrent Validity
Compares two methods of testing to see if results are comparable.
Ecological Validity
Can the findings be generalised to other settings beyond the lab?
Temporal Validity
Are they relevant/ true today?
Primary Data Evaluation
+ Specific
- Researcher bias
- Time consuming
Secondary Data Evaluation
+ Cheap
+ Quick
+ No preparation
+ Ethical
- Unfocused
- Might not be any data
- Out of date
- Incomplete
Meta-analysis
Using primary and secondary data to come to a conclusion.
Triangulation
Using multiple data sources/ research methods to improve reliability and validity.
Standard Deviation
Shows how far, on average, each score is from the mean.
Normal Distribution
Curve showing mode median and mean on a graph.
Positive Skew
Mode closer to origin, median and mean further.
Negative Skew
Mode further from the origin, median and mean closer.
Bimodal
Two modes on the graph.
Ethics
Standards of conduct outlined by the British Psychological Society that psychologists must follow.
Informed Consent
Participants must give consent with all important information provided so they can come to an educated decision.
Deception
Hiding information from or misleading participants.
Protection From Harm
Must protect the participants from being physically or psychologically injured.
Confidentiality
Participants information and behaviour must not be publicly shared.
Withdrawal
Participants have the right to withdraw at any time.
Debriefing
Investigators should explain everything after the investigation is concluded.
Peer Review
Evaluation of work by one or more people of similar competence to the producers. This helps maintain quality standards, improve performance and provide credibility.
Peer Review Evaluation
- Anonymity- may unfairly criticise those they have a grudge against.
- Publication Bias- file drawer effect- only publish statistically significant research.
- Burying ground breaking research- research that questions existing theories are often not published.
- Peers are in competition for scarce funding.
Attachment Research
Research shows that mothers can work, making more money for the economy.
Psychopathology Research
Anything to do with treatment and people ability to work and contribute to the economy as effective members of society.
Memory Research
Can help reduce costs in the justice system with cognitive interviewing and eye witness testimonies.
Statistically Significant
95% confidence
Type 1 Error 1+
False positive- rejecting the null when there is a possibility that the results were due to chance. Not being cautious enough.
Type 2 Error 2-
False negative- accepting the null when there’s a chance the results were significant. Being over cautious.
Statistical Test Table
. Diff. Aso
Level of Measure Ind. Rep. Cor
Nominal. Chi2. Sign. Chi2
Ordinal MW. Wil. Spear
Interval UT-t. RT-t. Pear
Nominal Data
Frequency/ categorisable e.g. food types
Ordinal Data
Rank order e.g. placements
Interval
On a scale e.g. minutes, percentage, number
Sign Test
- Determine if the test is one tailed or two tailed.
- Determine the direction of influence for the results (+-x)
- Add the number of times the least frequent sign appears. S = this number.
- Look at the critical value table and obtain the critical value for S looking in the 0.05 column and comparing with the number of participants - the no change ones.
- If S is the same or less than the critical value reject the null, if it is more, accept the null.
Content Analysis
Observe the presence of certain words, images or concepts within secondary data. Researchers count and examine the presence of certain words.
Content Analysis Evaluation
+ Quick and easy
+ Quantifies the meaning of text
+ Internally reliable
+ Ethical
- Accuracy of focus
- Researcher bias
Thematic Analysis
Looking for emergent themes in data, coding them and interpreting their meaning.
Thematic Analysis Evaluation
+ Flexible
+ Theoretically free
+ Detailed
- Researcher bias
- Time consuming
- Subjective
- Quality of coding
- Theme identification bias
- Specificity
Top-down analysis
Start with an idea and apply knowledge.
Bottom-up analysis
Open ended- no initial assumptions.
Case Study
An in-depth study using a range of methods on one person or a small group.
Case Study Evaluation
+ In-depth specific evaluation of data
+ Can help give insight into specific situations and scenarios
+ Can prompt further more generalisable research
- Little control over variables- poor reliability
- Extremely ungeneralisable
- Small sample in a specific situation
- Vulnerable to researcher bias
- Time consuming/ expensive
Thomas Kuhn- Definition of a Science
Should have a set of assumptions, scientific methods and terminology. Provide precise, operationalised hypotheses which can be tested.
Falsification
Should be able to test whether something is false or not.
Predictive Power
Theories should all predictions of future behaviour.
Nomothetic
Apply to who population
Idiographic
Apply to a set of individuals
Scientific Paradigms
Kuhn argues that psychology is fundamentally different to physical sciences, and does not have one unifying theory upon which ideas could be tested.
Ethical Considerations
We cannot isolate and test variables without being unethical.
Reports Structure
Title, Content page, Abstract, Introduction, Procedure, Results, Discussion, References, Appendices
Abstract
Brief summary of entire investigation in 100-150 words and is written last. Must include one sentence summary, sampling/ participants, procedure description, results description and implications.
Introduction
Review background research and quote your aim.
Procedure
Four sub-titled sub-sections:
- Design and overview- method, design, variables, controls, ethics.
- Participants and investigators- who were they and how were they selected?
- Apparatus- equipment
- Procedure- explain how you did the investigation from start to finish.
Results
Descriptive results with a summary and graphs.
Analysis of Results
Refer to inferential statistics/ statistical significance. Reject/ accept the null.
Discussion
- Explanation of findings
- Limitations and modifications
- Implications and suggestions
References
Last Name. Initials. (year) Title in italics. Publisher place: Publisher
Appendices
Any additional information like calculations, raw data, stimulus material, instructions and debrief etc.