Research Methods Flashcards
Lab experiment
Conducted in a very controlled environment.
Quasi experiment
Experimenter does not manipulate the IV. The IV is naturally occurring. This may be in a lab or field.
Field experiment
Carried out in someone’s natural environment/ everyday surroundings.
Advantages of a lab experiment
- highly controlled with reduced EVs
- replicable
- more objective
- see cause and effect
Disadvantages of a lab experiment
- low in ecological validity
- may cause demand characteristics
Advantages of a quasi experiment
- allows for research where the IV can’t be manipulated
- high in ecological validity, looking as ‘real’ problems
Disadvantages of a quasi experiment
- no control of IV
- less control of EVs (low validity)
- participants are aware they are studied (low validity)
Advantages of a field experiment
- high in ecological validity
- natural behaviour
- avoids demand characteristics and research bias/effects (high validity)
Disadvantages of a field experiment
- less control of EVs
- more time consuming
- ethical issues
Independent measures
Participants allocated to 2+ experimental groups, representing the different levels of the IV.
Advantages of Independent measures
- avoids order effects
- avoids the aim of the experiment being guessed (DM)
Disadvantages of Independent measures
- no control of participant variables
- more participants are needed
Repeated measures
Participants receive both parts of the IV. There may also be a control condition.
Advantages of Repeated measures
- good control of participant variables
- fewer participants needed
Disadvantages of Repeated measures
- order effects (but could be overcome but counterbalancing)
- purpose guessed (DM)
Matched pairs design
Participants receive one side of the IV but are matched with someone from the other groups on certain characteristics. Each group Is given one level of the IV.
Advantages of Matched pairs
- acts as a control for participant variables
- avoids order effects
Disadvantages of Matched pairs
- time consuming to match variables
- may not control all participant variables
Independent variable
The factor directly manipulated by the experimenter to oversee the effect on the variation of the DV. There are at least two levels of the IV.
Dependent variable
Measured by the experimenter to assess the effects on the IV.
Operationalisation
Defining the IV and DV in a way that they can be measured.
Confounding (extraneous) variables
Factors other than the IV that may cause a result.
Participant variables
A characteristic of the participants such as age or intelligence which could confound results.
Situational variables
A feature of the environment that may affect performance, such as a distracting noise or time of day.
Opportunity sampling
Selecting people most easily available at the time.
Advantages of opportunity sampling
- easy
- takes less time
Disadvantages of opportunity sampling
-bias as the sample is drawn from small part of the population
Self-selected sampling
Produced by asking for volunteers (in adverts).
Advantages of self-selected sampling
- convenient as it finds willing participants
- lower attrition rate
Disadvantages of self-selected sampling
-volunteer bias as participants will be a certain type of person (have extra time/ motivated)
Random sampling
Every member of a target population has an equal chance of being selected.
Advantages of random sampling
-unbiased as all have an equal chance
Disadvantages of random sampling
- more time consuming and effort as a lot of steps are required to contact people
Snowball sampling
Relies on referrals from initial participants to generate additional participants.
Advantages of snowball sampling
-hard to reach groups are easier to find (addicts)
Disadvantages of snowball sampling
- not a good cross section (representation) of population as it friends of friends
Null hypothesis (H0)
IV has no significant effect on the DV.
Alternative hypothesis (H1)
IV has a significant effect on the DV.
One-tailed hypothesis
Predicts a direction of the difference in the hypothesis. (e.g. will increase likelihood)
Two-tailed hypothesis
Predicts there will be a difference in the hypothesis. (e.g. will have an effect on likelihood)
Structured observation and 1+/1-
System used to restrict and organise the collection of information.
+inter-rater reliability improved as measures can be more consistent
-observers may ‘see’ or ‘hear’ what they expect- research/observer bias
Unstructured observation and 2+/2-
Observer records all relevant behaviour but has no system.
+useful when behaviour is unpredictable
+used in pilot studies
-not relevant or important behaviours recorded
-hard to do and may miss behaviours whilst recording
Covert observation and +/-
Observations made without a participant’s knowledge.
+natural behaviour
-raises ethical issues
Overt observation and +/-
Participant aware they are being observed.
+avoids lack of informed consent
-behaviour altered- observer effect
Naturalistic observation and +/-
Everything left as usual. Environment unstructured but may use structured techniques.
+high ecological validity
+natural behaviour
-little control of variables
CONTROLLED observation and 1+/3-
Some variables are manipulated by the researcher. Environment set up potentially.
+control focuses on particular aspects of behaviour (objective)
- feels unnatural
- lacks ecological validity
- vulnerable to demand characteristics.
Non-participant observation and +/-
Observer is not a participants in the behaviour being observed.
+increased objectivity because of psychological and psychological distance
-may misinterpret behaviour as an outsider which reduces validity
Participant observation and +/-
Observer is participant in the behaviour being observed.
+provides special insight into the behaviour
+monitor and record in closer detail
-objectivity reduced (observer bias)
-difficult to record un-obstructively in part of a group
Open questions and +/-
Invites respondent to provide own answers. Produces qualitative data.
\+rich detail \+not restricted \+increases validity -difficult to draw conclusions -look for trends rather than descriptive statistics
Closed questions and +/-
Fixed number of possible answers. Produces quantitative data.
+easy to analyse
+easy to draw conclusions
-oversimplifies reality
-low validity as can’t express precise feelings
Questionnaires
Respondent can record their own answers. Questions predetermined. Leading questions should be avoided.
+/- of questionnaires
+can be repeated quick and easily
+more willing to reveal closer info than in an interview
-less thoughtful response
-bias sample as only certain type of people do them
-limited as predetermined
Rating scales and +/-
Respondents aske to give assessment of views using a scale.
+objective towards feelings/attitudes
+quantitative data easy to analyse and graph
-may avoid extremed and go for the middle
-not true feelings so low validity
Likert scale
Indicated how much you agree or disagree. (very likely to very unlikely)
Semantic differential scale
Measures attitude towards something (motivated to depressed)
Interviews
Face-toface interaction which results in the collection of data.
+/- of interviews
+reveal more info as they have less time to think than when writing
+more thoughtful reponses are encouraged
-restricted number of participants as time consuming
-expensive to employ a trained interviewer
Structured interviews and +/-
Predetermined questions delivered by an interviewer.
+can be easily repeated
+easier to analyse as answers are predictable
-expectation may influence answer (interviewer bias)
Unstructured/ semi-structured interviews and +/-
Some predetermined questions but new ones developed as interview proceeds or no questions are decided in advance.
+more detail
+can access info that may not have been revealed
-expensive as requires trained interviewer
~
approximately
Type 1 error
False positive.
Rejects the null hypothesis when it is true.
Type 2 error
False negative.
Accepts the null hypothesis when it is not true.
Structured observational techniques and 2+/2-
Systematic procedures put in place to be objective and rigorous.
+more accurate on behaviour (not rusting what people say they do)
+objective
-may ‘see’ what is expected (observer bias)
-can’t provide info on what people think or feel
Behavioural categories and 3+/2-
What are they?
Separates a continuous stream of action into components, using a coding frame.
+important information isn’t overlooked
+easy to use
+increases inter-rater reliability
-may not cover all possibilities so to recorded (low validity)
-poorly designed coding frame reduced reliability
Event sampling and 2+/2-
How is it carried out?
List of behavioural categories. Count behaviour which occurs in a specific period.
+useful with occasional behaviour
+more manageable than continual recording
-miss things if too much happens at once
- may not be representative
Time sampling and 2+/2-
How is it carried out?
Record behaviour at regular intervals or at different times of day/month.
+good for time related behaviours
+more manageable way of recording than continual
-reduced validity as behaviour may occur outside of interval
-may not be representative
Concurrent validity
Could you compare a new method which an already established one? Should find a positive correlation.
Criterion validity
Does it allow you to predict future performance?
Content validity
Does it look like it is measuring what it is supposed to be?
Construct validity
Does it support to build on a theory?
Internal validity
The researcher is testing what they intend to test.
External validity
The extent to which results can be generalised to other people/settings.
Mundane realism
Extent to which it was alike to the real world/life.
Ecological validity
The ability to generalise results beyond the particular setting it is demonstrated in.
Population validity
The extent to which results can be generalised to other groups of people.
Internal reliability
Whether the procedure is consistent within itself.
External reliability
Whether something varies from one time to another.
Test-retest
The same test is given to the same participants on two occasions to see if the result is the same (external reliability.)
Split-half method
Test is split, scores compared and if the results are similar it is reliable (internal reliability.)
Inter-rater reliability
Consistence between two different interviewers/ observers (internal reliability.)
Respect-Informed consent
Participants are given information concerning the nature and purpose of the study and their role in it.
Respect- Right to withdraw
Participants should be told that they can stop participating if they are uncomfortable.
Respect-Confidentiality
A participants right to have personal info protected thorough withholding their name or keeping their info safe.
Respect- Privacy
Refers to a person’s right to control flow of information about themselves.
Responsibility- Protection
Participants should not experience negative psychical effects or negative psychological effects.
Responsibility- Debriefing
Post-research interview designed to inform about the true nature of a study. Restore the state they were on at the start.
Integrity- Deception
Occurs when a participant is not told the true research aims or not told what they are required to do.
Ethics committee
Groups of people who approve a study before it begins.
Presumptive consent
Method of dealing with informed consent/ deception. Would they agree to take part?
Correlation
Assess the degree to which two co-variables are related. Correlational analysis. Correlation coefficient describes to extent of correlation between the figures -1 and +1.
Nominal data
- results in categories
- can calculate mode
- display on bar chart or pie chart
Ordinal data
- ordering/ranking
- eg scores on a test out of 100
- can calculate median
- plot on a bar graph
Interval/ratio
- classifies, order and specifies distance between each interval
- can calculate mean
- (interval can go below 0 e.g. temperature)
- (ratio can be scores of a test which has a set maximum score)
- tend to treat as ordinal in stats test
Nominal
Independent measures
Chi squared-compare proportions of each category
Nominal
Repeated measures
Sign test- difference between two conditions for the same person
Nominal
Correlation/Association
Chi Squared-compare proportions of each category
Ordinal
Independent measures
Mann Whitney U test-compares size of differences between ranked data
Ordinal
Repeated measures
Wilcoxon Signed test-size of difference between conditions
Ordinal
Correlation/Association
Spearman’s Rho Correlation -strength of relationship between variables
Abstract
A brief summary of the whole report covering the aims/hypothesis, method, results and conclusions.
Introduction
Relevant theories and background research discussed allowing the researcher’s own aims/ hypotheses to be put forward.
Method
Includes: design,participants, materials, procedure.
Results
Reports finding of the study and makes use of descriptive statistics and inferential statistics.
Discussion
Summary of findings and possible explanations as well as possible issues with research and suggested improvements.
References
Give detail on all research documents, journals, books used within the research.
Appendices
Comprises all materials, raw data, full instruction or procedure details and statistical calculations.
Peer review
By professionals in the same field.
- allocation of research funding
- publication in scientific journals
- research rating of a university department
Mean +/-
+sensitive measure as it reflects the values of all the stat in the final calculation
- can’t be used with nominal data (used for interval/ratio)
- can be unrepresentative of the numbers is there are extreme values
Median +/-
+not affected by extreme scores (ordinal)
-does not reflect all values
Mode +/-
+useful for data in categories (nominal)
-not a useful way of describing data when there are several modes
Range +/-
+easy to calculate
- affected by extreme values
- fails to take into account the distribution of numbers
Variance and SD +/-
+precise measure of dispersion
+not difficult to calculate
-may hide some of the extreme characteristics of the data
Researcher bias
Expectations or beliefs may encourage certain behaviours in participants. This results in the researcher’s expectations being fulfilled.
(Observer bias= expectations affect what is recorded.)
Researcher effect
Anything the researcher does which has an effect on the participant’s performance, other than what is intended.
(Observer effect= present of an observer may affect the behaviour.)
Normal distribution curve
All the measures of central tendency are at the mid-point due to the nature of the variable which usually clusters here.
The curve is in a bell shape.
Skewed distribution curve
Measures of central tendency are not the same.
Positive skew= scores are bunched towards left. The mode is to the left of the mean.
Negative skew= scores are bunched towards the right. The mode is to the right of the mean.
Probability
A measure of likelihood that an event may occur.
Significance
Indicates if results can be used to reject the null hypothesis and accept the alternative (if not, then the null is accepted).
Significance level
Level of probability agreed to reject the null hypothesis, usually 0.05.
Table of critical value
Table which tells you whether an observed value is significant or not. Different table for each inferential test.
Correlation 3+/3-
+look at relationships between continuous variables
+can rule out a casual relationship
+strong correlation leads to further investigation
-cannot show cause and effect
-intervening variables can explain the link
-methods used to measure either co-variable may lack reliability or validity
What does the VARIANCE + SD tell us?
- measure of dispersion
- average difference between each score and the mean
- SD is in the same units as the mean
How science works
Objectivity
Uninfluenced by personal opinions or past experiences, free of bias.
How science works
Replicability
It should be possible to obtain the same result if the study was to be repeated. Measure of validity.
How science works
Control
The extent to which a variable is held constant or regulated.
How science works
Cause and effect
The aim of the research is to demonstrate a relationship between the IV which has been deliberately manipulated and observing tis effect on the DV.
How science works
Standardisation
All procedures are the same for every participant.
How science works
Induction
Reasoning form the particular to the general.
Observe something then generate a theory.
How science works
Deduction
Reasoning from the general to the particular.
Generate a theory and set out to test it.
How science works
Falsification
The attempt to prove something is wrong.
How science works
Quantifiable measures
Recording in numbers for ease of analysis.
Experimental realism
The extent to which the procedure is convincing.
