Research Methods Flashcards
State the 4 experimental methods
- Laboratoryexperiment
- Fieldexperiment
- Naturalexperiment
- Quasiexperiment
Lab experiment + A03
Controlled conditions, manipulates the
+ high degree of control = no EV, prevents EV from becoming confounding variables. High internal validity allowing cause + effect to be established
- lack external validity due to the artificial nature, meaning the study can lack ecological validity, meaning the findings cannot always be generalised beyond the lab and often lacks realism. PPTs know they are in the study and may change their behaviour.
Field experiment + A03
Natural conditions, manipulates the IV, ppt doesn’t typically know they’re taking part
+ natural setting = high ecological validity. Results are more representative of everyday behaviour. However, there is less control of EV, which can become confounding variables
- ethical issues, as ppts are unaware that they are taking part, so they cannot give informed consent. Research may involve a breach of their privacy.
Natural experiment + A03
Natural conditions, IV is naturally occurring (i.e flood and behaviour of people is compared)
+ higher external validity and high ecological validity given the real-life issues that are being studied and manipulated, HE, they have no control over the environment and EV meaning it is difficult to accurately asses (a confounding variable may have caused the cause/effect rs)
+ a unique insight gained into real-life situations HE may be a rare opportunity
Quasi experiment + A03?????
Controlled/Natural conditions, IV is a difference between people (i.e. gender and age)
- ppts can’t be randomly allocated to research conditions to remove bias. The IV is naturally occurring, so the level of
What is meant by an Extraneous Variable
anEVisavariablethatisnotmanipulatedbytheexperimenterbutmayhaveaneffectontheDV, making itdifficulttoestablishacauseandeffectrelationship.
Observational Techniques
Covert/ overt
Participant/ non-participant
Naturalistic/ controlled
Structured/ unstructured
Covert observation + A03
An undisclosed observation, observing people without their knowledge. The ppt may be told of their involvement after the study.
+ investigator effects are less likely, less chance that the ppt’s behvaiour will be impacted by the investigator, therefore no demand characteristics. more natural and representative of everyday behaviour
- ethical issues: ppts are unaware of their involvement in the study, meaning they cannot give their consent and cannot withdraw. Observing in a public area like a shopping centre is fine, so the researcher has to ensure that they don’t violate any privacy laws
Overt observations + A03
An observational technique in which the ppt knows they are being observed
+ more ethical than covert method. ppts can be informed of the aims and they can give consent. They can also exercise their right to withdraw, they are being protected
- possibility of investigator effects, bias can occur where the investigator influences the behaviour of the participants. Ppts will change their behaviour through demand characteristics. Natural behaviour is not being observed
Participant observations + A03
thepersonwhoisconductingtheobservationalsotakespartinthe activitybeingobserved.
+ researcher can obtain in-depth data. the close proximity allows a unique insight.
- possibility of investigator effects = demand characteristics = not natural behaviour
Non-participant observations + A03
The person conducting the study doesn’t participate in the activity being observed.
+ investigator effects are less likely, researcher may not be visible to the ppts, so the researchers behaviour will not have an imapct on behaviour, meaning it is more natural behvaiour
- due to a lack of proximity, the researcher may miss some behaviours
Naturalistic observations + A03
Controlled observations + A03
Structured observations + A03
Unstructured observations + A03
Difference between extraneous variables and confounding variables
How to write hypothesis’
Null hypothesis: no difference in [operationalized DV] (i.e. recall) between [condition one] (short men) and [condition two] (tall men)
Non-directional hypothesis: there will be a difference in [operationalized DV] between [condition 1] and [condition 2]
Directional hypothesis: Ppts who [condition 1] will get higher/lower [operationalized DV] compared to [condition 2]
Differentiate between the different variables
control: everything you want to remain the same
independent: the one thing you change
dependent: the change that happened because of the independent variable
operationalised: where it has turned into something that can be measured
Sign test method
1) Find whether the difference was an increase or decrease (put a plus or minus), if there was no change rule this data out
2) see which change occurred less, this is the S value.
3) if asked to calculate significant/not, use the table with the n value (RULE OUT NON CHANGES), and probability to find the critical value.
4) compare the s value and critical value (greater/less than will be given in the Q)
Explain difference in s values/difference
In a study, the bigger the difference between the before and after scores, the
lower/smaller the S-value and the
more likely we are to reject the null hypothesis.
Define validity
When a test measures what it claims to measure
There should be no extraneous or confounding variables
4 types of assessing validity
Face validity: looking at face value to see if it appears to measure what it claims to measure (common sense)
Concurrent validity: When a test produces similar results to the results obtained of another test of the same behaviour
Ecological validity:
Temporal validity:
Reliability
When the results are consistent every time the study is repeated
studies can be unreliable if there are EVs affecting the measurements and if the test used to measure the DV is unreliable itself
Internal and External reliability
internal: when the individual items within a test are consistent with each other.
external: when the results of a test are consistent every time it is repeated.
testing internal reliability: the split half method
split the results of the study in half to see if they correlate with each other
testing external reliability: the test-retest method
the researcher gets the same participants take the same questionnaire or test on two separate occasions and compares the results obtained both times, if the results are similar/same it has high external reliability
What is operationalised
Defining how the variables will be measured
t-test
calculates the probability of observing differences in samples if there is no difference between the populations
t value
the higher the t value, the lower the probability
gives the probability of observing our results if the null hypothesis is correct
factors affecting the t value
- dispersion (smaller dispersion, higher t)
- sample size (bigger sample, bigger t)
- mean
accepting/rejecting Ho
if the p value is lower than or equal to SL, reject Ho as there is still possibility of the difference being by chance
Type I error
incorrectly rejecting the null hypothesis, say there is a real difference between two experimental groups when there isn’t one.
Type II error
failing to reject the Ho when there was actually a difference
often due to very very low SL
SL and type I error
the higher the SL, the higher the chance of making a type I error
setting a very low SL results in a type II error
15% SL = 15% of making type I error
to reject Ho (t value)
obtained>=critical t
degrees of freedom for related/unrelated t test
unrelated: sample - 2
related: sample - 1
R: repeated, matched pairs
UR: independent gd
Mann-Whitney U
Is the difference in the sum of ranks big enough to reject the null hypothesis at a given significance level
u-value
the smaller the u value, the less likely it is that the null hyp is correct and more likely to reject
the bigger the difference of SoR, the smaller the u-value, the smaller p-value
factors affecting the u-value
