Research Methods Key Terms (Only Tricky Ones) Flashcards
Validity
The extent to which findings of a study are representative.
External Validity
Measures the extent to which the results can be generalised to the rest of the population.
Ecological Validity
Type of EXTERNAL validity
The extent to which the tasks participants are given align to a real, everyday experience.
Internal Validity
Measure the extent to which the results are due to manipulation of the IV, rather than other confounding variables.
Temporal Validity
Measures the extent to which research findings are still relevant in the current age.
Face Validity
Type of validity ASSESSMENT
Assesses whether a study measures what it was set out to measure.
Predictive Validity
Type of validity ASSESSMENT
Measures how well a study can predict future behaviour.
Concurrent Validity
Type of validity ASSESSMENT
How closely 2 different tests of the same behaviour/skill agree with each other.
(Like: TEST-RETEST reliability)
Improving Validity
- Standardised procedure
- Controlled conditions
- Use Single/double blind
procedure - Avoid investigator effects
(e.g. overfriendly) - Disguise aims
- Reduce chances of demand
characteristics - Reverse scoring
(questionnaire) - Covert (observation)
Reliability
How well a study has been set up to see whether the IV is seen to affect the DV.
Internal Reliability
Extent to which a measure is consistent with itself
Split-half method
Type of INTERNAL reliability ASSESSMENT
Participants are split in half and the responses from each group are compared.
External Reliability
Extent to which a measure is consistent over time.
Test-Retest Reliability
Type of EXTERNAL reliability ASSESSMENT
Repeating the same test with to see whether results are similar.
Inter-Observer Reliability
The level of consistency between two or more trained observers conducting the same observation.
Improving Reliability
- Using SH/TR methods
- Standardised procedure
- Controlled conditions
- Training researchers
- Operationalised variables
Levels of Measurement
Indicate the nature of the data, or how precise it is. Different stats tests require different LOMs.
NOIR (Basic → Complex)
Nominal Data
Named Categories, no true mathematical value.
(Most basic form of data)
(e.g. hair colour)
Ordinal Data
Ordered data which understands the relationship between places, no true mathematical values.
Provides no certainty between each value (e.g. difference between 1st and 2nd)
Uses SCALES.
(e.g. shoe size)
Interval Data
Most sensitive & sophisticated LOM, has true mathematical value.
It can go below zero, however zero does not figure as it means ‘nothing’.
Can be converted to ORDINAL, but not vice versa.
(e.g. temperature)
Ratio Data
Has a true value of zero, but stops at zero, has true mathematical values.
Relationship between data is unknown.
(E.g. height → cannot be -ve)
Extraneous Variables
Any variable, other than the IV that could influence the measurement of the DV, which could cause an error of a false causal relationship between IV & DV.
Confounding Variables
Variable, other than the IV, that changes systematically between levels of IV (i.e. changing the IV will change the CV) hiding the IV’s true effect on the DV.
Reducing Variable Impact
- Random Allocation
- Control Variables
- Matched pairs design
- Counterbalancing
- Standardisation
- Single/double blind trials
- Ensure research is objective & unbiased
