Validity Flashcards
internal/ experimental validity
- concerns the legitimacy + what is being measured in the actual study
- the extent to which a test study is investigating the effect of the IV on the DV
- researcher must make sure they are measuring what they intend to measure
factors that reduce internal/ experimental validity
- investigator effects
- demand characteristics
- confounding variables
- social desirability bias
- lack of operationalisation
investigator effects
the influence/ bias of the investigator on the study
- unintentional effect on participants + outcome of research
demand characteristics
when parts know the aim of the study and change their behaviour to affect the outcome - comply or ‘screw you’ effect
confounding variables
variables that have affected the DV apart from the IV
social desirability bias
when participants change their behaviour to be seen as socially desirable/ likeable/ not be judged
- affect DV + results gained
lack of operationalisation
not operationalised properly can result in low internal validity
ways of assessing internal validity
- concurrent validity
- face validity
assessing internal validity - concurrent validity
- way of establishing the internal validity of a new test
- scores gained from new test can be compared against an older, established test = validity is known
- if scores from both tests are similar and a positive correlation coefficient of +0.8 or greater is found, then the new test is judged as having high internal validity
assessing internal validity - face validity
- way to measure whether the test or measuring instrument is measuring what it should e.g. a questionnaire
- One or more researchers/experts in the field can examine the test items/questions to see whether they are measuring what it set out to
- do this by looking at the questions on a questionnaire and seeing, “on the face of it” whether the questions appear to be measuring what they should.
ways of improving internal validity
- concurrent validity
- face validity
- reducing investigator effects
- reduce demand characteristics
- confounding variables
improving internal validity - concurrent validity
- Concurrent validity can be improved (depending on the research method used)
- For instance if low concurrent validity is found on a questionnaire, then the researcher could remove questions that seem irrelevant or ambiguous
- then test the concurrent validity again
improving internal validity - face validity
- Face validity can be improved (for a questionnaire) by an expert in the field examining all of the questions on the questionnaire
- They might decide that some of the questions are not a good measure of the topic being investigated
- e.g. IQ, and they then might improve/rewrite/re-word certain questions again
- helps improve face validity.
improving internal validity - reduce investigator effects
- double blind procedure
- standardised instructions
- remove researcher so they’re not with parts at all = self administered
improving internal validity - reduce demand characteristics
- double blind procedure
- single blind procedure
improving internal validity - confounding variables
- lab experiment
- pilot study = change for actual procedure
- standardised procedure/ instructions
external validity
refers to factors outside of the research setting
- how well can the results be gained be generalised to other settings, people + time eras
types of external validity
- ecological validity
- temporal validity
ecological validity
able to generalise findings to other contexts/ settings
- if it has mundane realism = reflects real life
= increases ecological validity
temporal validity
the findings are true over a period of time + can be generalised to other historical time eras
ways of assessing external validity
- meta analysis
- ecological validity
- mundane realism
- demand characteristics
assessing external validity - meta analysis
- meta analysis conducted = comparison of findings from diff research studies that have investigated the same hypothesis, are compared
- consistent findings from diff research studies that have occurred across population, locations + periods of time = high ecological validity
assessing external validity - ecological validity
- environment should be naturalistic so there is high ecological validity
- a laboratory study might have low ecological validity cus the setting is not very natural + quite artificial
assessing external validity - mundane realism
- assess how the DV is measured
- task parts are given to do in the study + way its measured can affect the external validity
- task should have high mundane realism + should reflect a task that someone would do in everyday real life
assessing external validity - demand characteristics
- ensure that demand characteristics have been kept to a minimum
- by assessing how naturalistic the parts are behaving
- parts should not be aware of the true aim of the study = will change their behaviour + big effect on DV = low validity
ways of improving external validity
- demand characteristics reduced = double/ single blind procedure
- ecological validity improved = type of experiment and/ or observation
improving external validity - reducing demand characteristics
- demand characteristics reduced = improve external validity
- done using double blind/ single blind procedure
improving external validity - naturalistic settings
- naturalistic settings = improve ecological validity
- lab experiment can be carried out in more natural setting e.g. field experiment
- e.g. observation can be carried out in a covert manner instead
- these ensure parts behave more naturally + improve eco validity