research methods Flashcards
lab experiment
where the iv is directly manipulated in a highly controlled artificial environment
+- got control over extraneous variables (situational variables ), scientific, predictable, reliable
– lacks ecological validity, demand characteristics, lack of construct validity.
field experiment
where the iv is directly manipulated in a natural environment with some level of control
+- high in ecological validity, less chance of demand characteristics
– higher chance of extraneous variables, harder t test cause and effect, less scientific, harder to get consent.
quasi experiement
a naturally occurring iv that is not directly manipulated in either controlled or natural environment
+- more ethical, nothing manipulated , natural behaviour
–difficult to carry out as iv already exists , time consuming , extraneous variables likely - less reliable.
what is an experimental method in general (+/-)
manipulating IVs to test cause and effect so the hypothesis can be tested.
+- more scientific, objective ( measuring behaviour) (data)
– demand characteristics
structured observation (+/-)
predefined behaviour categories.
+- easier to record, easier to establish inter-rater reliability.
– reductionist, reduce validity, observer bias.
unstructured observation (+/-)
no predefined plan they record everyhting.
+- increases validity, wide range of context.
– harder to record, hard to establish reliability, observer bias.
controlled observation(+/-)
a research method where researchers watch participants in an artificial environment
+- increases reliability
–lowers ecological validity
naturalistic observation (+/-)
takes place in natural environment
+- less demand characteristics/ high EV
–low in reliabilty
participant observation(+/-)
where the researcher takes part in the experimt and activites.
+-, greater accuracy/detail, sometimes only way eg .gang
– ethical issues, reduces validity if behaviours influenced
non participant observation (+/-)
the researcher doesn’t take part.
+- remains objective, no influence on behaviour
–can’t observe certain behaviour, less detail/accuracy
overt observation(+/-)
participants are aware they are being watched.
+- more ethical ( consent)
– risk of demand characteristics/ decreases validity
covert observation(+/-)
participants are unaware they are being watched
+- less chance of demand characteristics/ increases validity
–less ethical ,they can’t consent
event sampling (+/-)
every occurrence of behaviour as specified on predefined checklist. observed at specified time.
+- quantitative data, easier to record
– reductionist, reduce validity, observer bias.
(similar to structured )
what is time sampling ? (+/-)
behaviour on predetermined checklist recorded at specific time intervals.
+- reliable
– time consuming, lowers validity as behaviour may be missed.
what is an observation
where researcher observes and records participants behaviour, but does not manipulate any variables
(can be use for a correlation )
target population
the audience a piece of research is generalised to from the sample
sample
group of participants that are representative of a target population
alternate hypothesis (h1)
there will be a difference.
null hypothesis (h0)
there is no difference
one tailed hypothesis
iv-dv-iv directional ( one will be higher/less than the other iv )
two-tailed hypothesis
there will/will not be a difference
non-directional
extraneous variable
something that is not the iv but
CAN affect the dv through lack of control.
confounding variable
something that not an iv but DOES affect the dv through lack of control
types of extraneous variables
individual differences( age, gender, culture)
situational variables ( environment )
conditions and two types?
‘groups of IV ‘
control , experimental
repeated measures ( +/-)
(within groups)
same participants used in all conditions
+- any changes in conditions are likely due to IV not participant
less individual differences
– more chance of demand characteristics/order effects
independent measures (+/-)
(between groups)
participants only take part in one condition
+-task variable, less chance of order effects/ demand characteristics
– more individual differences
any change to condition is more likely due to participant variables
matched pairs measure (+/-)
recruit participants and match them based on characteristics
+- individual differences controlled for
– time consuming (have to be observed/tested before matching)
opportunity sampling (+/-)
researcher chooses the most convenient people to the study.
+- easiest, less time to locate sample.
– biased sample as selected from limited geographical areas, not very representative.
random sampling (+/-)
every member in target population ahs equal chance of being chosen as names are pulled out a hat.
+- no bias as all have equal chance
– difficult and time consuming, representative sample not guaranteed.
snowball sampling (+/-)
researcher selects a person then they ask the person if they know anyone willing to take part within the target population. process continues.
+- possible to include members where no lists/identifiable clusters exists eg. criminals who aren’t easy to access.
– biased as sample drawn from small part of target population
self-selecting sampling (+/-)
people will volunteer themselves to take part in the research.
+- gives a wide range of ppts, convenient, ethical as leads to informed consent.
– unrepresentative as leads to bias sample because they won’t respond unless they have an interest so may lead to demand characteristics.
What is nominal data
Categorical data- frequency table eh number of …
A questionnaire with closed questions so generally a number of how many times something has occurred.
Usually a smaller table
What is a +/- of nominal data and what measure of central tendency is it ?
The mode:
+- Quickest, easy to generate from closed questions, increase reliability.
- - superficial
What is a +/- of ordinal data and what measure of central tendency is it ?
Median :
+- more info, takes into account anomalies
- -time consuming
What is ordinal data
Scale: data ranked or ordered form biggest to smallest, usually a bigger table
What is interval data
The divisions are equal and fixed like time, weight ,height ,temperature.
What is a +/- of interval data and what measure of central tendency is it
Mean :
+-Scientific, comparable, more reliable
- - abstract method may not use a fixed measure, anomalous result may affect results significantly.
What is level of significance ( what%)
This determines whether we accept or reject the null hypothesis. Our results will have an observed value that we compare to the criptic value that we don’t calculate(95%)
If the probability is less than or equal to 0.05 (95% confident ) we accept the null. The 5% may be fluke or extraneous variables.
what ethical considerations come under the principle: respect.
confidentiality, right to withdraw, informed consent, equality- gender, ethnicity.
what ethical considerations come under the principle: competence (the ability to do something successfully )
no ethical issue but it is about professionalism, recognising ethical issues, peer review.
what ethical considerations come under the principle: responsibility
psychological harm ,debriefing
what ethical considerations come under the principle: integrity ( the quality of being honest)
deception, inappropriate relationship
what is an open question (+/-)
no pre-determined response, can express opinion so we find out why.
discuss, outline, why.
+- high in construct validity- detail
– hard to analyse and spot trends/patterns.
what are closed questions (+/-)
pre-determined response (include answer in example)
+- easy to analyse and interpret
– don’t find out why so lacks construct validity.
what is a Likert scale and what does it do.
how far do you agree/ disagree.
gain opinion and numbers.
what is a rating scale and what does it do.
on a scale of eg. 1-5
numbers only.
what is a semantic differential scale and what does it does.
opposing adjectives.
usually moved electronically.
flexible dynamic scale- opinion.
eg. l11l11l
strengths and weaknesses of questionnaires
+- quicker with a large sample.
more ethical as is kept confidential
– social desirability
greater chance of leading questions.
strengths and weaknesses of interviews
+- less chance of social desirability.
– less ethical as they see them and take a name so isn’t confidential.
– takes longer with a large sample.
what is a structured interview (+/-)
pre-determined set of questions in professional environment.
+- more reliable- asked the same questions.
– more chance of social desirability.
hard to analyse.
what is an unstructured interview (+/-)
no pre-determined set of questions it can be like a conversation.
+- less chance of social desirability.
– more detail - follow interests.
– less reliable, hard to analyse.
what is a semi-structured interview (+/-)
some pre-determined questions but some relation to conversation.
+- more detail but is reliable as there is some consistency making it valid.
– social desirability.
what is a strength of interviews
people can ask if they are confused.
what is meant by validity
how trustworthy the procedure is .
does the research set out to measure what it was meant to measure.
what is internal validity
look how specifically in the study behaviour was measured.
what is external validity
(generalisability)
outside of the study where we can generalise it to.
what is face validity (in)
does something look like it will measure what it is supposed to.
what is construct validity (in)
does the measure relate to assumed characteristics of what is being assessed.
what is concurrent validity (in)
one test correlates well with a measure that’s previously been tested.
what is criterion validity (in)
to extent to which a measure can predict the performance or behaviour of measured thing.
what is ecological validity (EX)
whether it generalises to a real world environment outside of the study.
external is high in mundane realism- in the middle can argue both.
what is population validity (EX)
is the sample representative of wider target population.
beliefs.
what is reliability
how consistent a piece of research is.
can you get the same again ?
what is internal reliability
high levels of control in the study as things are kept the same- standardisation.
what is external reliability
is their consistency outside of the study at different times/places.
what is inter-rater reliability
two or more people agree on the results of the research.
consensus of average 0.8
what do you get with high and low levels of control
high- extraneous variables.
low- demand characteristics/ less natural behaviour.
ways of getting high/low levels of internal reliability
high- lab experiment, controlled/covert observation, questionnaire eg .same questions, structured interview
low- observation, self-report - can’t be the same if repeated as is opinion.
ways of getting high/low levels of external reliability
high- big samples- more consistent as can replicate more people.
low- time differences (longitudinal) attrition affect?
what is the split half method
comes from self-report.
two halves of questions are similar so we can assume the test is reliable.
eg. asking a repeat question but wording it differently
what is the test-reset method
experiment
test them more than once
eg. start/end of month to see if attitudes are different.
what are type 1 errors
false positive
we accept the alternate but should have accepted the null rather than rejecting it.
we assume results were down to iv but it was down to chance
this may be because the p-value is to lenient like p<0.10
what are type 11 errors
false negative
we reject the alternate when we should have accepted it.
assume results were down to chance but they were down to IV.
this may be because p-value is to stringent (tight/strict) but sometimes this is necessary in the medical industry.