Research methods 1 Flashcards
What is an alternate/ experimental hypothesis?
a clear, testable, precise statement that you wish to test which contains variables that have been operationalised.
Suggests there is a change
what is a null hypothesis?
all results are obtained due to chance not the iv.
What is a directional/ one tailed hypothesis? when is it used?
predicts direction change is expected to occur in eg bigger, smaller
used when previous research suggests and direction
What is a non-directional/ two tailed hypothesis? when is it used?
predicts change but no specific direction eg effect, change
used when no previous research or previous research is contradictory
what is operationalisation?
This term describes when a variable is defined by the researcher and a way of measuring that variable is developed for the research
what is a lab experiment?
eg?
takes place in a highly controlled environment
eg loftus and plamers study of effect of leading qs on memory
+ lab experiment
complete control of variables
forces pace of research
reliable (easy to replicate)
quantitative data
can use technical equptment
negatives of lab experiment
loss of ecological validity
drawbacks of experimental design
demand characteristics
sampling bias
ethics
what is a field experiment?
eg?
takes place in natural environment so reflects real life
eg pilliavans subway Samaritan study
+ field experiment
can witness non biased behaviour
more valid- better reflection of real life
negatives of field experiment
difficult to control variables
cant use as much technical equipment
cant control pace of research
not reliable (not easy to replicate)
what is a natural experiment?
eg?
researching the aftermath of something that has already occurred
eg effects of tsunami on future mental health
What is a quasi experiment?
eg?
researcher doesnt directly control iv but exploits naturally occurring differences
eg if iv is gender, this isnt technically an experimental method
+ quasi experiment
reduces demand characteristics
lack of direct intervention
allows comparison between groups
negatives of quasi experiment
loss of control over variables
what is an extraneous/cofounding variable?
Extraneous- nuisance/ extra variables that affect the DV- things we know about eg situational/ppt variables, demand characteristics, social desirability, order effects, researcher bias
confounding variables- variables that change with the IV- cant be sure if its the IV or the CV that is causing the change
name 6 extraneous variables
how to control
demand characteristics- changing behaviour to fit expectations
social desirability- changing behaviour to look good
investigator effect/ researcher bias- researcher somehow influences outcome eg through body lang or expectations
situational variables- outside influences eg time. control with standardisation (instructions, procedures etc same for each ppt)
order effects- boredom or fatigue. control with counterbalancing (ABBA technique- changes order of tasks each ppt uses in repeated measures) or randomisation (order of task etc decided by random method eg coin toss)
participant variables- difference between them eg age
what is an experimental design?
way of organising participants into groups
what is repeated measures?
using same participant in each condition
+ repeated measures
controls individual differences
fewer particiapnts needed
negatives of repeated measures
low validity
order effects
what is independent measures?
using different participants in each condition.
may select itself eg age or gender
+ independent measures
no order effects
more valid
negatives of independent measures
reduced control of individual differences
need more participants
what is matched pairs?
using similar but diff participants in each category eg twins
+ matched pairs
can clearly see differences between them
no order effects
control of individual differences
negatives of matched pairs
difficult to recruit
so time consuming
so expensive
what is a naturalistic obs?
studying people in their natural environment without interfering
what is a structured observation?
know in advance what you will look for- devise a checklist. quantitative data in form of tally
what is event sampling?
keeping count each time behavior occurs
what is time sampling?
note behaviours displayed at set time intervals eg every 20s
what is an unstructured observation?
when used?
how to increase accuracy?
researcher records all behaviour. quantitative and qualitative
used when dont know what to expect
inter-rater reliability- 2 observers trained in same way and must meet 80% agreement
what is a covert obs?
participants unaware theyre being observed
what is an overt obs?
participants aware theyre being observed
what is a participant obs?
researcher is part of action theyre observing
+ structured obs
quantitative data analyse
reliable (can be replicated)
negatives of structured obs
researcher bias
hard to decide what to look for
miss behaviours not on list
+ unstructured obs
good if dk what to expect
wont miss anything
negatives of unstructured obs
qualitative hard to record
less specific purpose
hard to record everyhting
+ covert obs
reduces demand characteristics
increase validity as no artificial behaviour
negatives of covert obs
participants havent consented
but in public anyway
+ overt obs
more ethical -consent
negatives of overt obs
demand characteristics
reduce validity
+ participant obs
in action so can see things you cant from far away
negatives of participant obs
researcher bias
may influence behaviour
hard to see everything from middle
difficult to make notes whilst in action
what is self reporting?
what needs to be done first?
asking participants questions through interviews or questionnaires
need pilot studies to check for ambiguity and increase validity and reliability
what are 4 types of question?
open- elaborate
closed- forced to pick
contingent- depend on previous answer
Likert scaling- assess strength but use odd nos so people often pick middle
why use a variety of q types in self report?
not bored
gather diff types of data eg qualitative and quantitative
what to avoid in questionnaires?
jargon
emotive language
leading qs
double barrelled qs
double negatives
+ interviews
highly valid
see body language- put them at ease so decrease demand characteristics
can elaborate on qs
+ structured interview
less researcher bias
reliable (standard qs so can have multiple interviewers)
valid, focused, targeted
can ask set qs then others
+ and - unstructured interview
+
flexible, follow up qs
broader info
more comfy- less demand characteristics
_
less reliable( hard to replicate)
and researcher bias
negatives of interview
smaller sample size
low reliability
difficult to analyse
can see and misjudge body lang
demand characteristics and researcher bias
+ questionnaire
quick, easy,cheap
reliable
ethical(consent)
gather lots of data
quantitative data
Ask range of qs so engaging
negatives of questionnaire
low validity (demand characterisics)
social desirability
lading questions- bias
can control sample
what is a corelation?
what is a corelation coefficient?
the relationship btwen two variables
tells you the strength of the relationship
+ of correlations
useful starting point for futher research
quick, easy,cheap
negatives of correlations
dont explain why it happens
validity flawed if method of measurment is poor
what is random sampling? + and -
all members of target pop have equal chance of being chosen
+
no researcher bias
equal chance of selection
-
difficult to access list of target pop
may not be representative
time consuming
people picked might not have consented
what is stratified sampling? + and -
target pop divided into subsets and random sample taken from subsets
+
more representative
free from researcher bias
-
difficult to get list of target pop
time consuming
subsets may not represent how people are diff
what is opportunity/convenience sampling? + and -
selecting anyone available from target pop
+
quick, easy, convenient
-
researcher bias
not representative
what is self selected/ volunteer sampling? + and -
participants volunteer sometimes not on purpose eg passersby
+
simple and quick
willing to engage
-
demand characteristics
participants similar eg extroverted so not represntative
ppts may have ulterior motive eg in schaffer and emmersons study, mothers may just want reassurance of their parenting
what is systematic sampling? +and -
picked from list at fixed intervals
+
no researcher bias
-
might as well do random sample
time consuming
participants may not consent
what is a snowball sample? + and -
getting to a hidden population
+
representative
find hidden pop
no researcher bias
-
may endanger participants
difficult to find hidden pop
less control
what is reliability?
interrater reliability
test retest
consistency of test/ procedure
would 2 people come to same conclusion with 80% concordance
would results be consistent in a second test
what is validity? what are the 5 types?
something being valid and comparable to real life
face validity-would common sense tell us it would work
predictive validity- predicts future performance w some accuracy
concurrent validity-compares 2 methods of testing to see if results are comparable
ecological validity- findings generalized beyond lab
temporal validity- results valid in past but not now
what is primary data? + and -
data collected directly by researcher
+
more up to date
more accurate
conrol over quality and accuracy
-
expensive
time consuming
researcger bias
what is secondary data? + and -
data that already exists but is analysed by researcher
+
cheap, quick, easy
historical, comparable data
benefit from pervious researchers expertise
-
may be inaccurate or irrelevant
no control over quality or accuracy
what is meta analysis?
what is triangulation?
combining primary and secondary to come to overall conclusion
looks for similarities and differences. combines strengths of both
linked to concept of triangulation which improves reliability and validity by using more than 1 research method or source of data
what is interval data?
Evenly spaced interval between each of the values eg minuets
what is ordinal data?
scores placed in rank order
what is nominal data?
frequency
what are the measures of central tendency? + and -
mode-most frequently occuring value
+not affected by extreme variables, easy to calculate
_ doesnt take all values into account, may not be a mode
median- middle value
+ not affected by extreme value, more representative
_ ignores other values, extreme values can be useful, just numerical
mean- arithmetic average
+ takes all values into account, more general and overall impression
_affected by extremes, not representative as easly distorted
what are measures of dispertion?
what are they?
examine variability in data set
help us understand if data is similar or diff
range- over how many numbers distribution is spread
standard deviation- how far each score is from mean
interquartile range- subtract Q1 from Q3
range+ mode
standard deviation+ mean
Interquartile range+ median
on which graphs is the x axis continuous? discrete?
continuous-histogram, line graph(frequency polygon)
discrete- bar chart
what is normal distribution? what are characteristics of the curve?
definition
- bell shaped
- symmetrical
- mean, median, mode all on same point
- two tails never reach horizontal axis
what are the 4 types of distribution? what do they look like?
normal- bell shaped
positive/right skew- coming towards you
negative/ left skew- going away from you
bi-modal- camel
what are orders of magnitude?
used to make up very approximate comparisons and reflect very large differences
what are ethics? what are the BPS guidelines?
standards of conduct distinguishing between right and wrong
Deception
Protection from harm
Confidentially
Withdrawal
Debriefing
what is peer review? what 4 things can be done?
evaluation of work by one or more people of similar competence to the producers of the work. maintain standards of quality, improve performance and provide credibility
1) unconditionally accept
2) accept but ask to improve (most common)
3) reject but encourage resubmission
4) reject outright
evaluate peer review
anonymity- reviewers are anonymous so may use this to unfairly criticize if hold grudge
publication bias- file drawer effect. only produce statistically sig findings
burying groundbreaking research- research that falsifies existing theories will find it hard to be published
what are the implications of psychological research for the economy?
how gov can use findings and how useful they are
1) attachment research- Bowlbys WHO report in 1950s stated babies needed mum for healthy psychological development but later evidence disproved this so mums went back to work
2) psychopathology research- to do with treatment and ability to work eg therapy initially more expensive than drug but get back to work sooner
3) memory research-
4) forensic psychology research-
what are inferential statistics?
discoveringif results are statistically significant. 5% or 1/20 or o.o5 chance that it is due to chance. express in terms of null- accepting or rejecting
what is a type 1 error? type 2?
type 1- false positive. not cautious enough. rejecting null when shouldn’t
type 2- false negative. being over cautious. accepting null when shouldn’t
describe the sign test
1) add up no of time least frequent sign appears ( doesnt matter if + or - as long as all that are same have same sign) this value is s
2) add no of participants (ignoring no changers) this value is n
see if hypothesis is 1 or 2 tailed
check critical value table for 1 or 2 tailed at the value of n
3) if our value of s is less than or equal to the value in the table, it is statistically significant and we can ignore the null. vice versa
describe how you would obtain a stratified sample (4)
- identify strata/sub-groups in their population, eg the four different schools
- calculate the required proportion from each stratum based on the proportion in the population
- select sample at random from each school/stratum/sub-group.
- use a random selection method, eg assign each student a number then use a computer,
calculator or random number table to select specified number of numbers between 0 and X (or
hat method).
