research methods Flashcards
1
Q
experiment definiton
A
a scientific procedure to confirm or make discovery
2
Q
variables definition
A
independent variables are what you change
dependant variables are what you measure
3
Q
what is normal distribution
A
a normal distribution is the name given to data from a sample that follows the normal spread of data
4
Q
normal distribution graph?
A
symmetrical
5
Q
skewed graph
A
-two sides aren’t the same
-one tail is longer
-can be positive (right) or negative (left)
6
Q
positive skew on graph
A
mean is higher = positive skew due to longer tail
skew to the left (close to y axis)
7
Q
negative skew on graph
A
mean is lower due to negative tail being longer
8
Q
measure of central tendancy
A
-middle/centre of a data set
-e.g. mean, median, mode
9
Q
measure of dispersion
A
-how spread out your data is
-e.g. range, or standard deviation
10
Q
standard deviation
A
-measure of dispersion
-how much variation there is from the mean
-low = close to the mean
high = data is spread out
11
Q
standard deviation steps
A
1) find mean
2) minus mean from each data point
3) square each result
4) add all squared numbers
5) work out mean for squared number
6) square root steps 5 result
12
Q
what is informed consent (code of ethics)
A
-signing to say they are ok to be a participant and that they know what the experiment is about
- if you can’t get this, ask someone similar and ‘presume’
13
Q
what are the code of ethics?
A
- informed consent
- no deception
- no physical or mental harm
- break in confidentiality
- privacy issues
- withdrawing
14
Q
what is no deception (code of ethics)
A
if there is deception, you need to debrief
they then have the right to withdraw from the study
15
Q
what is no physical or mental harm (code of ethics)
A
- physical harm is never acceptable, unless it’s minor
- if occurs, study is stopped and they are treated
- they might need therapy or counselling
16
Q
what is break in confidentiality (code of ethics)
A
- results are confidential
Participants info should never be public - pseudonym can be made up of name or initials (real or fake)
17
Q
what is privacy issues (code of ethics)
A
- not invade participants privacy
- only collect data which you gained informed consent
18
Q
what is withdrawing (code of ethics)
A
- right to withdraw at any point during start middle or end
- can’t withdraw once data has already been published or used
19
Q
what does congruent mean?
A
the same
20
Q
what are extraneous variables
A
variables in an environment that needs to be controlled
21
Q
what are confounding variables
A
variables the researcher failed to control and have confounded the results
22
Q
what is cause and effect
A
you can certainly say that the change in DV is caused by the manipulation of IV
Close as we can get to proof
23
Q
what is lab experiments
A
- strictly controlled environments
- IV is the only thing that changes
- IV change causes change in the DV
- as everything is controlled it can be stated that there is “cause and effect”
24
Q
lab experiments advantages
A
- high validity - tested what they meant to
- provide info that can show cause and effect - due to high control
25
lab experiments disadvantages
- lack ecological validity called real world validity or mundane realism
- means there is a trade off between high control and ecological validity
26
what is external validity
ecological validity, real world validity and mundane realism
27
what are field experiment
- controlled conditions in real world
- extraneous variables are controlled as much as possible
- similar to lab experiments but in less artificial environments
28
field experiment advantages
- fairly good internal validity
- provide cause and effect evidence
- higher ecological validity than lab experiments
29
field experiments disadvantages
- greater risk of confounding variables affecting DV and therefore reducing validity
- more expensive
- some research is impossible if not in a lab (e.g. brain scan)
30
what is a natural experiment
real world looking at naturally occuring variable - e.g. natural disaster
31
natural experiment advantages
- highest ecological validity
- can study something that's unethical or too difficult or expensive to run
e.g. affect of murder on victims family or natural disaster
32
natural experiment disadvantages
- no control
- tricky to show "cause and effect"
- easy to misinterpret what's being recorded as so many confounding variables - leads to increase risk of subjectivity
33
what is a quasi experiment
- IV that isn't controlled by experimenter
- usually occur naturally
- most common is age and gender
34
Repeated measures
-study has one group of participants who take part in every level of the independent variable
e.g. if juggling, one group learns how to juggle with balls and batons
35
strengths of repeated measures
- no participant differences between the conditions
- it might be the participants who are different from the others
36
weaknesses of repeated measures
dependant variables can be influenced by order effects
37
Counterbalancing (repeated measures)
- to prevent order effects, you counterbalance the order conditions
- half learns to do one first and the other half does the other one first
- then average the halves together
38
Independent measures
- different groups of participants completes each level of the independent variable
39
strengths of independent measures
- prevents order effects from affecting the experiment
40
weakness of independent measures
- dependent variable may be influenced by participant variables as there are different participants in each group
41
dealing with issues in independent measures
Random allocation - less likely that one group will have the 'best' participants .
- prevents the researcher from manipulating results
- Put all names into a computer and ask to make group
42
Matched pairs
- different participants complete each level of independent variable
- participants are matched in the same way
- paired on key variables (age, reaction speed)
43
strength of matched pairs
- no order effects
- participant differences have been controlled
44
weakness of matched pairs
- you need lots of participants to choose from
45
double blind trial
- experimenter and participants don't know what group they are in
- reduces the 'screw you' and 'halo' effect and bias
46
single blind trial
- participant doesn't know what level of the IV they're in
- helps prevent participant knowing what the experiment is about
- prevents 'halo effect' and 'screw you' effect
47
why do we sample?
- its not possible to test everyone
- a sample is a small group of individuals who are representative of a whole group
48
Representative sample
- sample will have similar mode, mean, range and standard deviation - i.e. the spread of data
- trust the results more when the sample is representative of the population
49
Random sample
- every participant has equal chance of being selected
- assign a number to everyone in your target population. Put names/numbers into a computers random number generator. Ask generator to select the number of participants you require
50
Alternative method to random sampling
- write all names onto same size and colour paper
- fold each paper in same way and shake box
- complete a random draw until you have the number you need
51
Random sample strengths
- no bias as everyone has equal chance to be selected
52
random sample weakness
- when sample is small there is a chance it will be biased by chance
53
opportunity sample
- the researcher chooses participants who are there at the time they run the study
- people who are in right place at right time
- e.g. people who ask you questions while you are walking
54
opportunity sample opportunity
- easiest method
55
opportunity sample weakness
- sample is always biased as there are some who aren't there at that time
56
volunteer sample
- reasearcher advertises study and people volunteer to take part
57
volunteer sample strength
- reach wide variety - representative
58
volunteer sample weakness
- volunteer bias
- volunteers are more motivated - 'good participant effect'
59
systematic sampling
- choose participants through using a system
60
systematic sampling strength
- participants aren't selected in a biased way
61
systematic sampling weakness
- some people will never be selected
62
Stratified sampling
- when a population is made of groups and you want the sample to represent different groups, you need to take a stratified sample
63
how to stratified sampling
- get your list for each group in the population
- work out correct number needed for each group
- use random sampling to collect the number
64
strength of stratified sampling
- very likely to be representative
65
natural (field) observation
- take place in real world
- not controlled
66
natural (field) observation strength
- high ecological validity
- can observe things you cant in a lab
67
natural (field) observation weakness
- less control of the situtation
68
types of observation
- natural (field)
- lab
- participant
- non participant
- covert
- overt
69
lab observation
- observation is set up in a lab in highly controlled way
70
lab observation strength
- high control
- create the exact situation you want
71
lab observation weakness
- low ecological validity
- some things arent possible in labs
72
participant observation
- observer is taking part in whats happening
73
participant observation strength
- observe fine detail and emotion
74
participant observation weakness
- researcher misses the whole picture
75
non participant observation
- researcher doesnt take part
76
non participant strength
- researcher can observe whole situation
77
non participant weakness
- researcher might miss detail
78
covert observation
- people being observed don't know they're observed
79
covert observation strength
- no change in behaviour
80
covert observation weakness
- researcher might have to lie
- possible ethical issues and deception
81
overt observation
-participants know theyre being observed
82
overt observation strength
- no ethical issues
83
overt observation weakness
- likely to change their behaviour
84
ways to collect data
- event sampling
- creating categories
- time sampling
- inter-observer reliability
85
event sampling
- choose behaviour categories you would like to observe
- clearly defined behaviours to make easier to record
- might loose fine details
86
creating categories
- must cover all behaviours
- no overlap - clear enough for others to copy
- only one meaning
87
time sampling
- use when its difficult to know what categories tp use
- records in time periods e.g. every 120 seconds
- gives observer a system to make easier
88
event sampling and time sampling
- sometimes you can use both
- useful if observing a very chaotic environment
- might have created behaviour categories and only record certain times
89
inter-observer reliability
- use two observers and check results are equal
- reliable - 80% relationship between observers
- check strength of relationship with correlation
90
what is a hypothesis
a statement that predicts the expected influence of the change in the IV on the DV
91
steps to find the hypothesis
1) identify the researches aim and the type of study
2) identify exactly what is manipulated
3) identify the specific measure that will be taken
4) decide whether it should be directional or non directional
92
directional hypothesis
- states the specific difference you expect to find
- only if there is previous research
- allow you to use a one tailed test
93
non-directional hypothesis
- states there will be a difference but not a direction
- no previous research as this allows for any result to be accepted
94
operationalizing
- to write a variable in a specific way
- state the variable in an operable way
- very detailed
- so other researchers know exactly what has changed and what's measured
95
how to operationalise the IV
- MUST state clearly what's being manipulated
e.g. 70-95IQ and 105-130 IQ
96
how to operationalise the DV
usually a unit is included
- seconds
- estimates of distance in metres
- estimates of temp in °C
- a tally, count or frequency
97
correlation hypothesis
- in a correlation we have co-variables instead of IV and DV
- co-variables are linked to same things
- e.g. weight and height of each person in a group can be correlated
98
non directional correlational hypothesis
- just state there will be a relationship
- e.g. there will be a relationship between UK shoe size and a person's height in cm
99
directional correlational hypothesis
state if there will be a positive or negative relationship
100
null hypothesis
- every research has a hypothesis and a null hypothesis
- hypothesis is what researchers aims to test
- null hypothesis states the opposite of the hypothesis
- either the hypothesis or null hypothesis are correct
101
what is content analysis
research method in the social sciences where the content of some form of recorded communication is analysed
102
what should good interview/questionnare questions be like
- should be clear
- should be unambiguous
- should not be leading
- should not be double-barreled
103
closed questions
- yes / no
- circle / tick
- numerical answer
- likert scale (agree/disagree/strongly agree etc)
104
open questions
- qualitative
- ask for open ended info from ppt
- often have an answer box
- can be long or short
105
questionnaires
- useful for gathering large amounts of data quickly and cheap
- both open and closed questions
- questions are normally easier to answer
106
questionnaire strength
- quick and cheap
- easier to analyse as questions are less complicated
- easier to get larger sample
107
questionnaire weakness
- ppt can't ask questions to check their understanding
- if questions are confusing, unlikely to get valid data
- can't ask follow up questions
- can't prompt the ppt if a question confuses them
108
3 types of interviews
- structured - decide and only ask specific questions
- unstructured - researcher creates more questions in response to answers
- semi-structured - some pre chosen questions
109
strength of interviews
- can ask follow up questions, unless it's structured
- can prompt ppt
- get a lot more detail
110
weakness of interviews
- more training required
- have to meet ppt or online
- can take a very long time to analyse - 10x longer than interview itself
111
correlations
- comparing two 'linked' variables to see if they change together.
- these variables are called 'covariables'
112
correlation coefficients
- tell you how strong a relationship is
- tell you the direction
- coefficients use R as a symbol
- on a scatter graph, the closer the fot the high the correlation
- +1 is positive
-1 is negative
113
curvilinear
- as one increases the other increases until a point when as one increases the other decreases
114
advantage of correlation
- can be conducted on secondary data - cheap to run
- only method to show a relationship
- can be replicated easily - just have to share data
115
disadvantage of correlation
- often misinterpreted as cause and effect
- can't show cause and effect
- intervening variables - unidentified covariable that's better link than what's been chosen
116
case studies
studying one person in detail
117
strength of case studies
- allow researcher to study something that would be unethical otherwise (Phineas Gage)
118
weakness of case studies
- the ppt can't be compared to anyone else
119
longitudinal study
- studies that last months or years
- record changes over time
120
strength of longitudinal study
- only way to measure change in time
121
weakness of longitudinal study
- high number of ppt dropout (attrition)
- can be very expensive
122
meta-analysis
- collects secondary data
- then analyse this data
123
strength of metal analysis
- good chance of seeing large scale patterns
124
weakness of meta analysis
- as each researcher was doing something different, results aren't always valid
125
thematic analysis
- looking at key theme
- much like English lit
- summary of themes is written
- quotes are used to support the themes
- pilot study might be done to test if it's worth doing the big study
126
strength of thematic analysis
- lots of detail and keeps data with rich detail
- high ecological validity
127
weakness of thematic analysis
- researcher is more likely to be subjective
- comparison is much more difficult
- very time consuming and therefore expensive
128
an alternative to thematic analysis
- turn into quantitative data
- complete content analysis
