research methods Flashcards
(191 cards)
1
Q
qualitative data
A
detailed data in the form of description
2
Q
quantitative data
A
numerical data that can be turned into statistical form
3
Q
longitudinal study
A
- study same group / person over time
- tracks development of behaviour
- collects multiple sets of data
4
Q
snap-shot data
A
- concluded at one point in time
- collects one set of data
- doesn’t track development of bhv
5
Q
ecological validity
A
whether the task and setting are representative of real life
6
Q
experiment
A
setting up a situation and studying behaviour
- lab
- quasi
- natrual
7
Q
observation
A
watching people with or without knowledge usually looking for certain pre-decided behaviour
8
Q
self-reports
A
asking ppts about their behaviour by using questionnaires or interviews
9
Q
correlations
A
looking at how 2 variables are related
10
Q
free will
A
human beings are entirely free to act as they chose and bare responsibility for the outcome of their behaviour
11
Q
determinism
A
suggests we lack control of our behaviour and it is pre-determined by factors such as genes and past experiences
12
Q
usefullness
A
research that enhances our knowledge and can be applied to real life situations
13
Q
limited usefulness
A
research that may lack credibility, generalisability and understanding or be difficult to apply to real life
14
Q
nature
A
behaviour is due to biological factors such as genetics, nervous systems
15
Q
nuture
A
sees behaviour as learnt or aquired through experiences in the environment
16
Q
interactionist
A
accepts both nature and nurture as being interconnected and human behaviour is a product of both
17
Q
reductionism
A
attempts to break down behaviour into constituent parts and uses single factors to account for a given behaviour eg genes
18
Q
holism
A
sees behaviour as too complex to be reduced and there are many factors contributing to behaviours
19
Q
individual
A
looks to a persons personality and dispositions as the cause of behaviour
20
Q
situational
A
behaviour is caused by situations around individuals eg group members or environmental context
21
Q
internal reliability
A
the extent to which we can replicate the procedure
22
Q
what does internal reliability concern
A
procedure (all, always, same)
23
Q
how is internal reliability increased
A
standardisation - keeping the procedure the same
24
Q
how is internal reliability checked
A
split-half method
test-retest method
25
external reliability
the extent to which we can replicate the findings and achieve consistency
26
what does external reliability concern
findings
27
how do we increase external reliability
quantitative data - makes comparison easier
28
how is external reliability checked
split- half method
| test-retest method
29
inter-rater reliability
whether 2 or more researchers find and conclude the same thing
30
when is inter-rater reliability used
mainly observations
31
how is inter-rater reliability increased
pilot studies
| use behavioural categories
32
internal validity
whether the study measures what it se out to measure
33
what does internal validity concern
procedure
34
what needs to be controlled to increase internal validity
extraneous variables
35
external validity
whether that findings can be generalised to real life
36
what does external validity concern
findings
37
test-restest validity
when ppts are tested more than once on separate occasions in the same condition
38
split half method
2 halves of a questionnaire are similar
39
population validity
whether the sample makes findings applicable to real life
40
ecological validity
task or setting true to real life meaning findings can be generalised
41
construct validity
extent to which a test measures all aspects of particular behaviour
42
criterion validity
the scores on one measure are able to predict the outcome on another related measure
43
predictive validity
a measure can predict future behaviour or attitude
44
3 types of extraneous variables
situational variables - env
individual variables - ppts
researcher effects
45
situational variables
lighting
distractions
time of year / day / month
location
46
individual differences
gender
age
personality
cognitive ability
47
researcher effects
facial expressions
body language
conscious/unconscious cues
48
face validity
extent something looks as if it will measure what its supposed to
49
measure of central tendency
measure of averages
| eg mean median mode
50
mean equation
∑(x÷n)
51
+ of using mean
all data is used
accurate rep of the data
first choice of central tendancy
52
- of mean
anomalous result can distort the values
| not appropriate if data is skewed
53
+ of the median
extreme scores don't distort the value
| use if data set is skewed by extreme values
54
- of the median
difficult and time consuming to calculate
| less representative of all values
55
+ of mode
can use when data is not numerical
| allows analysis for most occurring category
56
- of the mode
may not accurately reflect data set
| if no most popular answer , not useful
57
discrete data
can be placed into categories
58
continuous data
can be placed on a number line
59
bar chart
used for discrete data only
60
histogram
used for continuous data
| area of columns = frequency
61
frequency equation
frequency x class width
62
frequency density equation
frequency ÷ class width
63
pie chart
used for discrete data
| shows relative contribution to overall total
64
line graph
continuous data
continuous scale
shows change over time
65
scatter graph
continuous data
| measures relationship between 2 variables
66
operationalisation
how you make a variable measurable
67
what variables are used in experiments
independent variable
dependent variable
extraneous variables
68
what variables are used in observations
co-variables (2 behaviours being measured)
| extraneous variables
69
independent variable
manipulated / changed by the researcher
70
dependent variable
behaviour being measured
71
alternate hypothesis
statement of prediction between variables
72
correlational hypothesis
predict the relationship between 2 variables
73
one-tailed hypothesis
predicts specific direction of resulty
74
two-tailed hypothesis
predict a difference will be found but are non-directional in terms of what will be found
75
null hypothesis
states no difference will be found
76
target population
the group of people the psychologists want to be able to generalise their findings
77
4 methods of sampling
opportunity
volunteer
random
snowball
78
opportunity sampling
most common method
| uses people who are readily available
79
strength of opportunity sampling
easiest
quickest
most economical
80
weaknesses of opportunity sampling
likely to produce bias
| not very representative
81
volunteer sampling
psychologist makes an advert
| people who want to participate volunteer
82
strength of volunteer sampling
wide range of access to ppts
ethical => informed consent
convenient
83
weaknesses of volunteer sampling
unrepresentative
| people tend mot to respond unless interested
84
random sampling
everyone from TP has equal chance of selection
| all names are entered into a draw and randomly selected
85
strengths of random sampling
least biased method
86
weaknesses of random sampling
very difficult and time consuming
| limits TP
87
snowball sampling
used if population isn't easily contacted
| ask someone to ask someone else
88
strength of snowball sampling
possible to use members of groups where no lists or identifiable clusters exist
89
weaknesses of snowball sampling
no way of knowing if sample is representative of the population
90
ethical guidelines
rules which help to keep principles
91
4 principles
respect
responsibility
integrity
competence
92
principle of respect
respecting a persons individual rights
93
respect guidelines
informed consent
confidentiality
right to withdraw
94
strength of breaking 'informed consent'
allows further insight and prevents demand characteristics
95
strength of breaking confidentiality
may be able to offer therapy or treatment
96
strength of breaking right to withdraw
see the true behaviour and increases insight
97
principle of responsibility
upholding responsibility keeps participants safe
98
responsibility guidelines
protection from harm (physical and psychological)
| debrief
99
strength of breaking protection from harm
increases insight
| can lead to ground breaking discoveries
100
principle of integrity
being honest and truthful in research
101
integrity guidelines
deception
102
principle of competence
followed all guidelines properly or made suitable adjustments that are broken
103
competence guidelines
all of them
104
experimental design
control the influence of ppt variables in an experiment
105
types of experimental design
independent measures
repeated measures
matched patricipants
106
independent measures design
- 2 or more experimental conditions
- different ppts take part in each condition
- ppts are randomly allocated to one of the experimental conditions
107
strength of independent measures design
eliminates order effects eg practice, boredom, fatigue
108
weaknesses of independent measures design
- risk of individual differences
| - more ppts needed
109
repeated measures design
- 1 group of ppts
| - all ppts take part in all experimental conditions
110
strength of repeated measures design
- less risk of individual differences
| - more economical, less ppts needed
111
weakness of repeated measures design
increased risk of order effects
112
matched participants design
- 2 different groups of ppts
| - researcher allocates ppts to each group so that they all match in terms of key characteristics
113
strength of matched ppts design
- eliminates order effects
| - controls for individual differences
114
weaknesses of matched ppts design
- time consuming
| - ppts can't be fully matched
115
over come order effects
counter balancing => ABBA
116
how to control situational variables
lab experiment
117
how to control individual differences
repeated measures or matched ppts design
118
control researcher effects
double blind trial or training researchers
119
demand characteristic
changing behaviour to fit the aims of the experiment
120
how to control demand characteristics
use single blind procedure
field procedure
im or mp
121
social desirability bias
personality traits / bad habits hidden as not socially accepted
122
single-blind procedure
ppts not told the aims of the study
123
double-blind procedure
ppts not told the true aims of the study and experimenter not aware of which condition ppt is in
124
the aim of MY OWN experiment
to investigate whether males are better than females at visual spatial tasks
125
what is the iv and dv in MY OWN experiment
```
IV = gender - male vs female
DV = time - time taken in mins to complete maze
```
126
sample used in MY OWN experiment
opportunity sample
| 19 students from 6th form college ages 16 - 17
127
experimental method used in MY OWN experiment
quasi
128
experimental design used in MY OWN experiment
independent measures design
129
conclusion from MY OWN experiment
- males completed the maze quicker than females on avg
| - men are better than women at visual spatial tasks
130
naturalistic observation
observation carried out in a natural environment
| focuses on ppts naturally occurring bhv
131
controlled observation
conducted in a lab
allows control of env
mostly observe bhv through a one way mirror
132
participant observation
observer becomes part of the group / situation
| produces qualitative data of bhv
133
non-participant observation
observer remains external - watches from a distance without ppts knowledge
134
overt (disclosed) observation
ppts are aware they're being observed
135
covert (undisclosed) observation
ppts don't know they're being observed
136
unstructured observation
researcher write bhv as they see it
| analyse later by looking for patterns of behaviour
137
range
- how dispersed data is
- large range => very spread out
- small range => close together
138
equation for the range
highest score - lowest score + 1
139
variance
spread of scores around the mean
- large variance => scores are inconsistent and far from the mean
- small variance => sores are consistent and close to the mean
140
population variance equation
(∑(x-μ)^2 )÷n or 1÷n ∑(x-μ)^2
141
how to calculate variance (step by step)
1. calc mean
2. work out difference between each score and mean
3. square the differences
4. find sum of squared values
5. ÷ sum of differences by N an - 1
142
standard deviation
average amount numbers that differ from the mean
- large SD => scores are varied
- small SD => scores are consistent
143
strength of using the variance
takes all scores into account
| more precise and representative of dispersion
144
weakness of variance
mat hide characteristics of data which could skew it
145
strength of Standard deviation
- takes all scores into account - more precise and rep
| - returns the units to the same figure as the mean , easier to make direct judgements about data sets
146
what is skewed distribution
when data is skewed to one side of the bell curve
| - mean, median and mode aren't equal.
147
what causes positive skew
- extremely high scores pull mean to the RIGHT
| - there will be a long tail to the right
148
what causes a negative skew
- extremely low scores pulls the mean to the left
| there will be a long tail to the left
149
levels of measurement
how the data was measured and how precise it is
150
3 levels of measurement
1. nominal data
2. ordinal data
3. interval data
151
nominal data
- data in categories
| - shows number of times bhv occurred
152
ordinal data
- data put in rank order
| - the differences between each rank isn't known and doesn't have to be equal
153
interval data
- objective scale
- data comes from a scale of = or known units with equal intervals
- uses precise mathematical units
154
strength of nominal data
- easy to generate
| - large amount of data can quickly be collected; reliable
155
weakness of nominal data
- without linear scale = ppts unable to express degrees of response
- can only use the mode as measure of spread
156
strength of ordinal data
indicates relative values on a linear scale instead of just total
=>more informative then nominal data
157
weakness of ordinal data
dont know the size of the gap between values or if the gaps are equal
158
strengths of interval data
- more informative => points directly comparable because equal value
159
probability
- how likely something is going to happen
| - assesses the probability of an event; if the data is due to chance or not
160
significance
when low prob the difference between variables were due to chance => can accept the alternate hyp
161
link between significance and probability
probability of the results being due to chance increases => significance of results decreses
162
usual probability significance level
p<0.05
| probability results were due to chance is less than or = to 0.05/5%
163
type 2 error
when you wrongly accept the null hypothesis
| probabilities are too strict (p<0.01, p<0.001)
164
implications of type 2 error
null hypothesis is wrongly accepted => think there's no effect when there actually was
165
type 1 error
```
wrongly accept the alternate hypothesis
lenient probability (p<0.01/0.3)
```
166
implications of type 1 error
wrongly accept alternate hypothesis when null should be accepted
167
5 non-parametric hypothsis tests
1. mann Whitney-u
2. chi squared
3. binomial sign test
4. wilcoxon signed rank
5. spearman's rho
168
assumptions of parametric tests
- population should be ND
- var of pop should be approximately equal
- should have at least interval or ratio data
- no extreme scores
169
assumtions of non-parametric tests
- population isn't ND
- var of populations are unequal
- any level of data
- can incl extream scores
170
which test to use (checklist)
1. type of data
2. experimental design
3. difference in conditions
4. relationship or not
171
what does the mann Whitney-U measure
difference of 2 conditions of an IV
| ordinal or interval
172
checklist for mann Whitney-u
- DV produces ordinal or interval data
2. independent measures design
difference between each condition
173
equations used in mann whitney-u
```
Ua = NaNb+((Na(Na +1)÷2)-Ra
Ub = (NaNb)-Ua
```
174
steps for mann whitney-U
1. rank scores as one data set
2. add ranks for each group => Ra and Rb
3. Use formula to find Ua and Ub (Ub = observed value)
4. critical U values table (no. ppts in A on one axis B on other)
175
when is research significant in mann whitney-u
research is significant if observed value of U is EQUAL OR LESS than cv, at 5% sig level
[U = cv or U < cv] =>significant
176
when can null hypothesis be rejected
if U ≤ / < CV => null hypothesis is rejected
U > CV => null hyp can be accepted
177
experimental design used in wilcoxon signed rank
repeated measures
178
checklist for wilcoxon signed rank
1. DV produces ordinal or interval data
2. rm design
3. looking for difference between each condition
179
steps for wilcoxon signed rank
1. find difference between groups (add + or - )
2. rank differences (ignore signs and zeros)
3. count number of + & - separately
4. add ranks of less frequent sign (= observed value)
5. add number of differences (ignore 0's)
6. match n value on critical value table at 5% sig level, 2-tailed
180
when are results significant in wilcoxon signed rank
observed value should be LOWER than cv for significant results
- obs < cv =>reject ho
- obs > cv => accept ho
181
what does chi squared test measure
compares frequencies with occurring frequencies
182
checklist for chi squared test
1. DV produces nominal data
2. independent measures design
3. explores a difference between each condition / association
183
steps for chi squared test
1. add total for columns and rows
2. find expected values ((row T x columnT)÷overall T)
3. calculate chi squared using formula
4. calc DoF => (rows-1) x (columns-1)
5. find cv at 0.05 sig level using DoF
184
when is results significant for chi squared test
observed chi squared value is BIGGER or EQUAL TO cv
obs ≥/=cv =>significant, reject ho
obs < cv =>insignificant, accept ho
185
checklist for binomial test
1. dv produces nominal data
2. repeated measures design
3. exploring a difference between each condition
186
steps for a binomial sign test
1. add + &- ignore 0's
2. count each + and - (T+, T-)
3. smallest T value = observed
4. level of sig, no. of ppts (ignore 0's)
5. 0.05 for 1 tailed
187
when is binomial test significant
when observed value is SMALLER or EQUAL to the cv (T value)
obs significant, reject ho
obs > cv => not significant, accept ho
188
what does spearmans rho look at
examine the relationship between co-variables
| uses ranks
189
checklist for spearmans rho
- variables that produce ordinal data
- explores relationship between co-variables
- correlational design
190
spearmans rho steps
1. rank each column individually
2. rank scores for each group separately
3. find difference between ranks for each data set
5. square differences and add them
6. use formula to calc spearmans rho
191
when are results for spearmans rho significant
when calculated spearmans rho value is BIGGER than the critical value, results are significant
obs > cv =>significant, reject ho
obs insignificant, accept ho
