Research methods 3 Flashcards
2 types of data analysis
- Descriptive statistics
2) Inferential stats
quantitative data
numerical
qualitative data
descriptive
Primary data
Collected by reaearcher
Secondary data
collected by someone other than researcher ie: meta analyses
eval quantitative data
(+) objective
(-) lacks depth
qualitative data
(+) detailed
(-) Subjective interpretations
Eval primary data
(+) research can fit aim
(-) More time consuming
eval secondary data
(+) less expensive , less effort
(-) Research may not fully match the aims
(-) research may not be of high quality
Eval mean
(+)takes into account all data
(-) Affected by extreme values
Eval Median
(+) Not affected by extreme values
(-) doesn’t take all values into account
Eval Mode x4
(+) Not effected by extreme values
(+) useful for categorical data
(-)Doesn’t take all values into account
(-) sometimes there isn’t a mode
Eval standard deviation
(+) Not affected by extreme values
(-) Complicated to work out
What does the spread of data tell us
How consistent the data is
The smaller the spread, the more consistent = fewer individual difference s
Histogram
Used to present frequencies of continuous data
Bar chart
Used for non-continuous variables because bars are sperate from each other
Frequency polygon/ line graph
Used to show the frequencies of continuous data
Comment on mean and standard deviation question TEMPLATE
-PEE structure
-2 separate paragraphs
P - Which condition has a higher mean/standard deviation
Exa- Evidence from data. State which number is higher
Expl- What does this mean in the context of the study
3 types of (normal) Distributions of data
1) Normal distribution
2) Negative skew
3) positive skew
Normal distribution
-Symmetrical
-mean, median + mode in same place
-Bell shape
negative skew distribution
-Mode is high
-not symmetrical
Positive skew distribution
-Mode is low
-Not symmetrical
where is mean, mode median placed on a skewed graph
Mode placed at peak of graph
Mean placed in middle of graph
Median placed in between the two
Factors affecting the choice of stats test
1) Hypothesis- Difference or relationship ? (experiment or correlation)
2) Level of data - Nominal, ordinal, Interval/ratio
3) Research design - Related or unrelated?
Nominal data
Number of p.p falling into various categories
Not everyone gets a score
Ordinal data
Data can be ranked from lowest to highest
Each p.p gets a score
Measured on scales of unequal interval
(rating scales are ordinal)
Interval/Ratio data
Each p.p gets a score
Fixed equal intervals between the units of measurement on a continuous (established) Scale
Which research designs are Related or unrelated
Independant groups = unelated
Matched pairs = Related
Repeated measures = Related
How to conduct a sign test
1) Work out how many (+) , (-) , or (0) category by subtracting Condition B from A
2) Get rid of the zero (omit)
3) count up how many you have of each sign
4) ‘S’ is the sum of the less frequent sign
Stats tests for NOMINAL DATA
chi square
Sign test
Chi square
stats test for ORDINAL DATA
Mann Whitney
Wilcoxon
Spearman Rho
stats test for Interval data
Unrelated T test
Related T test
Pearson
stats test for UNRELATED design
Chi square
Mann whitney
Unrelated T test
stats test for Related design
Sign test
Wilcoxon
Related T test
stats test for correlation
Chi square
Spearman rho
Pearson
How to calculate degrees of freedom for chi square
(rows-1) x (columns-1)
Statements of Significance Template
-The calculated (observed) value of ….. (Symbol) was _________.
The critical value (p<___, n1=____ and n2=_____, 1/2 tailed test, df=___ )
is ___ .
-BEcause the calculated value was (greater/less than) the critical value, it was decided that the result was (significant/insignificant)
-Therefore, the ………. Hypothesis was (accepted/rejected)
-This means that….
Type 1 error
false positive, reject null when not true
More likely if the significance level is set too high. e.g: (p<0.10)
Type 2 error
False negtive, (accept nul when it’s false)
More likely if P is set too low e.g: (p<0.01)
Why do we use (p<0.05)
To reduce risk of type 1 and 2 errors
2 methods of analysing Qualitative data
1) content analysis
2) thematic analysis
Thematic analysis
- gather data from interviews, diaries or focus groups (which are flexible and open ended)
2)data transcribed
3) Analysis is attempted without any pre conceptions
4) the psychologist then codes the data by highlighting key themes that arise
5) Interpret meaning of the themes
Eval thematic analysis x3
(+) in depth + detail
(-) Time consuming
(-) subjective
