Statistics Flashcards
Descriptive statistics
used to describe/summarise the characteristics of a sample or set of data
- such as variables, mean, standard deviation or frequency
MEAN
MoCT
> most sensitive measure- exact distance between all values of data
X distorted by extreme values- less representative
X not NOMINAL data or discrete value (1.7 legs)
MEDIAN
MoCT
- middle value ordered in list
> ORDINAL data can be used
> not affected by extreme scores
X not as sensitive- exact values not reflected because not all data is used
MODE
MoCT
- most frequent occurring value in a set of scores
> discrete data, NOMINAL data can be used, not affected by extreme values
X not useful when there are several modes
RANGE
MoD
- arithmetic distance between top and bottom values
- customary to ADD 1
> identifies how spread out the data is- accounts for end values of data set
> extreme results, can’t see the distribution between numbers
STANDARD DEVIATION
MoD
- measures the spread out of the set of numbers, in effect the average distance of each number from the mean
> precise measure of dispersion
> all values taken into account
> know if the numbers are closely grouped or spread out
> not as affected by extreme results
X may hide characteristics- extreme values- indicates how far each value id rather than looking at values at each end of a data set
NORMAL distribution
BELL CURVE
- mean, median, mode are all in the exacts mid-point
- the distribution is symmetrical around this mid-point
SKEWED distribution
- the mode and the mean are NOT THE SAME SCORE
POSITIVE SKEW - few extreme scores affect mean
- mean always HIGHER than the median and mode
- most scores towards the low end
NEGATIVE SKEW - the mean is BELOW the median and mode
- if marks were plotted for an exam which was very easy so most people got a very high score
50= mean 80= mode
LEFT FOOT NEGATIVE RIGHT FOOT POSITIVE
All graphs need
- a title
- y axis title (DV)
- x axis title (IV)
- plot of graph
Histogram
- data should start from 0
- bars should be touching
- IV is numerical = CONTINUOUS
> interval OR ratio
Line graph
- show information is connected
- change in performance over time
- date must be continuous
Pie chart
- shows FREQUENCY data
- each slice = proportion of category
- total pie is 100%
- NOMINAL data
Scatter diagram
- only for correlations
- paired numerical data want to determine whether 2 co-variables are related
- positive, negative, uncorrelated
Bar chart
- data on the x axis = NOT CONTINUOUS
- usually in categories
- spaces between each bar
- NOMINAL
Frequency table
- may be given table containing raw data
- or MoCT
- mean, median, mode
Inferential statistics
infers probability and whether results are due to chance
are the significant?
Probability
5% chance of error or external cause or chance
95% certainty of results due to real change
When would you use p<_0.01?
1) testing a new medication
2) challenging established research
Significant result
not due to chance
- accept alternative and reject null
non-significant result
due to chance or other cause
- accept null and reject alternative
Observed value GREATER than critical value
Chi-squaRed
SpeaRman’s Rho
Observed value LESS than critical value
- wilcoxon
- sign
- mann whitney u
How to calculate SD
1) calculate the mean of each data set
2) take away the mean from each value in the data set
3) square this number
4) add all of these numbers together- sum of squared differences
5) divide this by n-1 9number of Ps involved|)
6) square root your final answer
Abstract
- snapshot of important info
- overview of Ps, procedures, results, conclusions etc
Introduction
- overview of previous research
- start broad then funnel to your specific study
- aims and hypothesis should leas on logically
Method
- detailed description on what they did, enough detail another person could replicate so include conclusions
Results
- what you found
- qualitative or quantitative data
- quantitative analysed with descriptive or inferential statistics
Discussion
- conclude
- compare to previous research - does it support or refute?
- evaluate- methodological issues, solutions, give prompts for future research
Reference
- full detail/ titles of all journals and book references used in the report
Appendix
- supplements
- include anything to long to include in main body of text
-e.g raw data, material used
The peer review process
- scrutinised by other experts in that field
1) study/ write about results
2) journal editor receives
3) peers review article and provide feedback
4) if it doesn’t maintain scientific standards it is rejected
What is the purpose of a peer review?
1) ALLOCATION OF RESEARCH FUNDING
- spend responsibly, decide what is worthwhile
2) PUBLICATION IN RESEARCH JOURNALS & BOOKS
- prevent any incorrect or faulty data entering the public domain
3) ASSESSING RESEARCH RATINGS OF UNI DEPARTMENTS
- funding depends on good ratings from previous REF peer reviews
Advantages of peer reviews
> validity/accuracy
further funding to make more improvements
problems highlighted
upholds principles of science and prevents fraud#
adds credibility to the research and field of study
Disadvantages of peer reviews
X takes time - delays research publication
X hard to find an expert
X peers bring their own bias
X conflict of interest
X publication bias- ‘file drawer effect’ - nulls never published