eModule 2 Flashcards
What are the three types of data?
- Quantitative
- Categorical
- Qualitative
What are the two types of categorical data?
Nominal (named) and Ordinal (ranked)
What are the two types of quantitative data?
Continuous and discrete
Define normality
A measure of central tendency. 2/3rd of the data is 1s.d. above ad below the mean and 50% of the data is above the mean and 50% below
Describe Gaussian distribution
Symmetric, central distribution with well-behaved tails. mean=median=mode
Describe skewness
asymmetrical distribution in which there are many data points at the high/ low end. Long, uneven tails
Describe Kurtosis
Data with heavy or light tails tails relative to those in a normal distribution
Normality can be assessed visually using which two graphs?
Histograms and Box and Whisker plots
Which stats tests can you use to assess normality? and when would you use them?
Shapiro-Wilks test (n<50)
Kolmogorov-Smirnov test (n>50)
Name three measures of central tendency
- Mean
- Median
- Mode
Name five measures of dispersion of data
- Standard deviation
- Variance
- Range
- SEM
- Confidence intervals
How is SEM calculated and what does it describe?
SD/ square root of n
Measures how well a sample mean approximates a population mean
How are Confidence intervals calculated and what does it describe?
SEM x 1.96
The range in which the true mean value might be found
When is data dependent (paired)?
Longitudinal experiments
When the same variables are measured in the same subjects at different time points.
When is data independent (unpaired)?
cross-sectional studies
When two groups are being compared with no common factors.
What two components are obtained from a person’s correlation and what do they indicate?
- an r-value: person’s correlation coefficient indicates strength of correlation
- a p- value: indicates the probability that the observed relation was due to chance
What are the two components of the Spearman’s correlation (used when data is not normally distributed)?
Spearman’s rho
p-value
Define correlation
indicates the strength of the relationship between two variables
Define regression
Quantifies the association between two variables. It tells us the impact that changing one variable will have on another variable. It is defined by y=a +bx
Why are parametric stats used over non-parametric? (3)
· Easier to understand
· Analyses are more powerful
· Less likely to incorrectly reject/fail to reject the hypothesis
When data is parametric which measure of central tendency is used?
Mean
When data is non-parametric which measure of central tendency is used?
Median
Rank the 7 different types of evidence from strongest to weakest
- Systematic reviews and meta-analyses
- Randomised control trials
- Well designed control trials
- Case control studies
- Systematic reviews of descriptive studies
- Descriptive cross-sectional studies
- Case report/series