Interpreting Data Flashcards
What are the main 2 types of data?
- qualitative
- quantitative
What are the 2 types of quantitative data?
- discrete (involves whole numbers)
- continuous (information measured on a continuum/scale)
What are the 3 measures of location?
- median (number in the middle of a set of numbers, used when there are large outliers or skewed data)
- mode (most common number)
- mean (average)
What are the 2 measures of spread?
- standard deviation
- interquartile range (when large outliers or skewed date)
What is Gaussian distribution?
Normal distribution
No outliers
Use mean and standard deviation
How does the Gaussian distribution change if the mean and standard deviation change?
Mean changes = graph moves left or right
S.d. changes = graph moves up or down (if it decreases graph increases) but area under line remains the same
What is a standard error?
Measure of statistical accuracy of an estimate
Standard deviation of the distribution of all possible sample means
Standard error of the mean = standard deviation/√sample size
How is the confidence interval ranges calculated?
Sample mean ± 1.96 x standard error
- get 2 values
What does confidence interval mean?
Expect the confidence interval amount of samples to be between the calculated range
- as sample size increases confidence interval gets narrower but range stays the same
What is the correlation coefficient?
- between -1 and 1
- 1 = perfect positive correlation
- -1 = perfect negative correlation
What is the significance of linear regression?
Y = a + bx Y - outcome (dependent variable) x - predictor (independent variable) b - slope a - intercept
What is statistical significance?
observed sample difference between groups may be due to chance
- statistically significant means unlikely due to chance
- determined by p value and confidence interval
How is the confidence interval calculated?
mean difference ± 1.96 x SE of mean difference
What is the confidence interval?
- mean difference and SE mean difference between 2 groups
What is the p value?
- probably of observing a result as or more extreme of the sample result if the underlying assumption in the population is true
- if greater than 0.05 (5%) cannot rule out chance effect
- if less than 0.05 probably not chance
How do you calculate standard deviation?
Average distance from the mean
(Value - mean)^2 for each, add up and divide by number of values
square the answer
What is the interquartile range?
Difference between 75th centile - 25th
- median is 50th
For a Gaussian distribution what are the reference ranges?
99% ranges lie within 2.58 SDs
95% ranges lie within 1.96 SDs from the mean
90% ranges lie within 1.64 SDs from the mean
When do we use standard deviation and when do we use standard error?
- s.d for ranges (individual values)
- SE for confidence intervals (means)
As sample size increases how does the range and confidence interval change?
Range stays the same
Confidence interval gets narrower
How do we determine whether an observed difference was due to chance?
Look at confidence intervals and p values
What are the 2 types of qualitative data?
- nominal (unordered)
- ordinal (ordered)
What are the 2 types of nominal data?
- binary (dead or alive)
- categorical (colours)
Define reference ranges
A constant proportion of values lying within any specified number of standard deviations above or below the mean
How do you calculate reference range?
mean ± 1.96 x standard deviation
What is the relationship between confidence intervals and p values?
If 99% CI for a difference excludes 0 then p<0.01, includes 0 then p ≥ 0.01
95% - excludes 0 then p<0.05, contains then p≥0.05
90% - excludes 0 then p<0.1, contains then p≥0.1