statistical testing Flashcards
(22 cards)
what does a low standard deviation indicate
greater precision in data collection
whats the formula for standard deviation and what does each bit mean
S= square root of= E(x_x-)^2 over n-1
x= individual value
x- = the mean value
n= number of data points
how many data points lie within one and two standard deviations of the mean
68% lie within one standard deviation
95% lie within two standard deviations
what is the spread
the mean +- standard deviation
what is data beyond two standard deviations classified as
anomalous
what can u use standard deviation for on a graph
error bars
what are error bars
-bars which are on either a bar or line graph
-they extend 1st above and below your mean
-the longer the bar, the more spread out data is from the mean
what is the t-test used for
identifying if there is a significant difference between two means
what does the t-test begin with
a null hypothesis that can later be accepted/rejected
what’s an unpaired t-test used for
if the data is not paired i.e. it doesn’t come from the same individual
what can you not say if you haven’t done a stats test
that there’s a “significant” difference
what do the different parts in an unpaired t-test mean
x- is the mean
s is the standard deviation
n is the number of data sets
the brackets are straight to show if there is a negative number u should ignore
steps to doing an unpaired t-test
- calculate standard deviation for each data set
- use the formula to calculate the t-value
- compare your t value to the value in the table of critical values using the correct degree of freedom
how to find the degree of freedom
sample size (of both data sets) - number of data sets (e.g. if A and B then its two)
when do u reject the null hypothesis
if t> critical value (t test)
when do you accept the null hypothesis
if t< critical value (t test)
what does the 5% significance level actually mean
it means that the apparent difference between the two data sets is down to pure chance only 5% of the time
what could you do to reach a more powerful conclusion
use a 1% or 0.5% significance level
what does each part of the pared t test stand for
d- is the mean of differences
root n is square root of number of individuals
s is standard deviation
when is the paired T-test used
if pairs of measurements are collected from the same individual
steps of doing a paired T test
1) calculate the mean of differences in before and after data
2) calculate the standard deviation of the data
3) use the t-test formulae to calculate the T value
4) compare the value to one from critical value table using correct degree of freedom
degree of freedom for paired T test
number of categories (e.g. number of individuals) minus one
