Interactive web application for performing one- and two-sample
statistical tests developed by the
Bioinformatics Core
at the Cancer Research UK Cambridge Institute for the
Introduction to Statistics
training course.
Upload a tabular data file or select one of the sample datasets from
the drop-down list and click on either the 'One sample test' or
'Two sample test' tabs above to select variables of interest, explore
and visualize the selected data, and carry out statistical tests.
The Q-Q (quantile-quantile) plot compares the data with a normal distribution by
plotting their quantiles against each other.
The theoretical quantiles are for a standard normal distribution with mean 0 and
standard deviation 1.
The points will lie approximately along the diagonal line (fitted through the
points for the first and third quartiles) if the data are normally distributed.
Significant deviations from the line may suggest the use of a non-parametric
test.
Shapiro-Wilk test of normality
The Shapiro-Wilk test tests the null hypothesis that the data come from a normally distributed population. The null hypothesis can be rejected if the p-value is less than 0.05, suggesting that the data come from a population that are not normally distributed. If the null hypothesis can't be rejected, this means there is insufficient evidence that the data are not normal. This is not the same as accepting that the data come from a normal distribution, i.e. it does not prove that the null hypothesis is true. Caution is advised when using a preliminary test for normality to decide whether a parametric or non-parametric test should subsequently be used, particularly when the sample size is small. It is often better to make your own assessment by looking at box plots, density plots, histograms and Q-Q plots.One sample t-test
Wilcoxon signed rank test
Tests whether the data come from a symmetric population centred around a specified median value.
The Q-Q (quantile-quantile) plot compares the data with a normal distribution by
plotting their quantiles against each other.
The theoretical quantiles are for a standard normal distribution with mean 0 and
standard deviation 1.
The points will lie approximately along the diagonal line (fitted through the
points for the first and third quartiles) if the data are normally distributed.
Significant deviations from the line may suggest the use of a non-parametric
test.
Shapiro-Wilk test of normality
The Shapiro-Wilk test tests the null hypothesis that the data come from a normally distributed population. The test is run for each of the two groups separately.F-test to compare two variances
The F-test of equality of variances is a test for the null hypothesis that two normal populations have the same variance. The Welch t-test, an adaptation of Student's t-test, may be more reliable when the samples have unequal variances or sample sizes. The null hypothesis can be rejected if the p-value is less than 0.05, suggesting that the data come from populations with different variance. Treat the result of this test with caution; it is often better assess differences in variance between the two groups by inspecting the box and density plots.
The Q-Q (quantile-quantile) plot compares the data with a normal distribution by
plotting their quantiles against each other. In the paired two sample case,
quantiles for the differences between pairs of measurements are used.
The theoretical quantiles are for a standard normal distribution with mean 0 and
standard deviation 1.
The points will lie approximately along the diagonal line (fitted through the
points for the first and third quartiles) if the data are normally distributed.
Significant deviations from the line may suggest the use of a non-parametric
test.
Shapiro-Wilk test of normality
The Shapiro-Wilk test tests the null hypothesis that the data come from a normally distributed population. In the paired two sample case, the test is run on the differences between pairs of measurements. The null hypothesis can be rejected if the p-value is less than 0.05, suggesting that the data come from a population that are not normally distributed. If the null hypothesis can't be rejected, this means there is insufficient evidence that the data are not normal. This is not the same as accepting that the data come from a normal distribution, i.e. it does not prove that the null hypothesis is true. Caution is advised when using a preliminary test for normality to decide whether a parametric or non-parametric test should subsequently be used, particularly when the sample size is small. It is often better to make your own assessment by looking at box plots, density plots, histograms and Q-Q plots.Two sample t-test
Welch two sample t-test
Wilcoxon rank sum test
Also known as the Mann-Whitney U test.Paired t-test
Wilcoxon signed rank test
