Statistical analysis

BASIC STATISTICAL ANALYSIS & GRAPHING
For Discrete Data*:
  • Use Fisher's exact test for data with a total sample size of less than 1000
  • Use Chi-square or G–test for data with a total sample size of greater than 1000
* Whole, concrete numbers with specific and fixed data values determined by counting

For Continuous Data:
  • When comparing two samples (e.g., control versus treated):
    • Use Student’s t-test for normal (parametric) distributions with equal variance then present as a bar graph
    • Use Mann-Whitney U Test for non-parametric analyses then present as a box plot
  • When comparing more than two samples (e.g., control, treatment A, treatment B, etc.):
    • Use One Way ANOVA followed by Tukey for normal (parametric) distributions with equal variance then present as a bar graph
    • Use One Way ANOVA followed by Kruskal-Wallis for non-parametric analyses then present as a box plot
  • When comparing samples with two variables (e.g., control for 1h & 24h, treatment A for 1h & 24h, treatment B for 1h & 24h, etc.):
    • Use Two Way ANOVA followed by Tukey for normal (parametric) distributions with equal variance then present as a grouped bar graph
    • Use Two Way ANOVA followed by Friedman's test for non-parametric analyses then present as a grouped box plot
*Watch these videos to learn how to do basic statistical analyses in Graph Pad
UP Diliman | UPD Institute of Biology
Copyright ©2016, University of the Philippines Diliman, Institute of Biology, Regenerative Biology Research Laboratory
​