Introduction to Experimental Design and Hypothesis testing
Why do we perform experiments? What conclusions would we like to be able to draw from these experiments? Who are we trying to convince? How does the “magic” of statistics help us reach conclusions?
As we attempt to make discoveries, we test new hypotheses using experimental data and hope that a skeptical and discerning person would start to believe the claims we make. Hypothesis testing, a branch of statistics, is a step we can rely on to bolster our claims.
In this 2-day workshop you will gain a fundamental understanding of common hypothesis-testing concepts and terms, such as null hypothesis, alternative hypothesis, type I error, type II error, p-value, and power.
No background in statistics is required.
This is an introductory workshop in the Biostats series . No prior experience required. No prerequisites.
- Stats and HT presentation and Experimental design presentations: We will go over these slides during this workshop.
- Code to go over: We will work through this code during the workshop.
- Code output: The output of the above code is generated in this file.
- Primer on testing: This is a word document detailing a flowchart guiding you towards simple hypothesis tests. This was prepared with Lennart Mucke and his lab.
Please download the materials for this workshop Please open Rstudio and install the following R packages copying and pasting the following commands:
install.packages("dplyr")
install.packages("onewaytests")
install.packages("pwr")
install.packages("ggplot2")
if (!require("BiocManager", quietly = TRUE)) install.packages("BiocManager")
BiocManager::install("multtest")
Check if the installation was successful copying and pasting the following commands to upload the libraries in R:
library(dplyr)
library(onewaytests)
library(pwr)
library(ggplot2)
library(multtest)
You can access these materials remotely at any time and go through them at your own pace. Here's how:
- Go over the slides.
- Work through the code in the R markdown file. You should be able to open it using RStudio installed on your computer. You will need to install the dplyr, onewaytests, pwr, ggplot2, multtest packages to be able to run all the lines of the code. Please refer to the workshop instructions above.
- As and when you need to test different hypotheses in your research and need guidance, try and consult the primer
- Email Reuben or Michela if you have any questions.