Before we can start talking about different types of experimental designs – we all have to be on the same page as to what is an experimental design.

I often asked my students the following questions:

Do you collect data “willy-nilly” for your projects? Do you just go out and collect data that you THINK is relevant for your projects?

You all know what the answer is going to be! A big

NO!

But then – how do you decide how to collect your data? Do you collect it the same way your lab has done for years? Or maybe try a really cool way that someone mentioned in passing?

NO!

To try and add context to any research and experiment I consult on – I will always ask “What is your research question?” Everything in your project stems from this question. Yes! I mean everything. From this question, you know what data you need to collect, from this you also can determine what experimental design you should be using, and lo and behold, I will argue that you can set out your output tables for your reports, paper, and theses, since data type, experimental design, and research question – will dictate statistical analyses! Yup you read that correctly – before you have any data in hand, you can already start preparing your final products – all because you have that Research Question defined.

So let’s step back a bit and talk about these “experimental designs”. What are the primary aims of having an experimental design?

1. First and foremost to answer your research question
2. To explain as much of the experimental error as possible by:
   • Controlling as much variation as possible
   • Replicating to estimate that experimental error
   • Randomizing to provide a valid measure of experimental error

Steps to Designing an Experiment

I love steps and todo lists. Items that are “easy” to follow and maybe check off (provides a fabulous sense of accomplishment when you can check it off a list – right???). So here are 4 steps to consider when designing an experiment/.

1. Define the EXPERIMENTAL UNIT
   • Consider the following question: What is the difference between EXPERIMENTAL unit and SAMPLING unit?
2. Identify the types of variables you are collecting
3. Define the treatment structure
4. Design the overall structure

Rather than repeat what I have written on the AgStats Blog – I’m going to link to the appropriate post and build on it next time. Experimental Designs. I will look more closely at each experimental design over the coming months.

Review the 4 steps and we’ll dig into a Completely Randomized Design next time!