Four easy tips for preparing a field experiment (Part I)

Preparing a successful field project can be boiled down into two major themes, a good experimental design and organization of time and materials. To keep things short and sweet I will go over the experimental design here (Part 1) and go over tips of time and materials in a sequel blog ;). In a nutshell, the experimental design is a research project’s blueprints, it provides the foundation of how experimental units are arrayed between treatments, spatially and temporally. The goal of any experimental design is to optimize the probability of addressing one’s specific research questions with statistical confidence. That said, a big deal that revolves around the development of  experimental design is having the adequate treatments established and sample sizes big enough to capture the variation in the data. Deciding what are the appropriate treatments and sample sizes within treatment for any study will obviously vary depending on the researcher’s main question and level of interest (e.g., do you care more about the species level or population level?) As such, I cannot give any detailed tips on what your experiment should look like but I can provide some broad advice that has proven helpful for me and may be helpful to follow when brainstorming your experiment! 

First tip, outline your research question and hypotheses in words. Second tip, sketch it out! What do you expect the outcome of your data to look like? Draw out the predictor and response variables and if you have different grouping levels (e.g., species type) and/or treatments think about ways of representing these factors. Because ecology and evolutionary biology data is generally complicated and messy, hypothetical figures of your results can potentially make it easier to decide on how you set up your experiment (see my doodle below!). I would be remiss, however, if I didn’t encourage you to read the literature well because chances are someone has had similar questions as you and/or similar study system etc., and in that case it is always a good idea to figure out what treatments were needed to address a particular question and a ballpark estimate of a decent sample size within each treatment. So, tip number three is do your literature review in support of your treatments and sample sizes. That said, a good thing to keep your eye on when doing your literature review is examining summary statistics that indicate more or less how variable the data is for the variables you are interested in examining such as the standard error and sum of squares. My fourth and final tip, which has been somewhat useful for me, particularly in the field setting, is have extra samples if you can afford it because you never know if there happens to be a couple of hungry voles near some your experimental plants …