Study design and data analysis

"You canít fix by analysis what you bungled by design."

In spite of its title, this section - and our Boot Camps - have mostly been about data analysis. But too often people come along with data they have collected, and what can be got out of it is limited by the study design.

We have decided to focus much more on the design of wildlife studies, and to cut down on the time spent on analysis.


The quotation above is from: Light, Singer and Willet (1990) By design: planning research in higher education, Harvard University, Boston MA.

The icon indicates material which we only include in the basic Boot Camp if there is sufficient interest - and time! Topics without hyperlinks are still being worked on. Some of the materials still need more work and trial runs.


Rationale
Why is a specialist course (or web site) on wildlife study design and data analysis needed?

  

Teaching approach

"Many of our biology students are refugees from high-school mathematics." - John Ollason

We try to get away from sit-and-listen lectures and include lots of practical activities, games and exercises.


R software
We recommend R and use it in most of our workshops.

Details of other software packages we use are here.


Inferences based on samples
We can rarely measure the whole population or area, so we take a sample and use this to make inferences about the whole. Our inferences will not exactly equal the values for the population, due to random errors or noise introduced by sampling.


Models and model selection
This is a key topic and underlies most approaches to detection probability


Study design
Without good study design, your data may turn out to be of little use. The design stage of a research project is usually skimped as people rush out to start collecting data.

Click to go to the study design page


Bayesian approaches
These methods quantify the uncertainty about estimates and produce information which can be fed into decision-making processes for management.


Checking quadrats on the lawn for ant occupancy. Estimating occupancy
Simple 'presence/absence' data can give useful results, if corrected for detection probability

You may also like to look at the exercises on occupancy at U of Vermont.


Estimating animal densities from mark-recapture data
Density can be estimated from mark-recapture data, provided the data include information on the location of capture. Methods for analyzing Spatially Explicit Capture-Recapture data (SECR) have been developed in the last few years.


Estimating animal survival with mark-recapture methods
Mark-recapture studies spread over several years allow us to study survival and other demographic parameters of populations.


 

Jump to:
 
Next Boot Camps:

Boot Camps are planned in Phnom Penh in March and in Bangkok in July. More...

Special topics:
 
 
 

 


Creative Commons LicenseContent on this site, except for photographs, is licensed under a Creative Commons Attribution 3.0 License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author is acknowledged.
wcsmalaysia.org home Text by Mike Meredith, updated 2 Feb 2015