Final tasks and exercises

Day 9

Felix May

Freie Universität Berlin @ Theoretical Ecology

Agenda

1. Information on the final report

2. Exercises on linear and generalized linear models

  • Same structure and type of tasks as final report

3. Feedback on the course

4. Distribution of final tasks

Your independent tasks and report

  • 2 people per group (1 or 3 in exceptions)
  • In English
  • Ca. 3 pages of text plus graphics (max. 4) and tables (max. 2)
  • Do not include tables of raw data
  • Add captions to your graphics and tables
  • Do graphics in R (of course!)
  • Give R version in text, ask R how to cite: citation()
  • Describe methods universally and not in R specific language

Tip

Read some current biological articles and do it the same way (just shorter)

Structure of the report

  • Introduction:
    • Short
    • Describe what you know about the context of the data
    • Explain research question(s) and/or hypotheses
  • Methods:
    • Describe what you know about the collection of the data
    • Mainly describe statistical methods
  • Results:
    • Explain what you found
    • Rather little text, refer to figures and/or tables!

Structure of the report

  • Discussion:
    • Interpret the results with respect to the questions/hypotheses
    • Potentially compare your findings with studies you know
  • Literature
    • Cite all references that you used
  • R code:
    • Hand in, too!
    • With comprehensive comments explaining what you did

Structure of the report

  • Declaration that you did the analysis and report by yourself and without the help of generative AI tools
    • We will circulate a template for the declaration
  • Upload a pdf or word file and the R code (*.R-file) in Blackboard
  • Check deadline on Blackboard: 3.5 weeks from now

Hints for Methods section:

Important points when you describe the data and statistical analyses:

  • What is the response variable?
  • What are the potential predictor variables?
  • What is the sample size?
  • Which models and/or which tests did you use?
  • In case of GLMs:
    • Which error distribution did you use and why?
    • Which link function was used?
    • Did you check for overdispersion?
    • If you detected overdispersion, how did you deal with it?

Criteria for grading

Introduction (10%)

  • Short motivation of the research question
  • Goals of the study
  • Clear questions and/or hypotheses

Methods (40%)

  • Short introduction of the data, and the models including a description of error distribution etc.
  • Reproducible R code (we should be able to run the code)
  • Agreement between R code and its description in the text
  • Sound methods corresponding to the data and their distribution

Criteria for grading

Results (40%)

  • Precise description of the results
  • Proper graphical visualisation
  • Are all relevant results (in terms of the research question) presented?

Discussion (10%)

  • Biological interpretation of the results
  • Link to the research question
  • Link to the context of the study

Select the appropriate model type

Practicing data analysis

  • Choose one (or both) of the following exercises
  • Analyse the data as you would do it for the exam
  • Create comprehensive figures
  • Be ready to present your analyses in the seminar
  • Can be done in RStudio directly, no need to prepare slides (but feel free to do so when you have time and motivation)
  • Bring general questions on data analysis for the seminar!

Graphics first

Recommendation: Always plot your data, before you start modelling!

Tasks & exercises