Statistics training II

Two-day intensive statistics course II

Current information on statistics training courses: see
www.uni-kassel.de/fb11agrar/de/fachgebiete-einrichtungen/umweltchemie/statistikschulungen.html

Statistical modeling and regression with R for soil scientists

Recommended prerequisite for participation: Basic knowledge of statistics and R (see contents of the two-day statistics intensive course I on March 9 and 10, especially a confident handling of vectors and data frames and reading data into R would be helpful)

In the media and journals of various disciplines, reference is repeatedly made to erroneous research due to inadequate statistical knowledge (see e.g. Ainsworth (2007, Nature 448, 849): "The abysmal standard of statistical analysis in much of genetic epidemiology is little short of scandalous". There may also be similar problems in the soil sciences.

A major potential problem in the soil sciences could be that some researchers are not even aware that there may be gaps in knowledge that could lead to incorrect evaluations. Problem areas may include (i) ignorance of the importance of residual checks, (ii) lack of understanding of the differences between a minimally adequate model and a maximally adequate model, (iii) ignorance of the differences between calibration, cross-validation and validation of a model & (iv) lack of knowledge of important special topics, such as performing Box-Cox transformations, polynomial regressions and model comparisons.

In the intensive course II, the statistical basics for soil scientists are to be further deepened so that the above-mentioned problem areas are less frequently encountered in soil sciences in the future.

Date: March 13, 2017 10:00 am to March 14, 2017 6:30 pm.

Location: University of Kassel, 37213 Witzenhausen, Nordbahnhofstr. 1a.

Costs: on request.

Number of participants: The number of participants is limited to 20.

Laptops: Please bring your own laptops with R and RStudio. If necessary, laptops can be borrowed by prior arrangement.

Materials: Course script, exercises and sample solutions will be provided.

Recommended literature: e.g. Crawley (2012). The R Book. 2nd Ed., Wiley.

Lecturer: Prof. Dr. Bernard Ludwig

Schedule of the intensive course:

Mon., 13.03.2017

10:00 - 12:00 Welcome, regressions and residual checks

12:00 - 13:00 Lunch break

13:00 - 14:00 Statistical modeling: model types & model simplification

14:00 - 16:30 Procedure for model simplification, exercises in R

16:30 - 17:00 Coffee break

17:00 - 18:30 Transformations (Sqrt, ln & Box-Cox) and exercises in R

Tue,14.03.2017

10:00 - 12:00 Dealing with variability and predictions (confidence & prediction bands; validation & cross-validation)

12:00 - 13:00 Lunch break

13:00 - 16:30 Exercises in R with practical examples

16:30 - 17:00 Coffee break

17:00 - 18:30 Polynomial & non-linear regressions and exercises in R with practical examples

Please send registrations and queries by 15.02.2017 to: Prof. Bernard Ludwig,University of Kassel, bludwig@uni-kassel.de

Please understand that the intensive course may be canceled if there are not enough registrations. In this case, or if the course is canceled due to illness of the lecturer or other events beyond our control, no further claims can be made.