NBIK14016U Experimental Design and Statistical Methods in Biology (StatBio)

Volume 2014/2015
Education
MSc Programme in Biology
Content

The course is intended to give a broad overview of experimental designs and statistical methods in order for students to plan their own experiments and to analyze existing data. Students are recommended to follow the course shortly before they start their bachelor project, their M.Sc. thesis project, or as a part of a Ph.D. study programme.

 

Learning Outcome

Knowledge:

The student will get an overview a range of statistical concepts and tools:

  • Principles in parametric statistics.
  • General linear statistical models.
  • Qualitative and quantitative variables.
  • Interaction between factors.
  • Block design.
  • Factorial design.
  • Polynomial regression.
  • Multiple regression.
  • Analysis of variance and covariance
  • Models with systematic and random effects.
  • Nested analysis of variance.
  • Estimation of parameters in general linear models.
  • Confidence limits of the parameters.
  • Prerequisites of analysis of variance.
  • Transformation of data.
  • Multiple comparisons of means.
  • Repeated measures ANOVA.
  • Multivariate ANOVA.
  • Data presentation.
  • Logistical regression.
  • Log-linear models.
  • Use of statistical software (SAS).


Skills:

A student that has successfully finished the course will possess the following qualifications:

  • be able to select the appropriate statistical model for the design in question.
  • be able to apply General Linear Models (analysis of variance (ANOVA), analysis of covariance (ancova), polynomial and multiple regression, nested and mixed anovas)
  • be able to describe multivariate anova (manova), repeated measurements anova, logistic regression and log-linear models.
  • be able to develop and apply statistical models that incorporate qualitative (both fixed and random effects) and quantitative variables, main effects, interactions, and second or higher order terms.
  • be able to use statistical software (currently SAS) to load a data set, to sort and summarize data, to perform relevant statistical analyses, and to report the results either graphically or in tables.
  • be able to estimate the parameters of a statistical model and their standard errors, and to test whether they are significantly different from 0.
  • be able to identify significant and non-significant factors so as to simplify the statistical model using various criteria for best fit.
  • be able to apply a priori and a posteriori tests to identify treatment differences.
  • be able to use the statistical model as a predictive tool to forecast the expected outcome of an observation from a set of independent variables.
  • be able to check whether data meet the assumptions of the model and, if needed, to select an appropriate data transformation.

 

Competences:

The student will learn the most commonly used experimental designs and appreciate their advantages with respect to the subsequent statistical analysis of data. The student will be able to select or, if necessary, to develop a statistical model for the experimental design, state the relevant statistical hypotheses, conduct the statistical analysis (generally using statistical software), present the results in a clear and understandable way, and finally interpret the results in a biological context to reach a sound conclusion based on the empirical evidence. In addition, the student should possess the necessary theoretical insight in statistics to be able to understand and comment critically on the use of statistics by others.

See Absalon.

The students are assumed to possess a basic knowledge of statistics at a level corresponding to at least “Matematik/Statistik” (1st year of bachelor study). The time schedule does not allow for repetition of basic statistics, so students lacking an up-to-date knowledge are requested to refresh fundamental statistical concepts and principles prior to the course. Although the course puts emphasis on applying statistics, it is unavoidable that some theory will be encountered, so students with poor mathematical skills should consider whether the course fulfils their needs.
Lectures: 36 hours (2+2 hours per week for 9 weeks, of which c. 1 hour per week is student presentations). Computer exercises: 27 hours (3 hours per week for 9 weeks).
The students are expected to do considerable homework.
The course is part of the qualification profiles 'Ecology and Evolution', 'Microbiology' and 'Ecosystem Functioning and Management'.
  • Category
  • Hours
  • Colloquia
  • 12
  • Lectures
  • 24
  • Practical exercises
  • 27
  • Preparation
  • 129
  • Project work
  • 14
  • Total
  • 206
Credit
7,5 ECTS
Type of assessment
Continuous assessment
During the course the student has to give a short (10-15 minutes) individual presentation of a case study that has involved statistical analysis. The study may either be taken from literature or be his/her own project.

In order to pass, the student should have participated actively in the course by being present at both the lectures and the exercises (max 20% absence from the exercises).

Furthermore, the students should (in groups of 2 or 3) have solved and handed in three homework exercises of an acceptable standard. A homework exercise is accepted (equivalent to grade 2) if the group is able to develop a reasonable statistical model that describes the experiment, to formulate the appropriate hypotheses, to identify the significant factors behind the data, to report the results, and to derive at a convincing conclusion including a biological interpretation. The model and the analyses need not be perfect, but the results should agree with those obtained from an analysis that exploits data optimally.
Marking scale
passed/not passed
Censorship form
No external censorship
Several internal examiners
Re-exam
Renewed hand over of home work exercises and new presentation of case study.
Criteria for exam assesment

See Learning Outcome