HEL-8047 Statistical models, conclusions and uncertainty for scientific data analysis - 7 stp

The course will provide in-depth knowledge and firm skills in statistical analysis in three main blocks:

(1) Data analysis: choosing the appropriate model based on research question and data availability and quality.

(2) Statistical methods: perform correct statistical analysis, and draw valid conclusions.

(3) Communication: describing the methods chosen and the conclusions drawn for peer review publication.

The course is a PhD level for students in health sciences doing quantitative research, students in other sciences doing quantitative research on medium sized/large data sets, and students in applied natural and mathematical sciences with health-related applications, or others who want to obtain such knowledge and skills.

The course is an expansion of basic skills from master level and focuses on in-depth knowledge of data analysis and statistical methods for research purposes, where the data itself can be complex and noisy, the correct approach is not obvious, and the results are not as expected. As opposed to statistical methods taught at master level, where the data (possibly simulated) has been matched to the exercise, the research approach to statistical methods requires an in-depth understanding, which this course offers.

The main topics covered in depth are: probability, sample and distribution, assumptions, point estimators and confidence intervals, p-values, statistical significance, hypothesis testing, power, model fitting (linear regression), and interpretation of results. Pointers towards more advanced methods are given.

An introduction to R (RStudio) is given, and all exercises will have suggested solutions in R. It is therefore recommended for students to use R, although analyses can be carried out by the students' preferred tool. Students are expected to bring their own data set, and assignments and home exam are directed towards developing relevant parts of the students’ own manuscript drafts.

PhD students and students at the Student Research Programme, or holders of a Norwegian master´s degree of five years or 3+2 years (or equivalent) may be admitted.

External PhD students and students at other student research programmes, must upload a document from their university stating that they are registered students.

Knowledge:

• In-depth knowledge of common data analysis and statistical methods in quantitative research.
• Assess and rank the adequateness of different methods, depending on research questions and data availability and quality.
• In-depth knowledge of the requirements for the validity of statistical analysis results.

Skills:

• Articulate and evaluate research questions, evaluate data and data quality. Re-evaluate research question according to preliminary data analysis, and modify the hypothesis.
• Appraise adequacy of statistical methods, and conclude regarding the validity of the results.
• Plan and perform analysis, defend choice of method. Present and arrange analysis, results and their validity for peer-review publication. This includes so-called "null findings", i.e., not rejecting the null hypothesis.

Competence:

• Identify non-valid statistical analysis by violation of requirements, and identify alternative, correct approaches.
• Identify basic components in a research question with matching data set, to build more complex analysis schemes.
• Recognise own limitations in statistical analysis, and communicate needs in an adequate manner to more skilled experts for interdisciplinary research collaborations.

The course consists of weekly lectures and seminars for 14 consecutive weeks.

The lectures will cover the course contents, the skills will be acquired through teacher-led seminars with exercises and discussions. The students are expected to hand in answers to exercises each week. Two announced exercises (assignments) are mandatory, and passing is required for taking the exam at the end of the semester. Although attendance to lectures and seminars are not mandatory, students are strongly advised to attend.