XiaLab at University of Ottawa

Skip Navigation Links

BIO8102B: Biostatistics (Fall term)

Course description:

(This is now obsolete because the restructuring of biostats courses in Biology Department. This course will now cover mainly multivariate statistics, but with student-requested topics having priority)

 Below you will find an outline of topics intended to be covered by the Applied Statistics grad course. The main goal is to understand the rationale of these statistical analyses, to use SAS to carry out the relevant statistical analysis and to properly interpret SAS output.

However, what is the most important for a biologist is to analyze and re-analyze the biological problem before any analysis of the data (or ideally before the experiment and data collection.

R. A. Fisher once remarked that "To consult the statistician after an experiment is finished is often merely to ask him to conduct a post mortem examination. He can perhaps say what the experiment died of."

(I may not be able to cover all the topics listed below.)

  1. Statistics and decision making
  2. SAS software
  3. Statistical estimation and significance tests (descriptive and hypothesis-drive research)
  4. Probability and statistical distributions
  5. One-sample hypothesis, power and sample size in statistical tests
  6. Dependent and independent variables (DV and IV), continuous and categorical variables
  7. Statistics involving one set of continuous DVs and one set of categorical IVs
    1. Two-sample hypotheses (one continuous DV with one categorical IV of two categories)
    2. ANOVA (analysis of variance, one continuous DV with one or more categorical IVs of two or more categories), multiple comparisons, family-wise error rate and false discovery rate.
    3. MANOVA (multivariate analysis of variance, multiple continuous DVs with one or more categorical IVs of two or more categories)
  8. ANCOVA (analysis of covariance, one continuous DV with one or more categorical IV plus one or more continuous covariates)
  9. Data transformation and normalization
  10. Statistics involving two sets of continuous variables
    1. With one set is DV and another is IV
      1. One continuous DV and one continuous IV: simple linear and non-linear regression
      2. One continuous DV and multiple continuous IVs: multiple linear regression, polynomial regression
      3. Multiple continuous DVs and multiple continuous IVs: multivariate regression (optional)
      4. One binary DV and one or more continuous IVs: logistic regression.
    2. With no IV/DV designation
      1. Bivariate correlation
      2. One set contains a single variable, the other contains multiple variables: multiple correlation and partial correlation
      3. Both sets contains multiple variables: canonical correlation
  11. Other multivariate statistical methods
    1. Principal component analysis
    2. Correspondence analysis
    3. Cluster analysis
    4. Discriminant function analysis
  12. Data analysis of categorical variables
    1. Goodness of fit test
    2. Two-way contingency tables
    3. Log-linear models for multi-way contingency tables
  13. Statistical criteria for model selection

Reference book

  1. Sokal, R. R., Rohlf, F. J., 1995. Biometry. Freeman, New York. (Any edition of the book would be fine)
  2. Zar, J. H., 1999. Biostatistical analysis. Prentice Hall, Upper Saddle River, New Jersey. (Any edition of the book would be fine)
  3. Any textbook on multivariate statistics for biologists

Evaluation:

Continuous assessment with assignments and classroom quizes.

Time and location:

Check

Web page:

bio8102b.aspx

Instructor:

Dr. Xuhua Xia, Rm. 278, 30 Marie Curie.
E-mail: xxia@uottawa.ca
Phone: 562-5800 ext 6886
Office hour: Friday between 9:00 am - 12:00 pm

© 2016. XiaLab. All Rights Reserved.
27209