Trend analysis and Environmental Impact Assessment course

• Start Date: September 28, 2010
• End Date: September 30, 2010
• City: Revelstoke BC
• Instructor: Dr. Carl Schwarz

This course was taught by Dr. Carl Schwarz from Simon Fraser University, to a maximum class size of 16 people. Carl has a wealth of information about statistics for biologists posted on his website at: http://www.stat.sfu.ca/~cschwarz/.

Course description

Environmental monitoring often looks at trends over time. Environmental impact assessments want to know if trends over time differ between control and impact sites. Statistical methods for the analysis of trends over time use many of the same methods as the analysis of experimental data (e.g. ANOVA, regression) but must now deal with problems such as autocorrelation and process error.

Course content

Review of statistical concepts on estimates, standard errors, confidence intervals, p-values, bias, precision, accuracy, missing values, etc.

Detecting step changes over time

• ANOVA and non-parametric methods
• Power and how many years to monitor
• Dealing with seasonality and autocorrelation
• Sampling vs process error

Detecting linear changes over time

• Regression and non-parametric methods
• Power and how many years to monitor
• Dealing with seaonality and autocorrelation
• Sampling vs process error
• Pooling slopes (ANCOVA)

BACI designs and variants

• Simple Before and After Controlled Impact design (BACI) – Introduction to two factor ANOVA
• BACI with multiple sites
• BACI with multiple years/multiple sites before/after
• Paired BACI with multiple years/multiple sites before/after
• Regression BACI (analysis of covariance)
• Power/sample size