Renewable Resources 480 Experimental Design & Data Analysis in - - PowerPoint PPT Presentation

Renewable Resources 480

Experimental Design & Data Analysis in Environmental Sciences

“Univariate Statistics and Data Management with R”

Laura Gray

Post-doctoral Fellow

www.ualberta.ca/~lkgray BSc in Mathematics, Statistics and Environmental Studies (UVIC, 2005) MSc in Forest Biology and Management (UofA, 2008) PhD in Forest Biology and Management (UofA, 2011)

Designated Professional Statistician with The Statistical Society of Canada Research: Adaptation, mitigation, and risk management for forest resources under climate change

Something about me…

Stefan Schreiber

Post-doctoral Fellow

David Roberts

Post-doctoral Fellow

www.ualberta.ca/~drr3

Gabriel Oltean

Master’s Student

Other people you should know…

Now let’s hear about you…

How do we understand/combine these things? How do we use these tools to conduct solid scientific research?

Academic Integrity

• 1. Conduct research well and with honesty
• 2. Give credit to other people′s work

University of Alberta Policy

http://www.ualberta.ca/TIE

I’m required to report any academic misconduct to the dean and the associate chair of grad studies. Even if there is a minimal penalty for a minor infraction, it leaves a record in your file. Just avoid this.

30.3.2(1) Plagiarism: No Student shall submit the words, ideas, images or data of another person as the Student’s own in any academic writing, essay, thesis, project, assignment, presentation or poster in a course or program of study.

Plagiarism

If you didn’t write it, don’t submit it without a reference

Examples:

Frowned-upon, but forgivable Copy a few generic sentences (no citation) Copy key sentences/phrases (with citation but without quotation marks) Self-plagiarism (with citation) Scientific Misconduct Copy and paste of major sections (with or without citation, with or without modification) Key content/phrases/ideas without citation (copied or even in your own words) Strict definition: 7 words in a row Practical definition: I can find the source (google, turnitin, desktop search) Applies to ALL Published and Unpublished work!

30.3.2(2)Cheating: No Student shall in the course of an examination or other similar activity, obtain or attempt to obtain information from another Student or other unauthorized source, give or attempt to give information to another Student, or use, attempt to use or possess for the purposes of use any unauthorized material.

Cheating

Exams are an individual sport, not a team effort. 30.3.2(2) d Cheating: No Student shall submit in any course or program of study, without the written approval of the course Instructor, all or a substantial portion of any academic writing, essay, thesis, research report, project, assignment, presentation or poster for which credit has previously been obtained by the Student or which has been or is being submitted by the Student in another course or program of study in the University or elsewhere. You can only submit the same material ONCE. 30.3.2(2) e Cheating: No Student shall submit in any course or program of study any academic writing, essay, thesis, report, project, assignment, presentation or poster containing a statement of fact known by the Student to be false or a reference to a source the Student knows to contain fabricated claims (unless acknowledged by the Student), or a fabricated reference to a source. Don’t make up your research. – With an exception Academic suicide

How do we understand/combine these things? How do we use these tools to conduct solid scientific research?

Criteria for Good Research Questions

• 1. You don’t already know the answer
• 2. If you are wrong, still interesting

Cory

RenR student 2010

• MSc. Student

Hydrology

Cory’s research: Movement of E.Coli through soil

What if testing was flawed?

Hypothesis: smooth surfaced cells migrate faster through soil

• > No initial detection

Cory’s research: Movement of E.Coli through soil

Smooth surfaced cells Rough surfaced cells

What about applied science?

• Effect of fertilizer on yield
• Effect of silviculture on forest growth
• Effect of medical treatment

Scientific hypothesis ≠ Statistical hypothesis

Research Question What you design your study around “Bigger picture” Used to make inferences on research questions though statistical tests Multiple within research Accepting or rejecting does not necessarily provide support for the scientific hypothesis tested

Scientific hypothesis ≠ Statistical hypothesis

Research Question What you design your study around “Bigger picture” Used to make inferences on research questions though statistical tests Multiple within research Accepting or rejecting does not necessarily provide support for the scientific hypothesis tested

Example from Cory’s research

• E. Coli contamination at Walkerton

Lakes ON in 2000 was unpreventable due to poor detection Smooth surface E.coli cells move faster through soil than rough surface cells preventing detection from standard testing methods

What can statistical tests do?

• 1. Answer if there is a

difference/relationship between groups

• 2. Answer if the observed

difference/relationship is due to random chance

Statistical techniques for this class

Parametric tests – e.g. t-test, ANOVA, etc. Non-parametric tests – tests for data that does not meet parametric

assumptions

Tests for binary data – tests for yes/no or proportions data Regression – linear, non-linear, multivariate Experimental design & simple principles

Advanced design RenR 580 Multivariate stats RenR 690

Project Data Options

1. Ideally data collected as part of your thesis work – get a chapter analyzed while you’re here

• 2. Talk with your supervisor about another project you can

analyze – they ALWAYS have data lying around

• 3. “Generate data” (this will be acknowledged upfront) –

learn how to analyze data similar to what you expect from your thesis research We will be using this in the labs as well to practice techniques.