Designing & Analysing High Quality Experiments in the Social Sciences

*****************************************

Note re laptops:

Laptops are not required for this course, however this course will take place in a classroom that is conducive to laptops (Wifi, plugs, etc) so participants are welcome to bring one if they feel more comfortable taking notes this way.

*****************************************

 

 

This course is designed for researchers who want to design experiments that provide generalizable findings.  It is ideal for post-graduate students who are preparing proposal or programs of study for their degree.  It is also ideal for professional researchers who want to increase their quantitative analysis research skill set.

 

This course will introduce students to experiments and quasi-experiments in the social sciences.  The emphasis is on designing experimental studies to achieve the highest calibre results to answer specific research questions common to quantitative research.  Additionally, students will be introduced to the statistical analyses necessary for the data generated from an experimental design.  This particular course will also explain how to conduct the necessary analyses using SPSS.  Following from a brief review of introductory statistics, the nature of experiments and causal inference will be discussed.  This will include a discussion of common threats to validity and methods to assess the quality of the experimental (or quasi-experimental) design.  Along with discussing true and quasi-experiments, this course will also address issues related to multiple regression.  The focus will be on the types of research questions that can be answered with multiple regression analysis, and how these forms of observational/correlational studies are distinct from true experiments.  Specific statistical analyses will be introduced, including 1-way, 2-way and factorial ANOVA with fixed, random and mixed effects; multiple comparison protocols for post hoc analyses; multiple regression for continuous and categorical predictors; ANCOVA; repeated measures ANOVA; and hierarchical or nested ANOVA designs.

 

Course Outline:

Day 1

     a.  Review of introductory statistics (histograms, central limit theorem, t-tests and p-values)

     b.  Causation and Experiments (revisit the independent measures t-test)

     c.  ANOVA:  the theory and notation, how it relates to experimental design

     d.  ANOVA:  using SPSS and interpreting the output

Day 2

     a.  Threats to validity

     b.  Quasi-experiments:  Locating potential threats to validity

     c.  Accounting for possible threats in quasi-experiments:  ANOVAs and t-tests

     d.  Introduction to multiple comparisons (planned vs. post hoc)

Day 3

     a.  2-way ANOVA:  introduction and interpretation (main effects and interactions)

     b.  Factorial ANOVA:  using SPSS and interpreting output

     c.  Random effects vs. Fixed effects

     d.  Mixed effects:  When SPSS “drops the ball”

Day 4

     a.  Multiple regression:  What it is and why it is not an experiment?

     b.  Multiple regression:  Research questions, SPSS and relationship to ANOVA

     c.  ANCOVA:  Measuring effects with validity threats present

     d.  Repeated Measures ANOVA:  Extending the paired-sample t-test

Day 5

     a.  Hierarchical (Nested) Designs

     b.  Course recap

     c.  Consultation session

 
Level 2 - runs over 5 days
Instructor: 

Dr Allen G Harbaugh is a Clinical Assistant Professor at Boston University where his principle duties include consultation on advanced quantitative methods and research design for faculty, staff and students.  He has over twenty years experience in educational research and teaching statistics; he regular presents and chairs at research conferences and workshops, both locally and overseas. He has provided consulting services for many grants and organisations, including projects with the Gates Foundation and UNESCO.  He teaches classes and short courses on numerous advanced quantitative research techniques, including structural equation modelling, multi-level modelling, and research design.  His research focuses on measurement development, self-efficacy to learn mathematics and statistics, and the relationships between epistemic beliefs and motivation.  Having worked as a Lecturer in the School of Education at Murdoch University, he is attuned to many of the issues related to doing quantitative research in Australia.

Course dates: Monday 9 February 2015 - Friday 13 February 2015
Course status: Course completed (no new applicants)
Week: 
Week 3
Recommended Background: 

Participants should have an understanding of elementary statistics equivalent to the syllabus of ‘Fundamentals of Statistics’.

Course fees
Member: 
$1,800
Non Member: 
$3,230
Full time student Member: 
$1,800
Program: 
Summer Program 2015