Q560: Experimental Methods in Cognitive Science
Fall 2021: Tues/Thurs, 9:25-10:40am, Geology 6050G
Download PDF Version of Syllabus
Instructor: Dr. Michael Jones
Office: PSY 370
Phone: 856-1490
Email: [email protected]
Office Hours: By appointment
Class webpage: http://www.compcog.com/courses.html
Overview
The purpose of Q560 is to bring Cognitive Science graduate students up to a common level of knowledge and skill in research design and analysis. It is most commonly taken by students without a strong background in empirical research. The course will focus on experimental design, behavioral statistics, methodologies, contemporary issues, and critical analysis skills.
The bulk of the course will be instructor-led lectures on statistical methodology relevant to cognitive science, sample experiments, and data analysis with computers. The remainder of the course will involve each student presenting a research project (experiment design and analysis) relevant to his/her research interests.
Goals
Materials
There is no textbook for this course. If you need a refresher on basic statistics about any undergraduate textbook with a title like “Statistics for Psychology” will contain the basics of inferential stats…they have all pretty much regressed to the mean of parametric stats over the years. There are many of these in used bookstores or lying in dumpsters around campus. A few good texts that are likely to be lying around the used bookstores are Aron & Aron, Gravetter & Wallnau, and a good one by David Howell.
There are also several online texts available, and I’ll put links to some good ones on our website. Your best option is probably to use the Introductory Statistics book at http://openstax.org/. My lecture slides will be posted on the course website as well.
I will be analyzing data using the SPSS statistical software package. I’m not using SPSS because I think it is the best but, rather, (like parametric statistics) it is very simple and the most commonly used. Most data analysis packages work similarly, and transfer between them is usually very easy. I will do examples with SPSS, but you are free to submit homework assignments and to analyze your final project data with any other package that has the required routines (SAS, S-Plus, Minitab, Statistica, R, Matlab, an abacus etc.).
SPSS is available in any UITS-maintained computer lab on campus, or via IU AnyWare.
We will also be learning to use a variety of new online tools for design and conduct of behavioral experiments (e.g., jsPsych, PsiTurk, Pushkin, etc.)
Evaluation
There will be five homework assignments involving article critiques, hand calculations, and computer analyses. You will also receive credit for your prospective experiment presentation, and your final paper after the experiment has been completed. The evaluation will be allocated as follows:
Homework Assignments (5 x 7%)…35%
Project Proposal Presentation….…25%
Final Paper……………………….…35%
Participation…………………………5%
Individual Project
Each student will select a research problem of interest and design a simple experiment to test a hypothesis. The experiment need not be administered by computer (but it is often much easier in the long run). The instructor and other students will be the subjects of your experiment, and you will analyze the data and write up a paper as if you are submitting to a Cognitive Science journal for publication.
A brief summary of the phenomenon that you are interested in exploring is due October 5. Simply state the domain (e.g., memory), and the phenomenon you are interested in (e.g., the mirror effect) and generally how you would like to explore it. Alternatively, you can pick something that is more exploratory (e.g., you want to explore whether people remember material better when presented with flashy graphics vs. plain text in Powerpoint). Give a few references so I can learn more about the area.
You will give a 20-minute presentation to the class on your proposed experiment. Be sure to include things like:
We will give constructive feedback about your experiment, and try to identify any potential flaws. When the experiment is airtight, you will actually conduct it—we will pick a week after the presentations to run all the experiments in my lab. You are also welcome to collect data online via a system such as PsiTurk.
With the experiment complete, you will analyze the data to determine the results, and write up a paper (max 16 pages) double-spaced APA format as if you were submitting your research report to a journal for publication. Be sure to include:
There are links to APA style on our website. I know it may not be the major editorial style of your particular discipline, but it is the way all behavioral experiments are disseminated in cognitive science.
You will find several sample projects, from start to finish, posted on our website.
Instructor: Dr. Michael Jones
Office: PSY 370
Phone: 856-1490
Email: [email protected]
Office Hours: By appointment
Class webpage: http://www.compcog.com/courses.html
Overview
The purpose of Q560 is to bring Cognitive Science graduate students up to a common level of knowledge and skill in research design and analysis. It is most commonly taken by students without a strong background in empirical research. The course will focus on experimental design, behavioral statistics, methodologies, contemporary issues, and critical analysis skills.
The bulk of the course will be instructor-led lectures on statistical methodology relevant to cognitive science, sample experiments, and data analysis with computers. The remainder of the course will involve each student presenting a research project (experiment design and analysis) relevant to his/her research interests.
Goals
- A firm grasp of the theory behind experimental design and analysis
- Ability to design and conduct simple experiments
- Ability to critically read and interpret research literature
- Experience with software for programming experiments and analyzing data
- Exposure to contemporary issues in the behavioral methodology literature
Materials
There is no textbook for this course. If you need a refresher on basic statistics about any undergraduate textbook with a title like “Statistics for Psychology” will contain the basics of inferential stats…they have all pretty much regressed to the mean of parametric stats over the years. There are many of these in used bookstores or lying in dumpsters around campus. A few good texts that are likely to be lying around the used bookstores are Aron & Aron, Gravetter & Wallnau, and a good one by David Howell.
There are also several online texts available, and I’ll put links to some good ones on our website. Your best option is probably to use the Introductory Statistics book at http://openstax.org/. My lecture slides will be posted on the course website as well.
I will be analyzing data using the SPSS statistical software package. I’m not using SPSS because I think it is the best but, rather, (like parametric statistics) it is very simple and the most commonly used. Most data analysis packages work similarly, and transfer between them is usually very easy. I will do examples with SPSS, but you are free to submit homework assignments and to analyze your final project data with any other package that has the required routines (SAS, S-Plus, Minitab, Statistica, R, Matlab, an abacus etc.).
SPSS is available in any UITS-maintained computer lab on campus, or via IU AnyWare.
We will also be learning to use a variety of new online tools for design and conduct of behavioral experiments (e.g., jsPsych, PsiTurk, Pushkin, etc.)
Evaluation
There will be five homework assignments involving article critiques, hand calculations, and computer analyses. You will also receive credit for your prospective experiment presentation, and your final paper after the experiment has been completed. The evaluation will be allocated as follows:
Homework Assignments (5 x 7%)…35%
Project Proposal Presentation….…25%
Final Paper……………………….…35%
Participation…………………………5%
Individual Project
Each student will select a research problem of interest and design a simple experiment to test a hypothesis. The experiment need not be administered by computer (but it is often much easier in the long run). The instructor and other students will be the subjects of your experiment, and you will analyze the data and write up a paper as if you are submitting to a Cognitive Science journal for publication.
A brief summary of the phenomenon that you are interested in exploring is due October 5. Simply state the domain (e.g., memory), and the phenomenon you are interested in (e.g., the mirror effect) and generally how you would like to explore it. Alternatively, you can pick something that is more exploratory (e.g., you want to explore whether people remember material better when presented with flashy graphics vs. plain text in Powerpoint). Give a few references so I can learn more about the area.
You will give a 20-minute presentation to the class on your proposed experiment. Be sure to include things like:
- Why anyone should care about this problem (link to real world examples)
- Relevant work that has been done to study the phenomenon
- Relevant theories or models that make predictions about the phenomenon
- Your proposed experiment: design, stimuli, hypotheses, variables, measures, potential confounds, and proposed analyses.
We will give constructive feedback about your experiment, and try to identify any potential flaws. When the experiment is airtight, you will actually conduct it—we will pick a week after the presentations to run all the experiments in my lab. You are also welcome to collect data online via a system such as PsiTurk.
With the experiment complete, you will analyze the data to determine the results, and write up a paper (max 16 pages) double-spaced APA format as if you were submitting your research report to a journal for publication. Be sure to include:
- An Abstract of ≤ 250 words describing the paper and results
- An Introduction summarizing the background literature around this phenomenon and clearly stating your research hypotheses
- A Methods section including the relevant subsections (Participants, Stimuli, Design, Procedure)
- A Results section containing your statistical analyses of the data
- A Discussion section
- References
- Figures and tables may be inline (embedded in text)
There are links to APA style on our website. I know it may not be the major editorial style of your particular discipline, but it is the way all behavioral experiments are disseminated in cognitive science.
You will find several sample projects, from start to finish, posted on our website.