Summary: what this course will cover
This course will cover some core skills that are necessary for becoming a psycholinguist. Basically, it will make you competent
in the ability to analyze and plot data (eyetracking, ERP, SPR, acceptability ratings and magnitude estimation
data will be used).
As a side effect of the course you will learn a new programming language called R.
How to install R on Windows
Course information
Time: Tuesdays 13:00 - 15:00
Teaching period in summer semester: 14.04.2006-20.07.2007
Place: Golm campus, Haus 24, CIP Pool)
Office hours: by appointment
Readings: The main text will be Gelman and Hill's
Data analysis using regression and multilevel/hierarchical models. Please buy a copy. Note that we will have to work in the Windows OS in order to use WinBUGS. There will be other readings, provided as the class proceeds.
Tentative Schedule [we will probably go slower]
April 14: Pavel Logacev will lead this class because I am away today
HW 0: work through
this tutorial up to (including) section 4
April 21: Linear regression introduction and review
HW1: Read Chapters 1-4 of Gelman and Hill and do problem 5 in chapter 3. In the models you fit (ratings as a function of beauty and/or sex), explain what all the estimated parameters mean.
Code + Data: child IQ as a function of mother's IQ and schooling.
More on regression: here.
April 29: More regression background (esp. multicollinearity)
HW2: Chapter 2, Problem 4.
May 6: Fake data simulation (parts of Ch 7 and 8), and
(maybe) Introduction to mixed-effects models (Ch 11)
Code+Data: here
May 13: Continue with last week's topics: fake data simulation, mixed-effects models. We will also discuss why we divide by n-1 and not n in the definition for standard deviation.
June 10: The goal today is to get some practice in practical data analysis using lmer. We will work with three datasets: (i) Eyetracking data from Vasishth, Bruessow, Lewis, Drenhaus 2008, (ii) Drenhaus et al ERP data (to be emailed), (iii) plotting the coefficients and standard errors computed from an lmer model (data to be emailed); here is an example plot (also download this data).
June 17: We will review the meaning of a confidence interval. Then we will learn to do some plotting. See here for code.
Grading
Weekly exercises and reading and writing assignments.
Final
scores will be based on the following mapping described in the
Studienordnung: 95-100%=1,0 (A);90-94=1,3 (A-);85-89=1,7
(B+);80-84=2,0 (B);75-79=2,3 (B-);70-74=2,7 (C+);65-69=3,0
(C);60-64=3,3 (C-);55-59=3,7 (D+);50-54=4,0 (D);45-49=5,0 (F). If a
student's score falls between the cracks, it will be treated as
falling in the higher bin. Also, if a student is in a higher semester
than 1st, I will adjust their scores so that 1st semesters are not at a
disadvantage (this holds only if the advanced students outperform the
1st semesters).
Here is an excerpt from the Studienordnung on what these major
categories are supposed to mean:
1 = sehr gut (eine hervorragende Leistung)
2 = gut (eine Leistung, die erheblich ueber den
durchschnittlichen Anforderungen liegt)
3 = befriedigend (eine Leistung, die durchschnittlichen Anforderungen entspricht)
4 = ausreichend (eine Leistung, die trotz ihrer Maengel noch den Anforderungen genuegt)
5 = nicht ausreichend (eine Leistung, die wegen erheblicher Maengel den Anforderungen nicht genuegt)
Students are expected to attend class regularly. If a student misses a
class, the student is responsible for finding out what the homework
assignment was, what readings were assigned, and what material was
covered.
Homework submission dates are strict (usually beginning of class the
following week). Late homework will not receive any credit.
Conduct in the classroom
Please do not engage in private conversations during class.
All cell phones must be switched off (except by permission from me).
Please do not walk into class after it starts (13:15AM is the deadline to be ready for class).
Questions to me during class are actively encouraged.
Official languages of the course
English and German.