Advanced Quantitative Methods Summer School 2019
Export to calendarBio
The Quantitative Methods Hub at the Department of Education, University of Oxford, offers its annual Advanced Quantitative Methods Summer School, this year consisting of eight different course days in week 3 and 4 of Trinity Term (14-24/5/2019). The courses require an understanding of multiple regression modelling or other multivariate techniques.
Students and staff are welcome to attend one, some or all days. Please register using link on right-hand side of page.
The courses will take place at the Department of Education at 15 Norham Gardens, OX2 6PY, Oxford. Please bring along your own laptop as we will not be in a computer class.
The programme is:
Week 1
Tue 14 May 10:00-16:30 Thees Spreckelsen, Intro to R
Wed 15 May 10:00-15:30 Joshua McGrane, Intro to Rasch modelling
Thur 16 May 11:00-17:00 Rens van de Schoot, Bayesian modelling, part 1
Fri 17 May 9:00-14:00 Rens van de Schoot, Bayesian modelling, part 2
Week 2
Tue 21 May 10:00-15:30 Kit Double, Multilevel modelling
Wed 22 May 10:00-15:30 Lars Malmberg, Intro to structural equation modelling (SEM)
Thu 23 May 10:00-15:30 Lars Malmberg, Longitudinal SEM
Fri 24 May 10:00-15:30 Lars Malmberg, Multilevel SEM
************** WEEK 1 **************
Tuesday 14 May 2019, 10:00 to 16.30
Introduction to R for the Analysis of Educational Data
Instructor: Thees F Spreckelsen
This course will introduce R for statistical analysis using the RStudio graphical user interface. The morning session will provide an introduction to the R language, the RStudio interface, and basic functions for data analysis. Applied examples will demonstrate R/RStudio broad functionalities. Participants will learn to work with objects, including how to create, save, and inspect objects, particularly data frame objects. They will create a simulated data frame and perform basic analysis functions with this object. In the afternoon, in a code-along session, participants will work with existing educational data from large-scale assessment studies. They will learn to read data into R and will apply data analysis functions presented in the morning session to real educational data. Participants will do basic datamanagement steps, calculate descriptive statistics and perform regression analysis. The third part of the course will be dedicated to a do-it-yourself assignment. The final session, will provide participants with a number of helpful resources and a set of next steps. In particular participants will be introduced to Rmarkdown – an easy way to document results as word- or html-document
Course prerequisites: It is assumed that participants will have a background in basic statistical methods up to, and including, regression analysis. Familiarity with syntax language from other statistical packages (eg. Stata, SPSS) is desirable.
Contents:
10:00 – 10:45 Introducing R and RStudio
10:45 – 11:00 Break (software installation)
11:00 – 12:30 R basics for data analysts
12:30 – 13:00 Lunch
13:00 – 14:30 Working with educational data
14:30 – 14:45 Break
14:45 – 15:45 Practise session
15:45 – 16.30 Next steps in R/Q&A
Cost: £25 for OU and Grand Union students, £100 for staff and external students (£80 / £320 for the week). Lunch and refreshments included in price.
Wednesday 15 May 2019, 10.00 to 15.30
Introduction to Item Response Theory and Rasch modelling
Instructor: Joshua McGrane
Over the past decades, Item Response Theory has become increasingly used as a validation tool for assessments and questionnaires in a range of fields, including education, psychology and health sciences. This one-day workshop will introduce Item Response Theory for assessment and survey data, with a particular focus upon Rasch modelling. Firstly, an overview of the conceptual and philosophical foundations of Item Response Theory will be provided, including the key differences between the models and their applications. The workshop will then focus upon Rasch modelling in more detail, including the different models for test and survey data, the information these models provide regarding the reliability and validity of assessments, and in particular, how the Rasch model is a powerful tool for diagnosing anomalies in assessment and survey data. Participants will then be introduced to Rasch modelling in R using the RStudio graphical user interface.
The workshop will be both theoretical and hands-on, and participants are encouraged to bring their own data for the practise session in the afternoon. An example dataset will be provided for those who do not bring their own data.
Course prerequisites: It is assumed that participants will have a background in basic statistical methods and familiarity with statistical software.
Contents:
10:00 – 10:15 Welcome, agenda
10:15 – 11:00 Foundations of psychometrics and Item Response Theory: basic concepts, different IRT models, paradigmatic differences
11:00 – 11:15 Break
11:15 – 12:30 Introduction to Rasch modelling: dichotomous and polytomous models, model estimation, tests of model fit
12:30 – 13:00 Lunch
13:00 – 13:45 Rasch modelling using R
13:45 – 14:00 Break
14:00 – 15:30 Practise session using R
Cost: £25 for OU and Grand Union students, £100 for staff and external students (£80 / £320 for the week). Lunch and refreshments included in price.
Thursday 15 May, 12.00-17.00 and Friday 16 May 2019, 9.00-14.00
Bayesian modelling
Instructor: Rens van de Schoot, Utrecht University
The use of Bayesian estimation has increased over the years because this estimation framework can handle some commonly encountered problems in orthodox statistics. For example, Bayesian methods can be used for producing more accurate parameter estimates and aiding in situations where only small sample sizes are available. Or, some researchers believe in the Bayesian way of updating knowledge with new data instead of testing the null hypothesis over and over again assuming nothing is going on in the population. This course will introduce participants to the prevailing “best practices” for Bayesian estimation (including structural equation modeling) entailing direct application to the research questions of the participants. During this course, you will be gently introduced into Bayesian statistics using class examples. We will explore the benefits of Bayesian statistics and discuss what is needed to run your first Bayesian model.
Knowledge of regression analysis and basic SEM is required. No previous knowledge of Bayesian analysis is assumed. You do not need to know matrix algebra, calculus, or likelihood theory. Since the course offers a gentle introduction there are hardly any formulas used in the lectures. The main focus is on conceptually understanding Bayesian statistics and applying Bayes to your own data set. Participants from a variety of fields—including psychology, education, human development, public health, prevention science, sociology, marketing, business, biology, medicine, political science, and communication—will benefit from the course.
After engaging in course lectures and discussions as well as completing the hands-on practice activities with real data, participants will be able to:
• Explain the differences between ‘classical’ and Bayesian statistics.
• Learn when to use to Bayesian analyses.
• Learn how to use Bayesian analyses for your own data.
• Learn how to apply the WAMBS-checklist (When to worry and how to Avoid the Misuse of Bayesian Statistics).
• Critically evaluate applications of Bayesian in scientific studies.
Contents:
Thurs 16
1130-1200 Lunch
1200 – 1400: Conceptual introduction + reasons for using Bayesian methods + discussion on interpretability of results when using p-values/95%intervals
1400-1430 break
1430 – 1530: Empirical example of a linear regression analysis in the Bayesian framework + where do priors come from
1530 – 1545 break
1545 – 1700: Q&A + which software to use + computer lab
Fri 17 May
900-1100: WAMBS-checklist: When to worry and how to Avoid the Misuse of Bayesian Statistics
11:00-11:15 Break
1115-1200: example Bayesian Latent Growth Mixture Model;
1200-1300 Lunch
1300-1400: Q&A + tips and tricks how to start with your own data
Cost: £25 for OU students per day, £100 for staff and external students per day (£80 / £320 for the week). Lunch, refreshments and handouts included in price.
************* WEEK 2 **********
Tuesday 21 May 2019, 10:00 to 15:30
Multilevel Modelling
Instructor: Kit Double
This course will introduce multilevel modelling (MLM) and provide an analytical framework for the study of research questions with MLM in R. During the morning session, lectures will introduce the rationale for MLM, provide applied examples in educational research and introduce a framework for the analysis of substantive research questions using MLM. Participants will learn how to set up, estimate, and interpret multilevel models in R. In the afternoon, participants will perform MLM on existing educational data, including specifying different model types, interpreting output and producing visual representations of MLM designs. The last part of the course will be dedicated to analysing growth curve models in R using an example dataset.
Course prerequisites: Participants need to understand the basics of multiple regression, or other relevant multivariate statistics. Ideally participants will have attended the ‘Introduction to R for the Analysis of Educational Data’ or have a basic knowledge of how to use R. Students should install R (https://www.r-project.org/) and Rstudio (https://www.rstudio.com/) on their own laptop prior to the course.
Contents:
10:00 – 11:30 Overview of MLM – Models Types and Research Questions
11:30 – 11:45 Break
11:45 – 12:30 Growth curve models and MLM resources 12:30 – 13:00 Lunch
13:00 – 14:30 Addressing research questions with MLM in R
14:30 – 14:45 Break
14:45 – 15:30 Growth Curve MLM in R
Cost: £25 for OU students, £100 for staff and external students (£80 / £320 for the week). Lunch, refreshments and handouts included in price.
Wednesday 22 May 2019, 10.00 to 15.30
Introduction to Structural Equation Modelling
Instructor: Lars-Erik Malmberg
The concept of a latent construct is central in the social sciences. A latent construct is a non-directly observed phenomenon (e.g., attitude, socioeconomic status) that we can model using manifest (observed) variables (e.g., survey and questionnaire responses, observation scores), by partitioning out residual (i.e., uniqueness, error variance). The structural equation model (SEM) is divided into two parts. In the measurement part of the model, we can inspect whether manifest variables measure the constructs they are intended to measure. This model is called confirmatory factor analysis (CFA) which allows the researcher to test whether an a priori model fits data, and whether this also holds across multiple groups. If measurement is satisfactory, the relationships between constructs can be estimated in the structural part of the SEM. Complex relationships between manifest variables and/or latent constructs can be tested in path-models not possible to specify in the multiple regression framework. During the course we will cover worked examples relevant for educational, psychological and social sciences.
Course Pre-requisites:Participants need to understand the basics of multiple regression, or other relevant multivariate statistics.
Contents
10:00 – 11:00 Introduction: Basic concepts, models and measurement. From multiple regression to path-models using manifest variables (in Mplus).
11:15 – 12:30 Observed (manifest) variables and unobserved (latent) constructs. Specification of measurement models for testing quality of measurement, using continuous and dichotomous manifest variables. Goodness-of-fit indices (in Mplus).
12:30 – 13:00 Lunch
13:00 – 14:15 Relationships between latent constructs. Specifying structural models to include directional (regression) paths between latent constructs.
14:15 – 14:30 Break
14:30 – 15:30 Practice session or work on own data
Software: We will mainly use the Mplus software. Some parallel code is available in R (Lavaan)
Cost: £25 for OU students, £100 for staff and external students (£80 / £320 for the week). Lunch, refreshments and handouts included in price.
Thursday 23 May 2019, 10.00 to 15.30
Structural Equation Modelling of longitudinal data
Instructor: Lars-Erik Malmberg
This follow-up of the introduction to SEM is an advanced course in which we focus on SEM for longitudinal data. Prospective longitudinal data is usually collected over longer periods of time (e.g., term or year) while intensive longitudinal is gathered within shorter time-spans (e.g., numerous times a day). Using SEM we can model repeated latent constructs over time using autoregressive models (i.e., a construct at the concurrent time-point is regressed on a construct at a previous time-point). We can also specify latent change models using “phantom constructs”. When particular interest is in individual differences in change over time, we can model time explicitly in the latent growth curve model. We will cover worked examples relevant for educational, psychological and social sciences.
Course Pre-requisites: Participants need to understand the basics of multiple regression, other relevant multivariate statistics, and have some exposure to either multilevel regression or SEM.
Contents
10:00 – 11:00 Overview of SEM for longitudinal data. Repeated measures and autoregressive models with manifest and latent constructs.
11:00-11:15 Break
11:15 – 12:30 The reciprocal effects model. The latent change model.
12:30 – 13:00 Lunch
13:00 – 14:15 The latent growth model. Coding of time and error structures.
14:15 – 14:30 Break
14:30 – 15:30 Practice session or work on own data
Software: We will mainly use the Mplus software. Some parallel code is available in R (Lavaan)
Cost: £25 for OU students, £100 for staff and external students (£80 / £320 for the week). Lunch, refreshments and handouts included in price.
Friday 24 May 2019, 10.00 to 15.30
Multilevel Structural Equation Modelling
Instructor: Lars-Erik Malmberg
In this follow-up of the introduction to SEM we focus on SEM for multilevel data. We start with an introduction to multilevel modelling, its notation, and model specification in the SEM framework. We will then consider two kinds of nested structures: students nested in institutions (e.g., students nested in schools), and time-points nested in persons.
Both cross-sectional and longitudinal data can be collected applying a nested structure (e.g., students in classrooms, parents in families, time-points in persons). Using SEM we can model repeated latent constructs over time, or across hierarchical levels net of measurement error. We will start by defining the multilevel model within the structural equation modelling framework. We will then introduce multilevel confirmatory factor analysis (MCFA) and full multilevel structural equation models (MSEM). In the worked examples we will use example data of both persons nested within organisations and time-points nested within persons.
During the course we will cover worked examples relevant for educational, psychological and social sciences. Participants need to understand the basics of multiple regression, other relevant multivariate statistics, and have some exposure to either multilevel regression or SEM.
Contents:
10:00 – 11:00 Introduction to MLM. Fixed effects and random effects. Comparison of MLM notation and visual depiction of MLM in the SEM framework.
11:00-11:15 Break
11:15 – 12:30 Factor structures across levels. Structural models at the within and between levels.
12:30 – 13:00 Lunch
13:00 – 14:15 Time-points nested in persons, parents nested in families
14:15 – 14:30 Break
14:30 – 15:30 Practice session or work on own data
Cost: £25 for OU students, £100 for staff and external students (£80 / £320 for the week). Lunch, refreshments and handouts included in price.
For enquiries contact: charlotte.trevillion@education.ox.ac.uk
About the Department of Education
In 2019, the University of Oxford’s Department of Education celebrates the 100th year since the passing of a statute creating what was known in 1919 as the University Department for the Training of Teachers. To celebrate our centenary a year-long series of activities will be delivered to address some of the department’s top initiatives for 2019, answer some of the big questions facing education today and to reveal the advancements the department has made to the study of and research in the field of education. Join us as we mark our 100th year and discover more about our anniversary here.
To receive more event details from the Department of Education, join our mailing list.