• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

Quantitative Data Analyses

2019/2020
Academic Year
ENG
Instruction in English
9
ECTS credits
Course type:
Elective course
When:
1 year, 3 module

Instructor

Course Syllabus

Abstract

The course covers the topics of diagnostics for linear models, binary, multinomial, ordered, Poisson regression, along with multilevel data analysis. Students are assumed to have basic knowledge of statistics and be familiar with several conventional statistical methods, most importantly linear regression, and with the R programming environment.
Learning Objectives

Learning Objectives

  • Give an introduction to a variety of extensions of linear regression analysis widely used in modern social sciences, as well as implementations of the respective methods in R, a popular, free programming language for statistical computing. By the end of the course, students will be able to choose relevant methods of analysis, implement necessary techniques, and interpret the results of modeling.
Expected Learning Outcomes

Expected Learning Outcomes

  • Able to read and understand most academic social sciences articles that use quantitative approach
  • Able to design a quantitative social study
  • Able to use R programming language for complex statistical computations
  • Able to choose statistical methods appropriate to their data and substantive research problem
Course Contents

Course Contents

  • Diagnostics for linear models
    Chapter 1. Linear regression assumptions Seminar 1. Regression model diagnostics Linearity of the data. Homogeneity of variance, non-constant error variance. variance inflation factors. Non-independence of Errors. Outliers, hat values, and high leverage points. Standardized residuals (studentized residuals), Cook’s distance, variance inflation factors, Durbin-Watson test. Seminar 2. Applying model diagnostics Specify a linear regression model using the data in ‘lab1.Rdata’ file. Check the assumptions of linearity, homogeneity of variance, multicollinearity, as well as outliers and influential values.
  • Binary logistic regression
    Chapter 2. Logistic regressions Seminar 3. Binary logistic regression Variables with binary outcomes. Standard logistic function. Bivariate and multiple logistic models. Logit and probit link functions. Latent variable interpretation. Interpretation of β-coefficients, odds-ratio. Model fit: pseudo R2, PCP, ePCP. Seminar 4. Applying binary logistic models Using the data in ‘lab2.Rdata’ specify a binary logistic model. Interpret coefficients for all predictors. Compute the odds ratios for all coefficient. Compute pseudo-R2 for the model and interpret the results. Calculate PCP and ePCP for the model and comment on the result.
  • Multinomial logistic regression
    Chapter 3. Logistic regression for multinomial data Seminar 5. Multinomial logistic regression Variables with multiple outcomes, the difference between ordered and unordered responses. Contrasts in outcome variables. Interpretation of β-coefficients in multinomial models. Seminar 6. Applying multinomial logistic regression Using the data in ‘lab3.Rdata’ specify a multinomial logistic model. Interpret coefficients for all predictors. Compute the odds ratios for all coefficient. Compute pseudo-R2 for the model and interpret the results. Seminar 7-8. Presentation of project proposals Prepare a short (10 min.) presentation with the description of the research you are going to develop as your final project.
  • Ordered logistic regression
    Chapter 4. Modelling ordinal data Seminar 9. Ordered logistic regression Cumulative probability. Latent variable interpretation, thresholds. Odds-ratio and the interpretation of the coefficients. Model selection and assumptions. Seminar 10. Applying ordered logistic regression Using the data in ‘lab4.Rdata’ specify an ordered logistic model. Interpret coefficients for all predictors. Compute the odds ratios for all coefficient. Compute pseudo-R2 for the model and interpret the results.
  • Models for count data
    Chapter 5. Models for count data Seminar 11. Poisson regression The difference between count and ordered responses. Poisson distribution. Interpretation of the coefficients, incident ratios. Overdispersion, offsets, excessive zeroes. Seminar 12. Applying Poisson regression Using the data in ‘lab5.Rdata’ specify a Poisson regression. Interpret coefficients for all predictors. Compute pseudo-R2 for the model and interpret the results. Seminars 13-14. Test 1 Complete the task using the data according to your test variant. The possible options are the binary logistic regression, multinomial logistic regression, cumulative model or model for count data.
  • Introduction to multilevel modeling
    Chapter 6. Introduction to multilevel modeling Seminar 15. Hierarchical linear models Hierarchy in data structure. Intra class correlation coefficient. Fixed and random effects. Random intercepts and random slopes. Linear models for hierarchical data. Seminar 16. Applying multilevel linear regression Using the data in ‘lab6.Rdata’ specify a multilevel linear regression. Comment on the results.
  • Multilevel logistic regressions
    Chapter 7. Multilevel logistic regressionsSeminar 17. Different types of generalized linear models with hierarchical structure ICC in logistic models. Random intercepts and random slopes in logistic models. Multilevel models for binary, ordinal and count responses. Seminar 18. Applying multilevel logistic regression Using the data in ‘lab6.Rdata’ specify a multilevel logistic regression. Comment on the results. Seminars 19-20. Test 2 Complete the task using the data according to your test variant. The possible options are linear multilevel model, binary logistic multilevel model, cumulative multilevel model or Poisson multilevel model.
Assessment Elements

Assessment Elements

  • non-blocking Test 1
  • non-blocking Test 2
  • non-blocking Home assignments
    An unweighted average of grades for home assignments including project proposal presentation
  • non-blocking In-class participation
  • non-blocking Written final project
  • non-blocking Presentation of the final project
Interim Assessment

Interim Assessment

  • Interim assessment (3 module)
    0.2 * Home assignments + 0.1 * In-class participation + 0.1 * Presentation of the final project + 0.2 * Test 1 + 0.2 * Test 2 + 0.2 * Written final project
Bibliography

Bibliography

Recommended Core Bibliography

  • Agresti, A., & Finlay, B. (2014). Statistical Methods for the Social Sciences: Pearson New International Edition (Vol. Pearson new international ed., 4. ed). Harlow England: Pearson. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=nlebk&AN=1418314
  • Field, A. V. (DE-588)128714581, (DE-627)378310763, (DE-576)186310501, aut. (2012). Discovering statistics using R Andy Field, Jeremy Miles, Zoë Field. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edswao&AN=edswao.363067604
  • Smith, R. B. (2011). Multilevel Modeling of Social Problems : A Causal Perspective. Dordrecht: Springer. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=371921

Recommended Additional Bibliography

  • Chatterjee, S., Hadi, A. S., & Ebooks Corporation. (2012). Regression Analysis by Example (Vol. Fifth edition). Hoboken, New Jersey: Wiley. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=959808
  • Freund, R. J., Wilson, W. J., & Sa, P. (2006). Regression Analysis (Vol. 2nd ed). AMsterdam: Academic Press. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=320724
  • Jiang, J. (2007). Linear and Generalized Linear Mixed Models and Their Applications. New York: Springer. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=212826
  • Upton, G. J. G. (2016). Categorical Data Analysis by Example. Hoboken, New Jersey: Wiley. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=1402878