Аспирантура 2024/2025

Методология исследования и базовая статистика

Статус: Курс обязательный

Кто читает: Департамент образовательных программ

Когда читается: 1-й курс, 1 семестр

Охват аудитории: для своего кампуса

Преподаватели: Захаров Андрей Борисович, Чархаби Мортеза

Язык: английский

Course Syllabus

Abstract

The students are expected to have some undergraduate background in quantitative methods (Descriptive statistics vs. inferential statistics; Measures of central tendency and dispersion; Normal distribution and standard scores; Principles of statistical inference, central limit theorem, and sampling distributions; Standard errors (for means, difference between means, correlation, regression coefficient, etc.; z-tests and t-tests; Covariance, correlation, and simple linear regression (regression with only one predictor – interpreting the regression coefficient, standard error of estimate, variance explained, etc.); Statistical power and effect size (what they are and why they’re important); Chi- square Goodness of fit and Test of association (two-way design); One way ANOVA, factorial ANOVA, and within-subjects and mixed ANOVA; Planned comparisons (a priori), contrasts, post - hoc comparisons (a posteriori)) and research designs (Experimental vs. quasi-experimental designs, Threats to validity and experimental control, Statistical inference).

Learning Objectives

• The course aims to review the principal research approaches and designs in social sciences, to provide students with a necessary foundation for developing their own research projects.

Expected Learning Outcomes

• Philosophy behind science and scientific method. Positivism and post-positivism. Nomothetic and idiographic strategies. Purposes for research in the social sciences. Human subjects research: ethics and IRB procedures. Protecting the rights of participants: informed consent, privacy and confidentiality.
• Literature reviews: purposes, procedures, and uses in science. Articulating research problems. Formulating research questions and research hypotheses: good and bad hypotheses. Using theory to turn a research question into a testable hypothesis.
• Variable types, uses, and their places in the research process: a) Observed vs. latent variables. b) Independent vs. dependent variables; Extraneous variables and confounds. c) Predictors, criteria, and covariates. d) Interactions, direct and indirect effects: mediators and moderators. Common variables of interest in social science research. Measurement and instrument validity. Scale types: nominal, ordinal, interval, ratio. Parametric vs non-parametric statistics
• Quantitative data characteristics. Continuous and discrete random variables. Theoretical and empirical (sampling) distributions. Populations vs. samples and the law of large numbers. Univariate vs. bivariate distributions. Measures of central tendency, variability, and distribution shape (skewness and kurtosis). Quantiles. Graphing data. The standard normal distribution as a probability distribution. Missing data analysis.
• Objectives and advantages of sampling. Sampling quality: Representativeness, unbiasedness, and precision. Sampling error. Sampling units – elements vs. cluster sampling. Non-probability vs. probability sampling designs. Non-probability sampling designs: Convenience sampling, purposive sampling, self-selection, snowball sampling. Systematic sampling
• Target vs. accessible populations. Simple random sampling-with and without replacement. Stratified random sampling. Cluster random sampling. Two-stage sampling. Complex samples. Effects of clustering on sampling error: effective sample size and design effect. Sampling weights: theory and procedures.
• Statistical hypothesis: Definition & articulation. The null and alternative hypothesis framework: making decisions using statistical inference. One-tailed and two-tailed hypotheses. Alpha (significance) level, critical vs observed values. Evaluating risk: Type I and type II errors. Confidence intervals and their relation to hypothesis testing. New statistics. A brief review of simple inferential statistics.
• Effect size measures: r, Cohen’s d, variance explained. Converting among effect size measures. Presenting effect sizes to policymakers: Common Language Effect Size (CLES), probability of correct classification, Binomial Effect Size Display (BESD). The concept of statistical power, the role of effect size, sample size, alpha level. The principal types of power analyses; using GPower.
• Descriptive research. Correlation and causality. Parametric and non-parametric correlations (Pearson, Spearman, and other): assumptions and sources of error. The effects of outliers, measurement reliability, aggregation bias, third variable problem. Correlation, regression, and coefficient of determination.
• Cross-sectional designs: research questions and threats to validity. Longitudinal & time series designs: research questions and threats to validity. Sequential designs. Single subject research. Case study research. Historical research.
• Univariate and multivariate statistics. Observed and latent variable models. Models of associations of 3 variables. Multivariate exploratory methods.
• Multiple regression: purpose, assumptions and limitations, steps, presenting results. Dummy coding and effect coding. Simultaneous and sequential (hierarchical) linear regression.
• Models of associations of 3 variables. Testing for simple moderation using GLM/ANOVA and hierarchical linear regression. Mediation: criteria and ways to establish. Complex hypotheses (moderated mediation and mediated moderation) and PROCESS macro. Overview of path analysis. Regression and causality.
• General linear model as a general framework for ANOVA and regression. ANCOVA: purpose, assumptions and limitations, steps, interpreting results, presenting results. MANOVA: purpose, assumptions and limitations, steps, interpreting results, presenting results. Using (M)AN(C)OVAs to analyze repeated-measures experimental data. Nesting.
• Conditions for cause-and-effect relationships. Neyman-Rubin causal model: units, treatments, potential outcomes. What effect is: ITE, ATE, ATT, ATU. SUTVA. Ways of unconfounded, ignorable and non-ignorable assignments. Endogeneity problem and bias: why randomized experiment is a golden standard in causal analysis.
• Threats to internal validity. Pre-experimental designs and how they answer these threats: one-group-posttest-only, one-group-pretest-posttest, nonequivalent-groups, removed-treatment, repeated-treatment, case-control-design.
• RCT designs and threats to internal validity: posttest-only-control-group, pretest-posttest-control-group, multiple-treatments-and-controls, factorial, repeated-measures, longitudinal. Situations when random assignment is desirable and when it is not feasible.
• Setting up a research question. Defining a population and sampling. Randomization: simple, clustered, stratified. Minimum detected effect size and power estimations for different randomization designs. Ways to set up an intervention: presence of a treatment, treatments of several types, dosages of a treatment. Fidelity. Estimation of a treatment effect.
• Overview and differences between qualitative and quantitative paradigms: their purposes, research procedures, and uses. Research in the quantitative paradigm: overview of major research designs. Mixed methods research
• Internal and external validity. Common threats to internal and external validity. Selection bias, maturation, regression to the mean, instrumentation etc. Role of statistical conclusion validity.

Course Contents

• Research defined
  The lecturer will be Morteza Charkhabi
• The problem of validity
  Lecterer will be Morteza Charkhabi
• Measurement: variables and scales, quantitative data characteristics
  The lecturer will be Захаров Андрей Борисович.
• Testing statistical hypotheses
  The lecturer will be Morteza Charkhabi
• A review of multivariate statistics
  The lecturer will be Morteza Charkhabi
• Correlations and descriptive research
  The lecturer will be Morteza Charkhabi
• Advanced regression topics: model specification and, statistical power issues
  The lecturer will be Захаров Андрей Борисович
• Neyman-Rubin Model of Causal Inference. SUTVA, unconfoundedness Assumptions
  The lecturer will be Захаров Андрей Борисович
• Experimental designs. Threats to internal validity
  The lecturer will be Захаров Андрей Борисович
• Randomization
  The lecturer will be Захаров Андрей Борисович
• A quick theoritical review about RCT and practicing previous statistical skills
  The lecturer will be Захаров Андрей Борисович

Assessment Elements

• Quantitative article
• Qualitative article

Interim Assessment

• 2024/2025 1st semester
  Shown below

Bibliography

Recommended Core Bibliography

• Causal inference in statistics, social, and biomedical sciences : an introduction, Imbens, G.W., 2020
• Experimental and quasi-experimental designs for generalized causal inference, Shadish, W. R., 2002
• Introduction to research methods in education, Punch, K. F., 2009
• Methods matter : improving causal inference in educational and social science research, Murnane, R. J., 2011
• Research methods for the behavioral sciences, Gravetter, F. J., 2012
• Using multivariate statistics, Tabachnick, B. G., 2007
• Using multivariate statistics, Tabachnik, B. G., 2007

Recommended Additional Bibliography

• Educational research and inquiry : qualitative and quantitative approaches, , 2010
• Introduction to research methods in education, Punch, K. F., 2014
• Using multivariate statistics, Tabachnick, B.G., 2014

Authors

• SHALOM KSENIYA VLADIMIROVNA
• CHARKHABI MORTEZA