Магистратура
2021/2022
Экспериметрика
Статус:
Курс по выбору (Экономика и экономическая политика)
Направление:
38.04.01. Экономика
Кто читает:
Департамент теоретической экономики
Где читается:
Факультет экономических наук
Когда читается:
1-й курс, 3, 4 модуль
Формат изучения:
без онлайн-курса
Охват аудитории:
для своего кампуса
Преподаватели:
Тремьюэн Джеймс Кристофер Росс
Прогр. обучения:
Экономика и экономическая политика
Язык:
английский
Кредиты:
6
Контактные часы:
80
Course Syllabus
Abstract
This course will cover a range of statistical tools useful for the analysis of experimental data, with a focus on text data and applications in Behavioural Economics. The majority of the course will be taken up with the estimation of structural behavioural models. The techniques developed in the course will be applied to the core topics of decision-making under uncertainty, social preferences, and bounded rationality. Students will learn how to use experimental data to estimate key parameters from prospect theory (e.g. risk and loss aversion), models of altruism and fairness-concerns, level-k models, and quantal response equilibria. We will explore how these parameters can vary within and and across sub-populations using finite-mixture and random-effects models. We will also discuss the "replication crisis", and potential approaches to increase the reliability of experimental findings. Programming in class will be performed in R, however prior knowledge is not required.
Learning Objectives
- Design experiments to maximize statistical power.
- Apply appropriate statistical tests to experimental data.
- Make correct inferences from statistical tests.
- Correctly account for non-independence of experimental data.
- Identify and correct for issues leading to non-replicability of experimental results.
- Estimate structural models derived from behavioural economics.
- Critique inference in experimental studies.
Expected Learning Outcomes
- To adjust for multiple hypotheses
- To Introduce to hypothesis testing. To learn Binomial test, exact Z-test, Mann-Whitney, stochastic inequality test, Wilcoxon rank sum, sign test, Spearman rank-correlation, Kendall rank correlation. To use Monte Carlo simulations to estimate size and power of tests
- To learn Heterogeneity and structural modelling
- To learn how to design experiments to maximize statistical power
- To use matching-group averages. To make Multi-level modelling
Course Contents
- Experimental design and experimetrics • How to design experiments to maximize statistical power.
- Reading and critiquing experimental studies
- Regression analysis and dealing with dependence of observations
- Heterogeneity and structural modelling
- The replication crisis
- Hypothesis testing and non-parametric tests • How to choose which test to apply to experimental data and what inferences can be drawn.
Assessment Elements
- Non-parametric tests
- Multiple hypothesis testing and accounting for non-independent data
- Estimating structural models
- Critique of experimental study
- Presentation of own critique and class discussion
- Final testRetake and comission will be held in the same format as the final test. Grades will be recalculated
Interim Assessment
- 2021/2022 4th module0.15 * Critique of experimental study + 0.15 * Estimating structural models + 0.3 * Final test + 0.15 * Multiple hypothesis testing and accounting for non-independent data + 0.15 * Non-parametric tests + 0.1 * Presentation of own critique and class discussion
