30 August 2013
Causality and Effect Evaluation within Public Administration
Effect evaluations are becoming increasingly popular within public administration both among practitioners and scholars. The aim of these studies is to provide insights into how public performance can be improved by answering questions such as “Do higher motivation among employee lead to higher performance?” or “Do manager type X do better than manager type Z?”. As these studies often imply policy recommendations, e.g. do X instead of Z, it is essential to avoid mistaking accidental association with causality. A thorough understanding of the theoretical and econometric background of effect evaluation is, hence, necessary. The aim of this seminar is to enable students to assess the quality of existing effect evaluations within public administration and to set up their own empirical effect analysis. In the first part of the seminar we will discuss what constitutes an effect and the strengths and weaknesses of the most commonly used tool for effect evaluation: OLS regression. In the second part of the seminar students are introduced to the more advanced econometric tools fixed effects, difference-in-difference and IV estimation in a non-technical way (with a minimum of mathematics). Hence, focus will be on teaching students how to use these tools, and why they can be useful when doing effect evaluation, not on mathematical derivation. In the third part of the seminar we will discuss strengths and weaknesses of existing effect evaluations dealing with one of three popular themes within public administration: managerial autonomy, use of performance related pay (PRP) and employee public service motivation (PSM). For each theme we will evaluate how and why results (and thereby policy recommendation) differ between studies as well as whether the methodological designs needs improvement. The methodological discussions during the seminar will be relevant for students doing or planning to do effect evaluation within other disciplines than public administration, and student will have the opportunity to get feedback on their own (future) effect evaluation regardless of discipline. There will be three to four practical afternoon session during the seminar. In these sessions we work with OLS regression, fixed effects, difference-in-difference and IV estimation in STATA. It is not a prerequisite for this seminar to be a STATA user.
The seminar is organized as follows:
Week 1: What constitutes an effect?
Intended and unintended effects
Experiments as the gold standard
OLS regression and endogeneity
Week 2: When OLS is broke, how do we fix it?
Week 3: What do we know of the performance effect of PSM, PRP and management autonomy?
Strengths and weaknesses of existing studies
Possibilities for improvement?
Students own (future) effect evaluation
Maria Falk Mikkelsen, Aarhus University
The objectives for this module are ability to:
- Describe and compare relevant theories as regards citizens’ perceptions of public services.
- Discuss the advantages and drawbacks of using different research designs and methods to study citizens’ perceptions of public services.
- Critically valuate and discuss theoretical and methodological strengths and weaknesses of existing studies of citizens’ perceptions of public services.
- Analyze and discuss to what extent citizens’ perceptions of public services is based on misperceptions.
- Analyze and discuss to what extent citizens’ perceptions can be shaped by new information.
- Analyze and discuss drivers of citizens’ satisfaction with public services
- Analyze and discuss how citizens react when dissatisfied with public services and why they do so.
Present their analysis in a clearly written language.
EUR 0: Students on a bilateral exchange programme do not have to pay. Freemovers are obliged to pay participation fees while tuition fees only apply to freemovers from countries outside the EU/EEA/Switzerland.