In pursuit of balance : randomization in practice in development field experiments, Vol. 1 of 1

About

Countries

Data & Research

Learning

News

Projects & Operations

Publications

Topics

Home > Data & Research > Research > Author Profile > Report Details

In pursuit of balance : randomization in practice in development field experiments

Author:	Bruhn, Miriam; McKenzie, David;	Collection Title:	Policy Research working paper ; no. WPS 4752
Country:	World;	Date Stored:	2008/10/15
Document Date:	2008/10/01	Document Type:	Policy Research Working Paper
Language:	English	Major Sector:	Finance
Rel. Proj ID:	1W-Program Research Proposal In Finance -- -- P107616;	Region:	The World Region
Report Number:	WPS4752	Sub Sectors:	General finance sector
SubTopics:	Climate Change and Environment; Scientific Research & Science Parks; Economic Theory & Research; Science Education; Statistical & Mathematical Sciences	Volume No:	1 of 1

Summary: Randomized experiments are increasingly used in development economics, with researchers now facing the question of not just whether to randomize, but how to do so. Pure random assignment guarantees that the treatment and control groups will have identical characteristics on average, but in any particular random allocation, the two groups will differ along some dimensions. Methods used to pursue greater balance include stratification, pair-wise matching, and re-randomization. This paper presents new evidence on the randomization methods used in existing randomized experiments, and carries out simulations in order to provide guidance for researchers. Three main results emerge. First, many researchers are not controlling for the method of randomization in their analysis. The authors show this leads to tests with incorrect size, and can result in lower power than if a pure random draw was used. Second, they find that in samples of 300 or more, the different randomization methods perform similarly in terms of achieving balance on many future outcomes of interest. However, for very persistent outcome variables and in smaller sample sizes, pair-wise matching and stratification perform best. Third, the analysis suggests that on balance the re-randomization methods common in practice are less desirable than other methods, such as matching.


Official Documents Official, scanned versions of documents (may include signatures, etc.)
File Type		Description	File Size (mb)

	PDF 59 pages	Official version*	4.13 (approx.)
	Text	Text version**

	See documents related to this project

* The official version is derived from scanning the final, paper copy of the document and is the official,
archived version including all signatures, charts, etc.

** The text version is the OCR text of the final scanned version and is not an accurate representation of the final text.
It is provided solely to benefit users with slow connectivity.

Permanent URL for this page: http://go.worldbank.org/OCVZK7GG70