Publication Type: Invited Paper

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: The level of compliance in a regulatory environment is often portrayed as a function of the cooperation that develops between regulated entities and regulators. In this paper, we argue that agency design plays a critical role in inducing such cooperative outcomes. Specifically we argue that agency design choices on two dimensions, regional scale and decentralization of authority, structure an agency’s ability to engage in flexible enforcement and consequently to secure flexible compliance from the regulated community. Using an original data set on individual-level regulatory compliance in air pollution control, we show that state regional offices that have both larger scale and decentralized authority secure higher flexible compliance compared to any other institutional combination. Our findings underline the importance of design choices in facilitating cooperation in regulatory environments.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: In surveys in which the primary
sampling units consist of individual elements each of which is listed
in the sample frame, a simple random sample is appropriate. However,
because of the impossibility of obtaining a complete list of the sample
frame and because of the high costs associated with obtaining a
complete list, large-scale studies, such as the American National
Election Studies (ANES), the National Crime Victimization Studies
(NCVS), the National Health and Nutrition Examination Surveys (NHANES),
the Consumer Expenditure Surveys (CES), and the Early Childhood
Longitudinal Study (ECLS), use PSUs that are not individuals but are
aggregates, each containing a cluster of individual elements. The
technique used to draw the samples in these large, national studies is
cluster sampling. In a cluster sample the researcher chooses the levels
of the clusters. Then he/she selects a stratified random sample
from each level, sequentially, thus focusing the sample down to the
basic unit to be interviewed. For example, in the ANES the primary
sampling unit is generally a county or metropolitan area. Using a
stratified random sample, a number of these PSUs are selected. These
selected PSUs are further divided into sample places, which are
sampled. The chosen sample places are subdivided into chunks which are
also randomly sampled. Chunks are divided into segments. Segments are
then sampled. Finally, from each of the sampled segments, dwelling
units are randomly selected to be included in the study. In any survey
employing complex designs, clustering, stratification, disproportionate
sampling, and samples with multiple stages, the standard errors are
much larger than a simple random sample of the same size. Most of the
standard statistical software packages, SPSS, SAS, assume a simple
random sample with independent and identically distributed
observations. The only exception is STATA. If the data were collected
using a complex sample design, these standard packages severely
underestimate the variance, resulting in too-small confidence
intervals, leading to the rejection of the null hypothesis when it is
true more frequently than the Type I error level would indicate. The
difference between the variances produced when treating a complex
sample as a simple random sample and the correct variance is called the
design effect and can be calculated. We discuss the reasons for the
complexity of these samples. Then, we examine a regression model
estimated with data collected by the multistage cluster sampling
technique (the American National Election Studies) and calculate the
design effects present when calculating the standard errors as though
it was a simple random sample. Finally, we discuss the implications of
this complexity for research and analysis.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Institutional creation and design; international bargaining; two-stage process; European Development Fund

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Portfolios have continued to become a growing presence in field-based teacher education programs in an attempt to track the progress of teacher candidates on their path to achieving standards. While the understanding of rubrics has improved portfolio evaluation, their reliability has proven tentative at best.