Publication Type: Single Paper Format
Abstract: Feedback-based modifications have defined and enriched our Teacher Work Sample assessment implementation. The detours and benefits of our TWS journey will be the focus of this session.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Based on a pilot study of 159 university undergraduates, this paper presents Respondent Driven Sampling, a new method of studying hidden populations, as a viable method of understanding overall network structure. The method was found to be highly effective and efficient. Examples identifying bridge groups, centrality, and multiplex association are discussed.

Supporting Publications:
Supporting Document

Publication Type: Paper Presentation
Abstract: Kaiser’s Measure of psychometric Sampling Adequacy (MSA) is an indicator of factorability for a collection of variables. Kaiser also had a variable MSA, but never provided guidelines for using it. This presentation discusses various uses for variable MSA asserting that the determination and analysis of variable MSA and general MSA for a collection of variables should be a requisite analysis before a factor analysis. The understanding and use of variable MSA should: facilitate more generalizable and stable exploratory and confirmatory factor analyses; help salvage collections of variables that appear to be “unfactorable” as well as collections of variables that are supposed to be “factorable”, but turn out to be convoluted and illogical with regard to the factors that emerge. Learning about the use of variable Measure of Sampling Adequacy should be helpful for beginning and experienced factor analysts.

Publication Type: Paper/Poster Proposal
Abstract: Survey designers often want to oversample specific sub-populations to improve precision and robustness. Nearly all existing methods in two stage sampling produce explicit problems, with specific weaknesses, because domains (sub-populations) targeted for over sampling usually cross-cut the clusters that form the PSUs.

This paper sets out a new method of sampling giving major gains over previous methods, in both effectiveness and efficiency, by simultaneously constraining three factors:

• - sample sizes within domains that cross-cut clusters
• - sample sizes within clusters
• - variation in selection probabilities, both within domains and overall.
This is achieved through use of a size measure which, instead of being a simple count of the second stage units, is a weighted sum of the counts of units in each domain within the cluster. This requires knowledge of the distribution of units over domains within each cluster, which is often possible when administrative data are available. These data can be exploited to produce samples with closely controlled composition of PSUs, that are more effective in boosting sub-samples than standard methods and have, through reduced design effects and increased effective sample size, greater cost efficiency than otherwise possible.

The paper comprises a description of the method, its underlying theoretical framework and a case study of its application in the design and implementation of the sample for a major longitudinal study of young people. The overall design was a conventional two-stage one with PPS, with schools being PSUs and individuals sampled from their registers. A critical objective was, however, boosting sample numbers for six ethnic minority groups. The paper sets out the advantages of the new method in this case, problems encountered and their solutions. The conclusion is that this method has substantial advantages over others and should be the default method for use in similar circumstances.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Internet surveys are rapidly becoming an accepted procedure in political science, yet questions about them remain. Using samples of college students aged 18-24 from four Internet surveys, three phone surveys, a Knowledge Networks web-enabled survey, and a mixed-method survey, we seek to understand the effects of Internet sampling of young adults on the assessment of voter registration, party identification, political ideology, and political engagement. We find that the Internet surveys showed significant differences with surveys conducted with other sampling methods. In particular, the Internet surveys attracted a considerably more activist sample of college students than the phone surveys. They also seemed to attract more ideologically extreme respondents. Less clear is whether they attracted strong partisans at the expense of Independent identifiers. While the Internet continues to provide an enticing and efficient mechanism for reaching young people, problems with opt-in panels suggest that serious attention needs to be given to improved ways of sampling.