Improving IRT parameter estimates with small sample sizes: Evaluating the efficacy of a new data augmentation technique

The 3PL model is a flexible and widely used tool in assessment. However, it suffers from limitations due to its need for large sample sizes. This study introduces and evaluates the efficacy of a new sample size augmentation technique called Duplicate, Erase, and Replace (DupER) Augmentation through a simulation study. Data are augmented using several variations of DupER Augmentation (based on different imputation methodologies, deletion rates, and duplication rates), analyzed in BILOG-MG 3, and results are compared to those obtained from analyzing the raw data. Additional manipulated variables include test length and sample size. Estimates are compared using seven different evaluative criteria. Results are mixed and inconclusive. DupER augmented data tend to result in larger root mean squared errors (RMSEs) and lower correlations between estimates and parameters for both item and ability parameters. However, some DupER variations produce estimates that are much less biased than those obtained from the raw data alone. For one DupER variation, it was found that DupER produced better results for low-ability simulees and worse results for those with high abilities. Findings, limitations, and recommendations for future studies are discussed. Specific recommendations for future studies include the application of Duper Augmentation (1) to empirical data, (2) with additional IRT models, and (3) the analysis of the efficacy of the procedure for different item and ability parameter distributions.

Ecology, stable isotopes, and management of grassland songbirds at National Park Service properties on the Great Plains

Grassland ecosystems have been severely reduced and grassland bird populations have experienced consistent declines. National Park Service (NPS) properties on the Great Plains provide breeding habitat for grassland songbirds, though little is known about the quality of this habitat. A short-term study on songbirds at three NPS properties complemented current monitoring, providing an among park comparison addressing grassland bird productivity and fidelity relative to NPS property size. During 2008-2009, I assessed avian species richness, and estimated bird density and grassland songbird nest success. Bird species richness was greatest at small and medium sites, while number of nesting obligate species was greatest at the large site. Species-specific densities varied among sites, with few grassland obligates occurring at all three sites. Nest success estimates for grassland obligates were highest at the small site and lower at the large site. Another method to quantify habitat quality is assessment of breeding site fidelity. Current extrinsic markers used in monitoring site fidelity are inadequate for small birds; stable isotope analyses provide an alternative. I compared two techniques for assigning stable isotope tissue origin and measured grassland songbird site fidelity. My method of assigning origin provided site-specific variances of expected stable isotope values, an improvement over the most commonly used method. Fidelity tended to be higher at the large site, which may indicate a more robust breeding community of grassland birds. The small size of two of my sites precluded large sample sizes and made strong inferences difficult. To quantify how scientists cope with weak inference, I conducted a literature review. Strong inference was rarely observed, and most authors of weak-inference papers provided specific management recommendations. I suggest that adaptive management is an ideal method to resolve uncertainty from weak inference. Managers should consider my results within the context of regional and global management and the extent to which their unit might aide songbird conservation.