Undergraduate Research Program in Statistics: Predicting Standardized Test Opt-Out in Michigan
During my junior year, myself and another undergraduate conducted preliminary research on missing education data for professor Ph.D. Ben Hansen’s use during an upcoming collaboration between the state of Michigan and the University of Michigan to analyze K-12 test score outcomes and implications.
Education providers, policymakers, and parents all use standardized state testing as an essential metric for evaluating student and school performances, and state testing remains one of the most common school-level metrics for comparison. State test results are often incorporated into formal and informal assessments of school and teacher success, and have informative potential when analyzed appropriately (educational citation). However, trends in scoring and the conclusions to which they may lead are vulnerable to selection bias in circumstances in which students were expected to take school standardized tests, but did not. This phenomenon of opting-out has been discussed as a response to various criticisms of statewide testing as well as school equity, and a major possible impact on holistic testing results.
The decision not to test encapsulates a variety of circumstances, including parental protest, performance pressure, illness, etc., therefore understanding trends in missing tests is an important prerequisite to any analysis of state testing scores. In this report, we sought to understand these missing tests by modeling the opt-out rates for state testing in Michigan for a given grade and school. Specifically, we investigate how grade level, school, and school district characteristics influence the rate of students expected to opt-out in Michigan using a variety of modeling methods such as Weighted Linear Regression, Ridge Regression, and Linear Mixed Models.
This investigation focused on the Michigan Student Test of Educational Progress (M-STEP), which the state of Michigan administers to students in grades 3rd-8th, and 11th, and the Michigan Merit Exam (MME) which is given to 11th-grade students. Any 12th-grade students who have not completed the required 11th-grade testing (either M-STEP or MME) must do so during their 12th-grade year.
Maple River Dam: Post-Removal Succession
In the summer between my sophomore and junior year, I spent a month at the University of Michigan Biological Station assisting with research and taking a class taught by Ph.D. Paul Moore.
I collected data in the field, ran chemical laboratory analyses, and conducted my own statistical analysis of the data compared to pre-removal sampling. From this research, I gained experience in algae sampling, macroinvertebrate sampling, water sampling, and the use of a flow meter and hydro lab.
Understanding the impacts of human intervention on the natural world is paramount to advising policy and environmental management decisions, particularly in the era of climate change. One of the foremost ways humans have impacted the biosphere is through severe habitat fragmentation. Dams are notorious for creating impassible reaches of habitat and changing the natural state of freshwater environments. This has cascading effects on the ecological health of an ecosystem, both through its biotic and abiotic factors. While dams have long been an effective means of flood control, river flow, and river restoration, there are compelling arguments for dam removal. These arguments, along with advances in ecological thinking and the emergence of the “Reconnecting Rivers” movement, led to the removal of the Maple River Dam in Pellston, Michigan in 2018. There has been increased interest in dam removal as a method for promoting more connected, biodiverse ecosystems. A comparison of data from before and after a major dam removal allows for the examination of the ecological consequences of such human ecosystem manipulation.
As part of a larger before-and-after study of the limnological conditions and species abundances in the Maple River (Boehm, 2015), the data collected has been compared to East, West, and Main Branch data from before the dam was removed. Quantifying the effects of ecosystem disturbance related to dam removal may be useful in establishing expectations for ecosystem recovery and succession elsewhere. Therefore, it is important to understand the effects of dam removal on various species, specifically macroinvertebrate populations, which are often used as indicators of the health and status of lotic ecosystems (Monaghan, 2012), as well as any changes in physical or chemical characteristics related to human intervention. This study investigated the effects of dam removal on the macroinvertebrate Functional Feeding Group (FFG) richness and distribution, discharge, and concentrations of nitrogen and phosphorus in the Maple River.
This study found notable differences between the state of the Maple River prior to dam removal and its current state, post dam removal. These differences primarily pertain to discharge, water chemistry, and P/R ratios, and are most prominent at sites downstream nearest to the former dam site, in accordance with the river continuum concept.
Lost in Translation: how the dominance of English in science inhibits conservation and sustainability
As an undergraduate, I have been working on a project with a co-author and post-doctoral research fellow in the School for Environment and Sustainability to investigate the effects of language barriers on conservation efforts in Latin America.
As issues regarding sustainability increase in urgency, access to environmental knowledge becomes paramount for governments and industries hoping to mitigate catastrophe. However, disparities in the language in which science is published impede the exchange of information between international scientists. English is the language of publication for the majority of conservation-related journals, and its dominance poses challenges for conservation science in the non-English-speaking world; missed opportunities for sharing important results and local knowledge can ultimately undermine conservation effectiveness. Given that twenty-eight of thirty-six global biodiversity hotspots are located in non-English-speaking regions, ensuring that relevant ecological knowledge is available to decision makers is critical.
Latin America, with sixty percent of Earth’s terrestrial life and nearly 900 million hectares of forest cover, is a region whose preservation is vital for global environmental stability due to its rich biodiversity and function as a carbon sink. The promotion of conservation and sustainable development in this region is unequivocally critical. However, research on the distribution of conservation journals by language and region reveals that when compared to English publications, Latin American Spanish publications are cited less often, suggesting that English-dominated conservation researchers do not reference local environmental knowledge due to language. The overall aim of this research was an analysis of a citation network comparing English and Spanish publications about Colombian sustainability issues, and understanding how access to scientific journals by different stakeholders can affect information dissemination and knowledge sharing in conservation science.
