170 likes | 193 Views
Explore the impact of statistics within education research, from reporting guidelines to practical applications and connections to scientific methodologies. Learn how to frame, measure, and analyze data effectively.
E N D
Using Statistics Effectively in Statistics Education Research Sterling C. Hilton Brigham Young University
OUTLINE • Background on SMER workshops • Summary of SMER final report • Connections to statistics education research • Discussion
BACKGROUND • Richard Scheaffer was PI on ASA sponsored study supported by NSF grant • Three workshops • Jan. 2005 • Sept. 2005 • March 2006 • 11 statisticians and 9 math education researchers
PURPOSE/MOTIVATION • No Child Left Behind Act • “scientifically based research” • “…involves the application of rigorous, systematic, and objective procedures to obtain reliable and valid knowledge…” • Call to improve mathematics education in K-12 grades (e.g. TIMMS 3) • Widespread use of qualitative research methods in mathematics education research
REPORT SUMMARY • Reporting guidelines for research program not research study • Goal to help build a useful and informative body of cumulative research • Help establish consistency in • Interventions • Measurement tools • Data collection • Data analysis • Reporting
REPORT SUMMARY • Influential documents • Scientific Research in Education (NRC, 2002) • Advancing Scientific Research in Education (NRC, 2005) • Mathematical Proficiency for All Students (RAND, 2003) • Influential perspective • Medical research model (clinical trials)
REPORT SUMMARY • Research program components (with characteristic activities and reporting guidelines) • Generate • Frame • Goals and Constructs • Measurement • Logistics and Feasibility • Examine • Generalize • Extend
REPORT SUMMARY • Generate ideas • Develop a conceptual framework • Consider own values, beliefs, biases • Characteristic activities • Identify interesting issues/questions • Search research literature • Build case for importance of issue • No reporting guidelines given
REPORT SUMMARY • Framing: Goals and Constructs • Characteristic activities • Formulate central ideas and underlying constructs • Formulate a research question • Identify relevant measures or need for new measures • Select and define research methods • Reporting guidelines • Define variables and measures used • State conjectures rather than causal statements • Exploratory and descriptive statistics
REPORT SUMMARY • Framing: Measurement • Validity, reliability, fairness • Characteristic activities • Examine previously used measures • Develop new measures • Test new measures • Reporting guidelines • Describe key details of development process • State validity, reliability and bias of measures
REPORT SUMMARY • Framing: Logistics and Feasibility • Characteristic activities • Formulate study design • Develop data collection protocols • Conduct pilot study • Reporting guidelines • Describe study design • Describe variables of interest • Describe the pilot tests of instruments / interventions
REPORT SUMMARY • Examine • If warranted, expansion of the pilot study to a larger community (single institution) • Main goal is to establish efficacy • Characteristic activities • Explore and define setting • Refine measures and protocols • Identify extraneous sources of variability • Reporting guidelines • Provide enough information to allow replication • Give estimates of parameters • Report formal statistical inference
REPORT SUMMARY • Generalize • If warranted, expansion of the study to larger community (multiple institutions) • Generally requires interdisciplinary work • Main goal is to generalize findings to larger population • Characteristic activities • Assess portability of measures and protocols • Design a multi-institutional study • Conduct rigorous statistical analysis • Reporting guidelines • Describe implementation • Describe efforts to minimize bias • Specify a statistical model
REPORT SUMMARY • Extend • If warranted, extend the study over time and to multiple communities • Generally requires interdisciplinary work • Main goal is to understand long-term effects • Characteristic activities • Design and conduct longitudinal study • Document the need for program improvements • Reporting guidelines • Describe implementation of longitudinal study • Describe dropout rate and its handling in the analysis
REPORT SUMMARY • Additional issues of interest • Stable unit treatment value assumption: the treatment is not dependent on the person administering it • The use of pre-post scores or gain scores is potentially problematic • Measurement development takes time and effort
REPORT SUMMARY • Connection to medical clinical trial model Generate Pre-clinical Frame Pre-clinical, Phase I Examine Phase I, Phase II Generalize Phase III Extend Phase IV
CONNECTIONS • Different research methods answer distinct research questions • Statistics education research is young (primarily in pre-clinical, phase I stages) • Researchers in statistics education need to: • talk the language of statisticians • allow the discipline to develop • not overstate their findings • have patience • persevere in doing quality research