David A. Klingbeil
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
David A. Klingbeil has a total of 115 co-authors. Here are the top co-authors by number of shared publications:
Ethan R. Van Norman
College of Education, Lehigh University, USA
15 shared publicationsPeter M. Nelson
ServeMinnesota, Minneapolis, MN, USA
12 shared publicationsDavid C. Parker
ServeMinnesota, Minneapolis, MN, USA
7 shared publications- Ethan R. Van Norman
Department of Education and Human Services Lehigh University Bethlehem Pennsylvania USA
7 shared publications - Bonnie Klimes-Dougan
Department of Psychology University of Minnesota Minneapolis Minnesota USA
4 shared publications - Matthew K. Burns
University of Florida
4 shared publications
Latest Publications
- 8 November 2026
- Published by SAGE Publications in Assessment for Effective Intervention
Ignorance Is Bliss: Evaluating the Impact of Priors, Correlations, and Base Rates in Universal Screening Cost–Accuracy Analysis
Universal screening in schools is a costly endeavor, but the use of cost–accuracy ratios may augment its positive impacts on school-based prevention. In addition, educational researchers have more recently begun to adopt Bayesian methods of calculating screener accuracy, such as post-test probabilities or fully Bayesian statistical models. However, there is no research evaluating how screener costs intersect with the properties of Bayesian statistical modeling. In this study, we conducted a simulation experiment of how priors within Bayesian logistic regression influenced the estimation of screening costs as an interaction between both the true correlation of a screener (e.g., from the normative sample) and base rates of actual risk. We found that the cost of falsely predicting a student was at risk was substantially higher with a lower base rate of .2, regardless of the strength of the true correlation between the screener and criterion outcome. Our results also demonstrated that costs were most sensitive to informative priors, even when the priors were designed to be very similar to the true simulation values. Results demonstrated the practical application of Bayesian analyses for evaluating how prior assumptions can influence screening cost estimation, and we provide practical guidance for how these methods can benefit practitioners and policymakers.
- 8 November 2026
- Published by SAGE Publications in Assessment for Effective Intervention
Ignorance Is Bliss: Evaluating the Impact of Priors, Correlations, and Base Rates in Universal Screening Cost–Accuracy Analysis
Universal screening in schools is a costly endeavor, but the use of cost–accuracy ratios may augment its positive impacts on school-based prevention. In addition, educational researchers have more recently begun to adopt Bayesian methods of calculating screener accuracy, such as post-test probabilities or fully Bayesian statistical models. However, there is no research evaluating how screener costs intersect with the properties of Bayesian statistical modeling. In this study, we conducted a simulation experiment of how priors within Bayesian logistic regression influenced the estimation of screening costs as an interaction between both the true correlation of a screener (e.g., from the normative sample) and base rates of actual risk. We found that the cost of falsely predicting a student was at risk was substantially higher with a lower base rate of .2, regardless of the strength of the true correlation between the screener and criterion outcome. Our results also demonstrated that costs were most sensitive to informative priors, even when the priors were designed to be very similar to the true simulation values. Results demonstrated the practical application of Bayesian analyses for evaluating how prior assumptions can influence screening cost estimation, and we provide practical guidance for how these methods can benefit practitioners and policymakers.
- Open Access
- 17 September 2025
- Published by Wiley in Psychology in the Schools
- Vol. 63 (2) , 423-432
Exploration of the Predictive Validity of Middle School Math Screeners
Universal screening is a key element of multi‐tiered systems of support. Relative to literacy, fewer studies have investigated the psychometric properties of math assessments particularly amongst middle school students. We investigated whether outcomes from two types of curriculum‐based measures (CBMs), computational fluency (M‐COMP) and concepts and application (M‐CAP) were predictive of end of year test scores after accounting for performance on a computer adaptive test of broad math performance (Measures of Academic Progress—MAP). We used quantile regression to determine whether M‐COMP and M‐CAP were differentially important in predicting end of year scores as a function of overall student skill level (e.g., students performing far below expectations) after accounting for MAP scores. Including M‐COMP consistently improved predictions of end of year state test scores for grade six students scoring below the 25th percentile on an end of year achievement test. M‐COMP and M‐CAP provided limited additional predictive utility across skill levels for students in grades seven and eight.
- 17 September 2025
- Published by Wiley in Psychology in the Schools
- Vol. 63 (2) , 423-432
Open AccessExploration of the Predictive Validity of Middle School Math Screeners
Universal screening is a key element of multi‐tiered systems of support. Relative to literacy, fewer studies have investigated the psychometric properties of math assessments particularly amongst middle school students. We investigated whether outcomes from two types of curriculum‐based measures (CBMs), computational fluency (M‐COMP) and concepts and application (M‐CAP) were predictive of end of year test scores after accounting for performance on a computer adaptive test of broad math performance (Measures of Academic Progress—MAP). We used quantile regression to determine whether M‐COMP and M‐CAP were differentially important in predicting end of year scores as a function of overall student skill level (e.g., students performing far below expectations) after accounting for MAP scores. Including M‐COMP consistently improved predictions of end of year state test scores for grade six students scoring below the 25th percentile on an end of year achievement test. M‐COMP and M‐CAP provided limited additional predictive utility across skill levels for students in grades seven and eight.
- Open Access
- 14 November 2024
- Published by Wiley in Psychology in the Schools
- Vol. 62 (3) , 721-733
Base Rates of Non‐Proficiency in Reading and Math: Implications for Universal Screening
Several researchers have called for schools to interpret universal screening results using posterior probabilities. Following this recommendation could require schools to move away from direct‐route, single‐measure screening unless base rates of risk fall within a narrow range. In this descriptive study, we investigated two questions surrounding the feasibility of adopting a probability‐based approach in schools. First, how many districts had base rates of risk that would allow educators to classify all students using a single screener? Second, to what extent would yearly fluctuation in base rates require planned flexibility in screening procedures? We used public state test data to estimate base rates of risk for each district in Florida, Oregon, and Wisconsin across four school years. Districts' base rates of risk were rarely within the required range for a single screener to classify all students based on posterior probabilities. Many districts had base rates conducive to direct route screening in at least 1 year but very few had base rates in the required zone across the 4 years. Using posterior probabilities to make screening decisions may require more changes to existing practices than simply interpreting screening results as probabilities. Additional research to promote the use of this approach in schools is needed.
- 14 November 2024
- Published by Wiley in Psychology in the Schools
- Vol. 62 (3) , 721-733
Open AccessBase Rates of Non‐Proficiency in Reading and Math: Implications for Universal Screening
Several researchers have called for schools to interpret universal screening results using posterior probabilities. Following this recommendation could require schools to move away from direct‐route, single‐measure screening unless base rates of risk fall within a narrow range. In this descriptive study, we investigated two questions surrounding the feasibility of adopting a probability‐based approach in schools. First, how many districts had base rates of risk that would allow educators to classify all students using a single screener? Second, to what extent would yearly fluctuation in base rates require planned flexibility in screening procedures? We used public state test data to estimate base rates of risk for each district in Florida, Oregon, and Wisconsin across four school years. Districts' base rates of risk were rarely within the required range for a single screener to classify all students based on posterior probabilities. Many districts had base rates conducive to direct route screening in at least 1 year but very few had base rates in the required zone across the 4 years. Using posterior probabilities to make screening decisions may require more changes to existing practices than simply interpreting screening results as probabilities. Additional research to promote the use of this approach in schools is needed.
- 1 November 2024
- Published by American Psychological Association (APA) in School Psychology
- Vol. 39 (6) , 579-588
Effects of portable interventions on school psychologists’ graph-rating inconsistency.
The visual analysis of data presented in time-series graphs are common in single-case design (SCD) research and applied practice in school psychology. A growing body of research suggests that visual analysts' ratings are often influenced by construct-irrelevant features including Y-axis truncation and compression of the number of data points per X- to Y-axis ratio. We developed and tested two brief interventions, based on the research in cognitive and visual science, to reduce visual analysts' inconsistency when viewing unstandardized graphs. Two hundred practicing school psychologists visually analyzed data presented on standardized graphs and the same data again on unstandardized graphs. Across all conditions, participants were more willing to identify meaningful effects on unstandardized graphs and rated the data as showing significantly larger effects than on the corresponding standardized graphs. However, participants who answered additional (task-relevant) questions about the level or trend of graphed data showed greater rating consistency across the types of graphs in comparison to participants who answered task-irrelevant but challenging questions or control participants. Our results replicated prior research demonstrating the impact of SCD graph construction on practicing school psychologists' interpretations and provide initial support for an intervention to minimize the impact of construct-irrelevant factors. Limitations and future directions for research are discussed. (PsycInfo Database Record (c) 2024 APA, all rights reserved).
- 1 November 2024
- Published by American Psychological Association (APA) in School Psychology
- Vol. 39 (6) , 579-588
Effects of portable interventions on school psychologists’ graph-rating inconsistency.
The visual analysis of data presented in time-series graphs are common in single-case design (SCD) research and applied practice in school psychology. A growing body of research suggests that visual analysts' ratings are often influenced by construct-irrelevant features including Y-axis truncation and compression of the number of data points per X- to Y-axis ratio. We developed and tested two brief interventions, based on the research in cognitive and visual science, to reduce visual analysts' inconsistency when viewing unstandardized graphs. Two hundred practicing school psychologists visually analyzed data presented on standardized graphs and the same data again on unstandardized graphs. Across all conditions, participants were more willing to identify meaningful effects on unstandardized graphs and rated the data as showing significantly larger effects than on the corresponding standardized graphs. However, participants who answered additional (task-relevant) questions about the level or trend of graphed data showed greater rating consistency across the types of graphs in comparison to participants who answered task-irrelevant but challenging questions or control participants. Our results replicated prior research demonstrating the impact of SCD graph construction on practicing school psychologists' interpretations and provide initial support for an intervention to minimize the impact of construct-irrelevant factors. Limitations and future directions for research are discussed. (PsycInfo Database Record (c) 2024 APA, all rights reserved).
- Open Access
- 22 August 2024
- Published by Center for Open Science in PsyArXiv
Flexible distributional models for meta-analysis of reading fluency outcomes from single-case designs: An examination using Bayesian methods
There is growing interest in statistical modeling of data from single-case design (SCD) research. However, currently available methods such as hierarchical linear models and generalized linear mixed models have assumptions that may limit their utility for applied SCDs, such as those that use curriculum-based measures of academic performance as outcomes. We demonstrate use of a flexible class of distributional models, known as generalized additive linear mixed models for location, scale, and shape (GAMLSS), to evaluate different distributional families and modeling specifications for reading curriculum-based measures of reading fluency data drawn from SCD studies of academic interventions. Using Bayesian methods and graphical posterior predictive checks, we evaluate GAMLSS based on normal (Gaussian), Poisson, and negative binomial distributional families. We also evaluate the extent to which the dispersion, or variability of outcomes, itself varies across studies and across participants within them. We find that negative binomial models with heterogeneous dispersions fit better than other distributional families and closely reproduce features of the observed data. Findings highlight the need to consider a broader set of distributional families when developing meta-analytic models of SCD data, as well as to consider how the degree of dispersion may vary from study to study. We discuss implications for future methodological research and for meta-analysis of SCDs.
- 22 August 2024
- Published by Center for Open Science in PsyArXiv
Open AccessFlexible distributional models for meta-analysis of reading fluency outcomes from single-case designs: An examination using Bayesian methods
There is growing interest in statistical modeling of data from single-case design (SCD) research. However, currently available methods such as hierarchical linear models and generalized linear mixed models have assumptions that may limit their utility for applied SCDs, such as those that use curriculum-based measures of academic performance as outcomes. We demonstrate use of a flexible class of distributional models, known as generalized additive linear mixed models for location, scale, and shape (GAMLSS), to evaluate different distributional families and modeling specifications for reading curriculum-based measures of reading fluency data drawn from SCD studies of academic interventions. Using Bayesian methods and graphical posterior predictive checks, we evaluate GAMLSS based on normal (Gaussian), Poisson, and negative binomial distributional families. We also evaluate the extent to which the dispersion, or variability of outcomes, itself varies across studies and across participants within them. We find that negative binomial models with heterogeneous dispersions fit better than other distributional families and closely reproduce features of the observed data. Findings highlight the need to consider a broader set of distributional families when developing meta-analytic models of SCD data, as well as to consider how the degree of dispersion may vary from study to study. We discuss implications for future methodological research and for meta-analysis of SCDs.
- 22 August 2024
- Published by Elsevier in Journal of School Psychology
- Vol. 106, 101356
Reading skill profiles of dysfluent readers in grades 2 and 3
- 22 August 2024
- Published by Elsevier in Journal of School Psychology
- Vol. 106, 101356
Reading skill profiles of dysfluent readers in grades 2 and 3
- 25 January 2024
- Published by SAGE Publications in Assessment for Effective Intervention
- Vol. 49 (3) , 148-158
The Influence of English Learner Status on Maintenance of Oral Reading Fluency Growth
Relatively little is known regarding post-intervention reading fluency outcomes for English learners (ELs) in comparison with non-EL peers, yet educators must be prepared to sustain growth for all students transitioning to less-intensive tiers of support. The purpose of this study was to investigate whether EL status moderated post-intervention maintenance effects among U.S. second- and third-grade students who transitioned back to Tier 1 instruction only due to successful performance during a Tier 2 reading fluency intervention. Piecewise multi-level models were estimated to address whether EL status uniquely predicted intervention growth patterns and the extent to which these patterns were maintained over a 12- to 13-week post-intervention period. Reading fluency scores were similar between EL and non-EL students prior to the start of and during the intervention, and all students’ performance decreased slightly immediately after support ceased. Regardless of grade level or EL status, post-intervention fluency gains generally remained smaller than those observed during intervention meriting attention to individual- and systems-level instructional considerations for ensuring continued growth.
- 25 January 2024
- Published by SAGE Publications in Assessment for Effective Intervention
- Vol. 49 (3) , 148-158
The Influence of English Learner Status on Maintenance of Oral Reading Fluency Growth
Relatively little is known regarding post-intervention reading fluency outcomes for English learners (ELs) in comparison with non-EL peers, yet educators must be prepared to sustain growth for all students transitioning to less-intensive tiers of support. The purpose of this study was to investigate whether EL status moderated post-intervention maintenance effects among U.S. second- and third-grade students who transitioned back to Tier 1 instruction only due to successful performance during a Tier 2 reading fluency intervention. Piecewise multi-level models were estimated to address whether EL status uniquely predicted intervention growth patterns and the extent to which these patterns were maintained over a 12- to 13-week post-intervention period. Reading fluency scores were similar between EL and non-EL students prior to the start of and during the intervention, and all students’ performance decreased slightly immediately after support ceased. Regardless of grade level or EL status, post-intervention fluency gains generally remained smaller than those observed during intervention meriting attention to individual- and systems-level instructional considerations for ensuring continued growth.
- Open Access
- 6 December 2023
- Published by Wiley in Psychology in the Schools
- Vol. 61 (4) , 1458-1473
Multiple baseline and multiple probe design studies targeting academic skills: Trends over time in effect sizes
Single‐case design (SCD) is a quantitative experimental technique in which participants serve as their own control. The use of an effect size in SCD allows evaluation of outcomes as well as comparison of outcomes via meta‐analyses. Characteristics of SCD research make the selection of an appropriate effect size complicated. Additionally, there are a number of factors that complicate the use of SCDs as a means to improve academic skills. The purpose of this study was to examine patterns by which SCD effect sizes are used to quantify outcomes from academic interventions. To do so, a descriptive analysis of an extant database of SCD studies was conducted. The authors created frequency tables for each effect size identified in the database as well as graphs to show the extent to which the use of effect sizes changed over time. The authors also determined whether the most frequently used effect sizes were appropriate for summarizing changes in academic outcomes. Although many effect sizes have been developed, only a small number are routinely used in SCD research for academic skills. The most frequently used effect sizes were those that come from standardized statistics, compared with those that utilized principles of regression or Bayesian analysis.
- 6 December 2023
- Published by Wiley in Psychology in the Schools
- Vol. 61 (4) , 1458-1473
Open AccessMultiple baseline and multiple probe design studies targeting academic skills: Trends over time in effect sizes
Single‐case design (SCD) is a quantitative experimental technique in which participants serve as their own control. The use of an effect size in SCD allows evaluation of outcomes as well as comparison of outcomes via meta‐analyses. Characteristics of SCD research make the selection of an appropriate effect size complicated. Additionally, there are a number of factors that complicate the use of SCDs as a means to improve academic skills. The purpose of this study was to examine patterns by which SCD effect sizes are used to quantify outcomes from academic interventions. To do so, a descriptive analysis of an extant database of SCD studies was conducted. The authors created frequency tables for each effect size identified in the database as well as graphs to show the extent to which the use of effect sizes changed over time. The authors also determined whether the most frequently used effect sizes were appropriate for summarizing changes in academic outcomes. Although many effect sizes have been developed, only a small number are routinely used in SCD research for academic skills. The most frequently used effect sizes were those that come from standardized statistics, compared with those that utilized principles of regression or Bayesian analysis.
- 22 November 2022
- Published by SAGE Publications in Journal of Psychoeducational Assessment
- Vol. 41 (2) , 194-208
Comparing AimswebPlus to the Benchmark Assessment System for Universal Screening in Upper Elementary Grades
Early identification of students needing additional support is a foundational component of Multi-Tiered Systems of Support (MTSS). Due to the resource-intensive nature of implementing MTSS, it is critical that universal screening procedures are maximally accurate and efficient. The purpose of this study was to compare the classification accuracy of aimswebPlus reading scores to the Benchmark Assessment System scores. We used data from a mid-size city in Texas to retrospectively compare the classification accuracy between fall aimswebPlus reading composites to the Benchmark Assessment System scores when predicting student performance on the statewide reading test. When classification decisions were made based on the vendor-recommended cut-scores, both measures were insufficiently sensitive for screening in MTSS. Following aimswebPlus’ recommended method for establishing local-cut scores improved the sensitivity of decisions, but the specificity values were well below minimally acceptable levels. Limitations, directions for future research, and implications for practice are discussed.
- 22 November 2022
- Published by SAGE Publications in Journal of Psychoeducational Assessment
- Vol. 41 (2) , 194-208
Comparing AimswebPlus to the Benchmark Assessment System for Universal Screening in Upper Elementary Grades
Early identification of students needing additional support is a foundational component of Multi-Tiered Systems of Support (MTSS). Due to the resource-intensive nature of implementing MTSS, it is critical that universal screening procedures are maximally accurate and efficient. The purpose of this study was to compare the classification accuracy of aimswebPlus reading scores to the Benchmark Assessment System scores. We used data from a mid-size city in Texas to retrospectively compare the classification accuracy between fall aimswebPlus reading composites to the Benchmark Assessment System scores when predicting student performance on the statewide reading test. When classification decisions were made based on the vendor-recommended cut-scores, both measures were insufficiently sensitive for screening in MTSS. Following aimswebPlus’ recommended method for establishing local-cut scores improved the sensitivity of decisions, but the specificity values were well below minimally acceptable levels. Limitations, directions for future research, and implications for practice are discussed.
Effectiveness of accelerated reader on children's reading outcomes: A meta‐analytic review
Accelerated reader (AR) is a computerized reading program commonly used in schools. The program aims to enhance students' reading achievement and encourage students to read more through goal setting and frequent reading practice. A meta-analytic review of the AR was conducted to analyse its effectiveness as an evidence-based intervention for improving student reading achievement, attitude, and motivation. This study investigated potential moderating variables, including publication type, participant, and study characteristics that impact student reading outcomes. A total of 44 studies from peer-reviewed journal articles and dissertations met the inclusion criteria. Participants included 16,653 students enrolled in elementary, middle, and high school. Hedges' g effect sizes measures suggest pretest-posttest one-group AR studies have moderate effects (g = 0.541) while comparison group AR studies have marginal effects (g = 0.278). A meta-regression model of six potential categorical moderators of comparison group studies indicted no significant moderators. Implications and the need for further research regarding evidence-based and culturally appropriate reading interventions are discussed.
Effectiveness of accelerated reader on children's reading outcomes: A meta‐analytic review
Accelerated reader (AR) is a computerized reading program commonly used in schools. The program aims to enhance students' reading achievement and encourage students to read more through goal setting and frequent reading practice. A meta-analytic review of the AR was conducted to analyse its effectiveness as an evidence-based intervention for improving student reading achievement, attitude, and motivation. This study investigated potential moderating variables, including publication type, participant, and study characteristics that impact student reading outcomes. A total of 44 studies from peer-reviewed journal articles and dissertations met the inclusion criteria. Participants included 16,653 students enrolled in elementary, middle, and high school. Hedges' g effect sizes measures suggest pretest-posttest one-group AR studies have moderate effects (g = 0.541) while comparison group AR studies have marginal effects (g = 0.278). A meta-regression model of six potential categorical moderators of comparison group studies indicted no significant moderators. Implications and the need for further research regarding evidence-based and culturally appropriate reading interventions are discussed.