Thinking about a summer vacation project? Academic prices for Maximizing Predictive Accuracy, and for MegaODA and CTA software, is now available to everyone: please see the Resources page.
Author Archives: paulyarnold
Computing Propensity Score Weights for CTA Models Involving Perfectly Predicted Endpoints
Paul R. Yarnold & Ariel Linden
Optimal Data Analysis, LLC & Linden Consulting Group, LLC
The use of CTA to construct propensity score weights is complicated by division by zero in models having any perfectly predicted endpoints: omitting undefined propensity scores yields a degenerate solution. This note presents an algorithmic remedy to this situation.
What is Optimal Data Analysis?
Paul R. Yarnold
Optimal Data Analysis, LLC
Preview begins by describing the ODA algorithm, requisite special-purpose software, and applied investigations using ODA to conduct statistical analysis. Discussion next addresses the development and application of multivariable linear and non-linear optimal models. Preview concludes by discussing current research and development foci, including causal inference methodology, system automation, new application arenas, and evolving methods and resources.
How to Find Articles in the ODA eJournal
Paul R. Yarnold
Optimal Data Analysis, LLC
This note offers tips on finding articles on specific topics in ODA.
Novometric Pairwise Comparisons in Consolidated Temporal Series
Paul R. Yarnold
Optimal Data Analysis, LLC
A total of 6,005 hospital patients rated their satisfaction with time taken for a nurse to respond to the call button (1=very dissatisfied; 2=somewhat dissatisfied; 3=neutral; 4=somewhat satisfied; 5=very satisfied). Responses were consolidated into four successive quarters (3rd and 4th Quarter of 1989, 1st and 2nd Quarter of 1990). Novometric analysis is used to evaluate statistical and ecological significance of all pairwise comparisons between the four quarters within this consolidated temporal series.
Novometric Comparison of Patient Satisfaction with Nurse Responsiveness Over Successive Quarters
Paul R. Yarnold
Optimal Data Analysis, LLC
A segmentation printout or “banner pass” constitutes a consolidated crosstabulation of user/customer responses to specific prompts, for desired reporting periods and/or organizational units. In this study a total of 6,005 hospital patients rated their satisfaction with time for a nurse to respond to the call button (1=very dissatisfied; 2=somewhat dissatisfied; 3=neutral; 4=somewhat satisfied; 5=very satisfied). Responses were consolidated into successive fiscal quarters (1=3rd Quarter, 1989; 2=4th Quarter, 1989; 3=1st Quarter, 1990; 4=2nd Quarter, 1990). Novometric analysis is used to evaluate the statistical and ecological significance of satisfaction gains observed over time.
Novometric Comparison of Markov Transition Matrices for Heterogeneous Populations
Paul R. Yarnold
Optimal Data Analysis, LLC
The American National Election Panel Study modeled transitions in social class (Working versus Middle) identification occurring between 1956, 1958, and 1960, and although a first-order Markov model was judged unsatisfactory, insufficient measurements were available to validate higher-order models. The sample was thus stratified with respect to whether respondents’ perceived their financial condition to have worsened or improved between data collection waves, in order to “…significantly decrease within-strata and increase between-strata heterogeneity of transition rates”. The probability of working class identification in 1958 was hypothesized to be greatest for respondents who “…felt their financial status declined, identification in 1956 held constant”, and was confirmed by visual examination. Non-directional omnibus chi-square analysis showed “stratification of the sample by perceived financial change (yielded) significantly different transition matrices” (chi-square=14.9, df=2, p<0.01). In contrast, novometric analysis found social class identification in 1958 was predicted by social class identification in 1956, and not by perceived financial status.
