Paul R. Yarnold & Ariel Linden
Optimal Data Analysis LLC & Linden Consulting Group LLC
This paper describes how an exploratory (post hoc, nondirectional, two-tailed) hypothesis involving a binary (dichotomous) class variable and a categorical ordinal (three-level) attribute is evaluated using MegaODA software using the new Stata package implementing ODA analysis.
Paul R. Yarnold & Ariel Linden
Optimal Data Analysis LLC & Linden Consulting Group LLC
This paper describes how an exploratory (post hoc, nondirectional, two-tailed) hypothesis involving a binary (dichotomous) class variable and a categorical ordinal (three-level) attribute is evaluated using MegaODA software using the new Stata package implementing ODA analysis.
Paul R. Yarnold & Ariel Linden
Optimal Data Analysis LLC & Linden Consulting Group LLC
This paper describes how a confirmatory (a priori, directional, one-tailed) hypothesis involving a binary (dichotomous) class variable and an ordinal (quintiles) attribute is evaluated using MegaODA software using the new Stata package implementing ODA analysis.
Paul R. Yarnold
Optimal Data Analysis, LLC
This study predicts change in magnitude of weekly RG-scores recorded over a year for a single individual. Attributes include weekly numbers of citations, recommendations, reads, and full-text downloads. Events in the transition table are weighted by the absolute product of ipsative standard scores for RG-score (treated as the class variable) and the attribute which is being assessed, thereby maximizing measurement precision, accuracy of prediction, and statistical power.
Paul R. Yarnold
Optimal Data Analysis, LLC
The study explores the serial nature of weekly RG-score values for a single individual over the period of one year. Events in the transition table are weighted by their corresponding change-in-value, thereby maximizing measurement precision and model accuracy.
Paul R. Yarnold
Optimal Data Analysis, LLC
Research Gate (RG) weekly summary statistics—including RG-score (the class or “dependent variable”), and number of citations, recommendations, and article views and downloads (the attributes or “independent variables”), were obtained for a single user. Single-case novometric classification tree analysis (CTA) was used to predict RG-score as a function of the number of citations, recommendations, and article views and downloads. Two analyses were conducted: one forced the model to have stable classification in leave-one-out (LOO) jackknife analysis; the second permitted jackknife instability so long as the LOO Type I error rate was p<0.05. A single-attribute model which achieved relatively strong LOO accuracy was identified.
Paul R. Yarnold & Ariel Linden
Optimal Data Analysis LLC & Linden Consulting Group LLC
This paper describes how an exploratory (i.e., post hoc, nondirectional, two-tailed) hypothesis involving a binary (i.e., dichotomous) class variable and an ordinal (rank) attribute is evaluated using MegaODA software vis-à-vis the new Stata package implementing ODA analysis.
