Splitting a Predictor at the Upper Quarter or Third and the Lower Quarter or Third

By Andrew Gelman, David K. Park. 2009

208 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In General Decision Sciences

Citations in the corpus

Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions (2018, background): “Defining an outcome as binary rather than multinomial or continuous can lose information [63].”
Gaydar and the Fallacy of Decontextualized Measurement (2018, background): “A guiding principle when trying to split a population into two groups is that you always need a third group to account for those individuals that are within a margin of error (Gelman and Park, 2009).”
Simple rules to guide expert classifications (2020, methodology): “Further, all continuous features are discretized into three approximately equal-sized bins representing (categorical) low, medium, and high values of the feature, following Gelman and Park (2009).”
Preference for Human (vs. Robotic) Labor is Stronger in Symbolic Consumption Contexts (2020, background): “Following Gelman and Park (2009), the lower panel of Figure 2 shows that the interaction between production mode and symbolic value is signi ﬁ cant for participants high in need for uniqueness ( F (1, 136) = 10.31, p < .01), but not for those low in need for uniqueness ( F (1, 126) = 0.48, p = .49).”
Models as Approximations - A Conspiracy of Random Regressors and Model Deviations Against Classical Inference in Regression (2015, background): “• Marginal centering, unconditional, is a special case of (15) due to the inclusion of an intercept in ~ X: (16) E[ η ] = E[ ] = E[ δ ] = 0.”
Models as Approximations: How Random Predictors and Model Violations Invalidate Classical Inference in Regression (2014, unknown): “Using E[ δ ~ X ] = 0 from (12) we have: N1/2 (β̂ − β) = ( 1 N XX )−1 ( 1 N1/2 Xδ ) = ( 1 N ∑ ~ Xi ~ X i )−1 ( 1 N1/2 ∑ ~ Xi δi) D −→ E[ ~ X ~ X ]−1N ( 0,E[ δ2 ~ X ~ X T ] ) = N ( 0,E[ ~ X ~ X T ]−1E[ δ2 ~ X ~ X T ]E[ ~ X ~ X T ]−1 ) , Note: The center parts of the first two asymptotic sandwich covar”
Two Lines: A Valid Alternative to the Invalid Testing of U-Shaped Relationships With Quadratic Regressions (2018, unknown): “See e.g., Gelman and Park (2008).”
Models as Approximations II: A Model-Free Theory of Parametric Regression (2016, background): “3 Given a variable Z, consider for concreteness a univariate Gaussian weight function of Z, centered at ξ on the Z axis: (8) wξ(z) = w ∗ ξ (z)/E[w ∗ ξ (Z)] , w ∗ ξ (z) ∝ exp(−(z − ξ)(2)/(2γ(2))) , (3)Mathematically, the restriction to weights as a function of univariate variables Z is no restriction”
The Median Split: Robust, Refined, and Revived (2015, background): “• The Gelman and Park (2009) idea of splitting a variable into three parts is fine, but note the relative efficiencies being discussed compare their dropping the middle part of a distribution of data, which reduces the effective sample size.”
You can win but I can't lose: Bias against high-status groups increases their zero-sum beliefs about discrimination (2015, unknown): “To probe the four-way interaction, we created a variable that examined the upper and lower quartiles of ZSBs (Gelman & Park, 2009).”
Is Augmented Reality Technology an Effective Tool for E-commerce? An Interactivity and Vividness Perspective (2017, methodology): “To identify the effect of high vs. low previous media experience, we used the method suggested by Gelman and Park (2009).”
A researcher's guide to regression, discretization, and median splits of continuous variables (2015, methodology): “In this setting, the approach of Gelman and Park (2009) of splitting the data into three groups and discarding the middle group seems even more advantageous for researchers considering dichotomization.”
Models as Approximations I: Consequences Illustrated with Linear Regression (2014, methodology): “y, a linear t can be used to infer the overall direction of the association between the response and the predictors. The above formulas were used and modied to produce alternative slope estimates by Gelman and Park (2008), also with the \Goal of Expressing Regressions as Comparisons that can be Unde”