Logistic Regression: Why We Cannot Do What We Think We Can Do, and What We Can Do About It

By C. Mood. 2010

3,178 citations recorded by the source corpus.

Subfields: Statistics and Probability, Statistical Learning Theory

How researchers see this paper

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

In Statistics and Probability

In Statistical Learning Theory

Citations in the corpus

Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI (2019, neutral): “Linked to this ﬁrst concern, [179] also states that a comparison between these ratios across models with diﬀerent variables might be problematic, since the unobserved heterogeneity is likely to vary, thereby invalidating the comparison.”
Continuation to upper secondary education in Finland: Children of immigrants and the majority compared (2011, methodology): “Therefore, log odds ratios are also included in the tables, as they have these characteristics, but are not comparable across models (Mood, 2010).”
Stock Market Literacy, Trust, and Participation (2015, neutral): “…di⁄erent predictor variables in the various probit model speci(cid:133)cations, the scaling of the response variable changes and therefore the changes in estimated coe¢ cients might not entirely be due to the suppressor e⁄ect (see Winship and Mare, 1984; Williams, 2009; Mood, 2010; among others).”
Boys’ and girls’ educational choices in secondary education. The role of gender ideology (2016, background): “These analyses give odds ratios, which are difficult to interpret due to unobserved heterogeneity (Mood 2010).”
Ordinal regression: A review and a taxonomy of models (2021, background): “Allison's article stimulated extensive research that aimed at avoiding such misleading results, see, for example, Williams (2009), Mood (2010), Karlson et al. (2012), Breen et al. (2014), Rohwer (2015), and Tutz (2019).”
Inaccurate Estimation of Disparities Due to Mischievous Responders (2014, methodology): “MAy 2014 175 Disparities were estimated using linear models or linear probability models, for ease of interpretation and comparability of scale across models with differing samples (Mood, 2010), with cluster-robust standard errors to account for the nesting of students within schools.”
Reputation Formation and the Evolution of Cooperation in Anonymous Online Markets (2014, positive): “…of covariates, we divide the process time into days and approximate the hazard rates using discrete-time linear probability models (Long 1997; Mood 2010) containing indicator variables for different segments of the process time (i.e., we model the baseline hazard rates shown in Figure 1 as…”
Early Career Outcomes for the “Best and the Brightest” (2015, methodology): “…regression models, as unlike ordinary least squares regression coefficients, comparisons of coefficients from model to model are affected both by the covariance structure of the variables as well as the scaling of the dependent variable under the logistic response assumptions (see Mood, 2010).”