An Alternative Method for Assessing Goodness-of-Fit for Logit Models

Abstract

An inference strategy is presented to assess the goodness-of-fit for logit models. The conventional approach has emphasized likelihood ratio tests in which a distribution for the test statistic is assumed. A more recent development, the prediction success table, is based upon a ratio of predicted to observed choice patterns, yet it is primarily a descriptive index not placed within the context of an inference strategy/hypothesis-testing framework. An alternative method is put forward which uses residuals. Although problematic in a parametric context, residuals can be used in an approach which incorporates both a nonparametric randomization procedure and a sample reuse design. A test statistic is defined and compared to a reference distribution created under the assumption of random variation in the responses with variation in the explanatory variable(s). A numerical example is carried out to investigate the efficacy of using this approach.

Consumer choice of broiler meat: The effects of country of origin and production methods

“…The R 2 statistics was relatively good (R 2 = 0.45) (see, e.g. Costanzo, Halperin, Gale, & Richardson, 1982). Price had a significant negative effect, as expected, and the country of origin had a significant impact on the probability of choice.…”

The relevance of origin information at the point of sale

“…The models exhibit pseudo-R 2 values of 0.11183 (full sample), 0.1840 (Bavaria) and 0.0570 (Hessen and Thüringen); thus, only the model for the Bavarian survey region demonstrates relatively strong performance (Table 5). According to Costanzo, Halperin, Gale, and Richardson (1982), a pseudo-R 2 value of 0.2-0.4 signifies a satisfactory level of model estimation.…”

Human Dimensions Approach towards Integrating Sustainable Transportation and Urban Planning Policies: A Decision Support System (DSS) Based on Stated Preferences Data

“…The model adequacy was measured using Pseudo-R 2 (0), which reached a value of 0.22 and documents a good model fit [33]. In the models, a positive coefficient value indicates that the attribute level has a positive effect of choosing one alternative over another, whilst a negative coefficient value displays a negative effect.…”