In this paper I present a class of discrete choice models for ordinal response variables based on a generalization of the stereotype model. The stereotype model can be derived and generalized as a random utility model for ordered alternatives. Random utility models can be specified to account for heteroscedastic and correlated utilities. In the case of the generalized stereotype model this includes category-specific random effects due to individual differences in response style. But unlike standard random utility models the generalized stereotype model is better suited for ordinal response variables and can be interpreted as a kind of unidimensional unfolding model. This paper discusses the specification, interpretation, identification, and estimation of generalized stereotype models. Two applications are provided for illustration.

This paper argues that test data are ordinal, that latent trait scores are only determined ordinally, and that test data are used largely for ordinal purposes. Therefore it is desirable to develop a test theory based only on ordinal assumptions. A set of ordinal assumptions is presented, including an ordinal version of local independence. From these assumptions it is first shown that the gamma-correlation between two tests is the product of their gamma-correlations with the true latent order. The theory is generalized to allow for heterogeneous tests by defining a weighted average local independence. The tau-correlations between total score and the latent order can be found in both homogeneous and heterogeneous cases, and a system of differential item weighting to maximize the tau-correlation between weighted items and the latent order is provided. Thus a purely ordinal test theory seems possible.

This paper proposes a general approach to accounting for individual differences in the extreme response style in statistical models for ordered response categories. This approach uses a hierarchical ordinal regression modeling framework with heterogeneous thresholds structures to account for individual differences in the response style. Markov chain Monte Carlo algorithms for Bayesian inference for models with heterogeneous thresholds structures are discussed in detail. A simulation and two examples based on ordinal probit models are given to illustrate the proposed methodology. The simulation and examples also demonstrate that failing to account for individual differences in the extreme response style can have adverse consequences for statistical inferences.

Silent pauses are a natural part of speech production and have consequences for speech perception. However, studies have shown mixed results regarding whether listeners process pauses in native and non-native speech similarly or differently. A possible explanation for these mixed results is that perceptual consequences of pauses differ depending on the type of processing that listeners engage in: a focus on the content/meaning of the speech versus style/form of the speech. Thus, the present study examines the effect of silent pauses of listeners’ perception of native and non-native speech in two different tasks: the perceived credibility and the perceived fluency of the speech. Specifically, we ask whether characteristics of silent pauses influence listeners’ perception differently for native versus non-native speech, and whether this pattern differs when listeners are rating the credibility versus the fluency of the speech. We find that while native speakers are rated as more fluent than non-native speakers, there is no evidence that native speakers are rated as more credible. Our findings suggest that the way a non-native accent and disfluency together impact speech perception differs depending on the type of processing that listeners are engaged in when listening to the speech.

Determining appropriate topics and target audiences is essential to design effective educational outreach programs. Based on landowner responses to a mail survey, we determined both the importance and the availability of wildlife and forest management information topics to Mississippi landowners. Combining this information clearly identified the appropriate subject matter for outreach programs—topics important to landowners and for which information was relatively unavailable. The importance of wildlife and forest management information relative to its availability depended on the region, land use patterns, and landowner characteristics, thus demonstrating which segments of the population should be targeted to maximize program impact.

John Moody published his first railroad security analysis and ratings manual in April 1909. This study analyzes several current issues by looking back at Moody's original intentions for constructing a ratings system. The study analyzes whether Moody intended his ratings to reflect his private information, or rather, to serve some alternative role, as with monitoring conflicts of interests or realizing informational economies of scale. The study uses an ordinal regression approach to evaluate a set of explanatory variables, constructed from both the manual itself and the panic months of 1907, to test the potential information content of Moody's ratings. At the time of Moody's first rating system, the illiquidity of the US Treasury market forced investors to seek alternative ‘high-quality’ securities. Indeed, Moody rated 38.94 percent of railroad bonds as Aaa, and rated 85.25 percent of railroad bonds as A, Aa or Aaa in his universe of railroad bonds rated. To further test the informational content of Moody's ratings, the study pursues a structural default analysis during the panic year of 1907, which yields results that indicate that the default risk of railroad securities was quite low at the time. These results provide justification for the high overall ratings that Moody assigned to railroad securities, and thus their role as near risk-free securities. Therefore, railroad securities, and Moody's ratings, played a particularly important role in the financial system at the time.