Background: The analysis of correlated binary data is commonly addressed through the use of conditional models with random effects included in the systematic component as opposed to generalized estimating equations (GEE) models that addressed the random component. Since the joint distribution of the observations is usually unknown, the conditional distribution is a natural approach. Our objective was to compare the fit of different binary models for correlated data in tobacco use. We advocate that the joint modeling of the mean and dispersion may be at times just as adequate. We assessed the ability of these models to account for the intraclass correlation. In so doing, we concentrated on fitting logistic regression models to address smoking behaviors.

Methods: Frequentist and Bayes’ hierarchical models were used to predict conditional probabilities, and the joint modeling (GLM and GAM) models were used to predict marginal probabilities. These models were fitted to National Longitudinal Study of Adolescent to Adult Health (Add Health) data for tobacco use.

Results: We found that people were less likely to smoke if they had higher income, high school or higher education and religious. Individuals were more likely to smoke if they had abused drug or alcohol, spent more time on TV and video games, and been arrested. Moreover, individuals who drank alcohol early in life were more likely to be a regular smoker. Children who experienced mistreatment from their parents were more likely to use tobacco regularly.

Conclusions: The joint modeling of the mean and dispersion models offered a flexible and meaningful method of addressing the intraclass correlation. They do not require one to identify random effects nor distinguish from one level of the hierarchy to the other. Moreover, once one can identify the significant random effects, one can obtain similar results to the random coefficient models. We found that the set of marginal models accounting for extravariation through the additional dispersion submodel produced similar results with regards to inferences and predictions. Moreover, both marginal and conditional models demonstrated similar predictive power.

It is well understood that innovation drives productivity growth in agriculture. Innovation, however, is a process that involves activities distributed throughout the supply chain. In this dissertation I investigate three topics that are at the core of the distribution and diffusion of innovation: optimal licensing of university-based inventions, new variety adoption among farmers, and consumers’ choice of new products within a social network environment.

University researchers assume an important role in innovation, particularly as a result of the Bayh-Dole Act, which allowed universities to license inventions funded by federal research dollars, to private industry. Aligning the incentives to innovate at the university level with the incentives to adopt downstream, I show that non-exclusive licensing is preferred under both fixed fee and royalty licensing. Finding support for non-exclusive licensing is important as it provides evidence that the concept underlying the Bayh-Dole Act has economic merit, namely that the goals of university-based researchers are consistent with those of society, and taxpayers, in general.

After licensing, new products enter the diffusion process. Using a case study of small holders in Mozambique, I observe substantial geographic clustering of new-variety adoption decisions. Controlling for the other potential factors, I find that information diffusion through space is largely responsible for variation in adoption. As predicted by a social learning model, spatial effects are not based on geographic distance, but rather on neighbor-relationships that follow from information exchange. My findings are consistent with others who find information to be the primary barrier to adoption, and means that adoption can be accelerated by improving information exchange among farmers.

Ultimately, innovation is only useful when adopted by end consumers. Consumers’ choices of new products are determined by many factors such as personal preferences, the attributes of the products, and more importantly, peer recommendations. My experimental data shows that peers are indeed important, but “weak ties” or information from friends-of-friends is more important than close friends. Further, others regarded as experts in the subject matter exert the strongest influence on peer choices.