Political scientists often wish to classify documents based on their content to measure variables, such as the ideology of political speeches or whether documents describe a Militarized Interstate Dispute. Simple classifiers often serve well in these tasks. However, if words occurring early in a document alter the meaning of words occurring later in the document, using a more complicated model that can incorporate these time-dependent relationships can increase classification accuracy. Long short-term memory (LSTM) models are a type of neural network model designed to work with data that contains time dependencies. We investigate the conditions under which these models are useful for political science text classification tasks with applications to Chinese social media posts as well as US newspaper articles. We also provide guidance for the use of LSTM models.

No abstract

No abstract

No abstract

No abstract

What do we know about the causes and outcomes of international military conflict? Decades of research from different theoretical traditions have explored the outbreak and conclusion of international conflict from a variety of angles. Broadly speaking, scholarship about international conflict has tended to orbit around three core concepts: power, institutions, and the source of the interstate dispute. The question that remains is how well verified are the most important theories? Three influential theories seek to predict patterns of international conflict: power transition theory, which argues that shifts in power increase the likelihood of war; selectorate theory, which predicts that states that have large winning coalitions are more selective about war; and theories about issue indivisibility and war, which predict that issues that states view as impossible to divide—such as a national homeland—are more likely to lead to conflict. Each of these theories produces specific predictions, allowing an assessment of how well the evidence supports the theories’ main conjectures. Central to understanding the causes of conflict is whether empirical work has tested these three theories using well-validated measures; whether a variety of scholars have tested the core propositions of the theory; and whether scholars have found evidence of the causal mechanisms proposed by each theory. Although each theory has garnered some support, they all fall short on one or more of these criteria. In particular, more work is needed in both measurement and evidence of causal mechanisms before scholars can be confident of the theories’ explanatory power.

Researchers are employing confirmatory factor analysis (CFA) with multitrait-multimethod (MTMM) matrices to estimate parameters representing trait, method, and error variance, as well as parameters representing the correlations among traits (or factors). This study utilizes CFA with MTMM matrices to assess the convergent validity, discriminant validity, and the presence and effects of method variance in the end-user computing satisfaction instrument (EUCSI) and the computer self-efficacy instrument (CSE). The results of the study indicate that, in these samples, the two instruments demonstrate adequate convergent and discriminant validity, but that method variance is present and accounts for a large proportion of the variance in both models. Further, the proposed factor structure of the EUCSI appears to be unstable as a result of the effects of multiple methods, while the proposed factor structure of the CSE remains stable in the presence of the methods.

Social science and management information systems (MIS) research have been criticized for failure to integrate theory construction and theory testing (see e.g., Subramanian & Nilakanta, 1994). In particular, concerns with MIS as a cohesive research discipline have long included inadequate construct development and lack of valid, reliable measuring instruments for those constructs (Keen, 1980). Understanding the theoretical basis of constructs and how they are developed and tested across the research continuum are fundamentals of a cohesive academic discipline. To provide a common research framework for the growth of MIS as a scientific discipline, this chapter proposes a framework for an integrated research continuum across the life cycle of the research process.