Publication Type: Proceeding

Publication Type: Session Paper
Abstract: The Internet may have paved a new way to the research on public opinion and media effects. Netizens not only receive news information from the Internet but also use the same medium to share, exchange, and dispute each other’s political views. This interactivity allows the Internet to function as a “public forum”, where important issues are introduced and discussed, so that the opinion of the majority finally emerges. Important in this process of forming public opinion is the question of which issues are prominent on the Internet. As agenda-setting hypothesizes, prominent issues in the media may set the agenda not only for private discussions but also for policy debates among legislators.
This study raises a question, how is the Internet agenda set? Specifically, we look at the influence of television news on news selections of Internet portals in Korea. This process, commonly referred to as intermedia agenda-setting, explains the influence of prestige media on the news coverage of other news outlets.
Data for this study came from an extensive content analysis of two primetime news programs (KBS and MBC) and two major Internet portals (Daum and Naver) in Korea. Internet portals, which function as the entrance point to the sea of online information, play a major role as an online source of news information.
Looking at the correspondence between the prominent issues in television and those in Internet portals, this study looks at the extent to which television news may shape the agenda of Internet portals. Also examined is the question of whether the intermedia influences are reciprocal. Whereas television news may influence the Internet agenda, it is also likely that Internet portals may, to certain extent, influence the agenda of television news. That is, an issue becomes controversial on the Internet first; and then later offline media may pick up the issue and report it intensively in their primetime newscasts.
A series of cross-lagged correlation analyses revealed that television news might influence the process of issue selection in Internet Portals. Internet portals, however, did not show such a significant influence on the agenda of television news. The estimated influences from Internet portals to television news were all smaller than the baseline correlations, indicating that these influences were likely spurious and produced simply by the cross-sectional and time lagged associations. The implications of the findings are discussed in detail.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: This paper asks to what extent transnational norms and institutions are influencing rights claims and rights legislation at the national level in Latin America. One particularly dynamic – and fraught -- arena of rights at the transnational level is that of gender rights, including women's rights and sexual and reproductive rights. Thus, this paper examines four types of gender rights in the region: political quotas for women; ending domestic violence; reproductive rights, particularly abortion; and LGBT or queer rights. It concludes by speculating on why the effects of the transnational have differed across the rights types.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: The Asian Financial Crisis in the 1990s alerted policymakers and academics to the importance of creating early warning systems such as Credit Rating Agencies (CRAs) to assess bond default risk and prevent financial crises. However, do these agencies predict crises or only respond to them? Moreover, what role does politics play in the early warning systems? Controlling for the common explanations in the literature, we use panel data and interviews to investigate the role of CRAs and political determinants fifty developing countries from 1987 to 2003. We find that CRA are effective forecasters for bond defaults and currency crises. Although most political variables appear to have little effect on the likelihood for defaults or currency crises, the fact that CRAs use political factors in their methodology underscores the importance of politics. A unified model combining political and economic criteria is needed to predict financial crises and defaults. The message for investors in developing countries is that CRA ratings are good predictors of the future because their models are more holistic and include both economic and political determinants.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Recent work in time series analysis emphasizes the importance of omitted exogenous variables as a cause of autocorrelation. While autocorrelation by itself does not affect point estimates, then, the omitted variables driving it may introduce substantial bias in the estimates of coefficients on included exogenous variables. This suggests that analysts correcting for autocorrelation should consider not only the effects on estimated standard errors, but also the extent to which the correction technique mitigates omitted variables bias. In this paper, I show that in the case of an omitted contemporaneous explanatory variable, the coefficients estimated from Ordinary Least Squares (OLS) and Newey-West methods are biased. If the omitted variable is strongly autocorrelated, which is typically the case, a lagged dependent variable (LDV) can be a good proxy for the omitted variable -- even if the omitted variable is unknown and/or unmeasurable -- and the LDV therefore mitigates the specification bias. Concerns that addressing
autocorrelation by adding a LDV results in attenuated coefficient(s) on the variable(s) of interest, then, have the
problem reversed: the LDV estimate is less biased than the OLS
estimate.