Abstract To examine the atmospheric responses to Arctic sea-ice variability in the Northern Hemisphere cold season (October to following March), this study uses a coordinated set of large-ensemble experiments of nine atmospheric general circulation models (AGCMs) forced with observed daily-varying sea-ice, sea-surface temperature, and radiative forcings prescribed during the 1979-2014 period, together with a parallel set of experiments where Arctic sea ice is substituted by its climatology. The simulations of the former set reproduce the near-surface temperature trends in reanalysis data, with similar amplitude, and their multi-model ensemble mean (MMEM) shows decreasing sea-level pressure over much of the polar cap and Eurasia in boreal autumn. The MMEM difference between the two experiments allows isolating the effects of Arctic sea-ice loss, which explain a large portion of the Arctic warming trends in the lower troposphere and drives a small but statistically significant weakening of the wintertime Arctic Oscillation. The observed interannual co-variability between sea-ice extent in the Barents-Kara Seas and lagged atmospheric circulation is distinguished from the effects of confounding factors based on multiple regression, and quantitatively compared to the co-variability in MMEMs. The interannual sea-ice decline followed by a negative North Atlantic Oscillation-like anomaly found in observations is also seen in the MMEM differences, with consistent spatial structure but much smaller amplitude. This result suggests that the sea-ice impacts on trends and interannual atmospheric variability simulated by AGCMs could be underestimated, but caution is needed because internal atmospheric variability may have affected the observed relationship.
more »
« less
A Bracketing Relationship between Difference-in-Differences and Lagged-Dependent-Variable Adjustment
Difference-in-differences is a widely used evaluation strategy that draws causal inference from observational panel data. Its causal identification relies on the assumption of parallel trends, which is scale-dependent and may be questionable in some applications. A common alternative is a regression model that adjusts for the lagged dependent variable, which rests on the assumption of ignorability conditional on past outcomes. In the context of linear models, Angrist and Pischke (2009) show that the difference-in-differences and lagged-dependent-variable regression estimates have a bracketing relationship. Namely, for a true positive effect, if ignorability is correct, then mistakenly assuming parallel trends will overestimate the effect; in contrast, if the parallel trends assumption is correct, then mistakenly assuming ignorability will underestimate the effect. We show that the same bracketing relationship holds in general nonparametric (model-free) settings. We also extend the result to semiparametric estimation based on inverse probability weighting. We provide three examples to illustrate the theoretical results with replication files in Ding and Li (2019).
more »
« less
- Award ID(s):
- 1713152
- NSF-PAR ID:
- 10167774
- Date Published:
- Journal Name:
- Political Analysis
- Volume:
- 27
- Issue:
- 4
- ISSN:
- 1047-1987
- Page Range / eLocation ID:
- 605 to 615
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
ABSTRACT CONTEXT This paper examines an engineering dynamics course at Purdue University that was specifically designed to create an active, blended, and collaborative environment. In addition to in-person classes and support, students have access to blended content such as solution videos, mechanics visualizations, a course discussion forum, and interactive simulations. PURPOSE Many studies have shown that students’ engagement in an online discussion forum enhances their learning performance (Davies & Graff, 2005; Hrastinski, 2008). However, our previous research showed that students’ engagement in the online forum of our dynamics course differed significantly across students’ demographics. We showed that women, white, or Asian American students were more likely to be involved in online discussions than men, international, or Hispanic students (Duan et al., 2018). In this paper, we take the previous analysis further by examining whether the observed differences in online student engagement mediate or moderate student performance. APPROACH To answer our research question, we will first investigate the mediation effect by creating two models. A first model with race/international status as the mediating variable and gender identity as a control variable, and a second model with gender identity as the mediating variable and race/international status as a control. Second, we will investigate the moderation effect of demographic factors by creating a regression model including interaction terms to show the relationship of each demographic’s discussion forum engagement to overall performance. The goal of investigating these interaction terms is to determine if a moderating relationship exists where demographic factors impact online engagement, which in turn impact course performance. CONCLUSIONS We find that gender identity is the only significant demographic factor that moderates the effect of a student’s engagement on their performance. Based on the findings of our previous work, students of various racial and ethnic identities do engage differently in the discussion forum. However, this analysis was unable to detect any significant difference in student engagement based on demographics. Our paper contributes to understanding the mechanisms through which students’ engagement can translate into academic performance by focusing on their demographic background. The moderating role of students’ demographic background calls for a more targeted design of instructional tools in blended and collaborative environments to better support students from various demographic backgrounds.more » « less
-
The goal of causal mediation analysis, often described within the potential outcomes framework, is to decompose the effect of an exposure on an outcome of interest along different causal pathways. Using the assumption of sequential ignorability to attain non-parametric identification, Imai et al. (2010) proposed a flexible approach to measuring mediation effects, focusing on parametric and semiparametric normal/Bernoulli models for the outcome and mediator. Less attention has been paid to the case where the outcome and/or mediator model are mixed-scale, ordinal, or otherwise fall outside the normal/Bernoulli setting. We develop a simple, but flexible, parametric modeling framework to accommodate the common situation where the responses are mixed continuous and binary, and, apply it to a zero-one inflated beta model for the outcome and mediator. Applying our proposed methods to the publicly-available JOBS II dataset, we (i) argue for the need for non-normal models, (ii) show how to estimate both average and quantile mediation effects for boundary-censored data, and (iii) show how to conduct a meaningful sensitivity analysis by introducing unidentified, scientifically meaningful, sensitivity parameters.more » « less
-
more » « less
Abstract For small tropical islands with limited freshwater resources, understanding how island hydrology is influenced by regional climate is important, considering projected hydroclimate and sea level changes as well as growing populations dependent on limited groundwater resources. However, the relationship between climate variability and hydrologic variability for many tropical islands remains uncertain due to local hydroclimatic data scarcity. Here, we present a case study from Kiritimati, Republic of Kiribati (2°N, 157°W), utilizing the normalized difference vegetation index to investigate variability in island surface water area, an important link between climate variability and groundwater storage. Kiritimati surface water area varies seasonally, following wet and dry seasons, and interannually, due to hydroclimate variability associated with the El Niño/Southern Oscillation. The NIÑO3.4 sea surface temperature index, satellite‐derived precipitation, precipitation minus evaporation, and local sea level all had significant positive correlations with surface water area. Lagged correlations show sea level changes and precipitation influence surface water area up to 6 months later. Differences in the timing of surface water area changes and variable climate‐surface water area correlations in island subregions indicate that surface hydrology on Kiritimati is not uniform in response to climate variations. Rather, the magnitude of the ocean–atmosphere anomalies and island–ocean connectivity determine the extent to which sea level and precipitation control surface water area. The very strong 2015–2016 El Niño event led to the largest surface water area measured in the 18‐year data set. Surface water area decreased to pre‐event values in a similarly rapid manner (<6 months) after both the very strong 2015–2016 event and the 2009–2010 moderate El Niño event. Future changes in the frequency and amplitude of interannual hydroclimate variability as well as seasonal duration will thus alter surface water coverage on Kiritimati, with implications for freshwater resources, flooding, and drought.
-
more » « less
Conventionally, regression discontinuity analysis contrasts a univariate regression’s limits as its independent variable, R, approaches a cut point, c, from either side. Alternative methods target the average treatment effect in a small region around c, at the cost of an assumption that treatment assignment, [Formula: see text], is ignorable vis-à-vis potential outcomes. Instead, the method presented in this article assumes “residual ignorability,” ignorability of treatment assignment vis-à-vis detrended potential outcomes. Detrending is effected not with ordinary least squares but with MM estimation, following a distinct phase of sample decontamination. The method’s inferences acknowledge uncertainty in both of these adjustments, despite its applicability whether R is discrete or continuous; it is uniquely robust to leading validity threats facing regression discontinuity designs.
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1017/pan.2019.25
