Caveats for Causal Interpretations of Linear Regression Coefficients for Fine Particulate (PM2.5) Air Pollution Health Effects

Recent linear regression analyses have concluded that decreasing levels of fine particulate matter (PM2.5) air pollution have increased life expectancy in the United States. These findings have left unresolved questions about the causal relation between reductions in PM2.5 levels and changes in cause‐specific (especially, cardiovascular disease, CVD) mortality risks. Their robustness (e.g., sensitivity to deletion of a single data point) has also been questioned. We investigate these issues in the National Mortality and Morbidity Air Pollution Study database. Comparing changes in PM2.5 levels and cause‐specific mortality rates for elderly people in 24 cities between two periods separated by a decade (1987–1989 and 1999–2000) shows that reductions in PM2.5 were significantly associated with increases in respiratory mortality rates and with decreases in CVD mortality rates. CVD and all‐cause mortality risks fell equally for all months of the year over this period, but average PM2.5 levels increased significantly for winter months. This casts doubts on the causal interpretation that declines in PM2.5 over the decade caused reduced short‐term mortality risks. Nonlinear regression suggests that reduced or negative marginal health benefits are associated with reductions of PM2.5 below 1999–2000 levels (about 15 μg/m³). Such nonlinear relations imply that risk communication statements that project a constant incremental reduction in mortality risks per unit reduction in PM2.5 do not adequately reflect the realistic possibility of nonlinear exposure‐response relations and diminishing returns to further exposure reductions.     

Associations Between 1980 U.S. Mortality Rates and Alternative Measures of Airborne Particle Concentration

We analyzed the 1980 U.S. vital statistics and available ambient air pollution data bases for sulfates and fine, inhalable, and total suspended particles. Using multiple regression analyses, we conducted a cross-sectional analysis of the association between various particle measures and total mortality. Results from the various analyses indicated the importance of considering particle size, composition, and source information in modeling of particle pollution health effects. Of the independent mortality predictors considered, particle exposure measures related to the respirable and/or toxic fraction of the aerosols, such as fine particles and sulfates, were most consistently and significantly associated with the reported SMSA-specific total annual mortality rates. On the other hand, particle mass measures that included coarse particles (e.g., total suspended particles and inhalable particles) were often found to be nonsignificant predictors of total mortality. Furthermore, based on the application of fine particle source apportionment, particles from industrial sources (e.g., from iron/steel emissions) and from coal combustion were suggested to be more significant contributors to human mortality than soil-derived particles.     

Methods for Detecting and Estimating Population Threshold Concentrations for Air Pollution–Related Mortality with Exposure Measurement Error

The association between daily fluctuations in ambient particulate matter and daily variations in nonaccidental mortality have been extensively investigated. Although it is now widely recognized that such an association exists, the form of the concentration–response model is still in question. Linear, no threshold and linear threshold models have been most commonly examined. In this paper we considered methods to detect and estimate threshold concentrations using time series data of daily mortality rates and air pollution concentrations. Because exposure is measured with error, we also considered the influence of measurement error in distinguishing between these two completing model specifications. The methods were illustrated on a 15-year daily time series of nonaccidental mortality and particulate air pollution data in Toronto, Canada. Nonparametric smoothed representations of the association between mortality and air pollution were adequate to graphically distinguish between these two forms. Weighted nonlinear regression methods for relative risk models were adequate to give nearly unbiased estimates of threshold concentrations even under conditions of extreme exposure measurement error. The uncertainty in the threshold estimates increased with the degree of exposure error. Regression models incorporating threshold concentrations could be clearly distinguished from linear relative risk models in the presence of exposure measurement error. The assumption of a linear model given that a threshold model was the correct form usually resulted in overestimates in the number of averted premature deaths, except for low threshold concentrations and large measurement error.     

区域合作污染控制及动态收益分配策略研究

近年来，相邻区域之间的跨界污染问题层出不穷，并呈现交叉伤害的外溢性。在考虑污染容量动态变化的基础上，建立了非对称两区域之间的微分博弈模型，探讨在非合作机制、治污成本分担机制以及协同合作三种情形下双方的反馈纳什均衡策略、污染容量最优轨迹及瞬时收益的变化。在此基础上，进一步引入纳什议价方法分析了具有时间一致性的动态合作收益最优分配策略。结果表明，当污染对环境的伤害较大时，治污成本分担机制可以实现环境与经济的帕累托改善，而协同合作机制在区域收益与环境方面都是最优的；其次，通过数值分析探讨了地区的"短视"行为和环境自净率对稳定状态下均衡策略的影响，分析表明，地区的"短视"行为导致短期收益最大化，从而减少对治污的投入而增加了污染容量；而随着环境自净率的增加，污染容量与地区收益均有改善，但对环境伤害更为敏感的地区收益增加更显著。最后，给出了动态最优收益分配的具体数值变化。研究结论对区域之间治污模式与合作分配机制具有一定的决策参考价值。     

大气污染协同治理能促进污染物减排吗?--基于城市的三重差分研究

大气污染具有流动性,跨区域特征明显,协同治理至关重要。基于2010-2015年275个城市的面板数据,运用三重差分法(DDD)研究了大气污染协同治理对污染物减排的影响。从全样本分析结果来看,大气污染协同治理降低了工业二氧化硫排放量和工业烟(粉)尘排放量,而此结果并没有通过显著性检验。从子样本分析结果来看,大气污染协同治理显著降低了工业二氧化硫排放量,但对工业烟(粉)尘排放量没有产生显著影响。城市在促进不同污染物减排时存在差异,减排难度较低的污染物往往被优先完成。从时间效应来看,大气污染协同治理对污染物减排的影响存在时滞性,而且这种时滞性在不同污染物减排之间存在差异。     

Regression Versus Causation,Revisited

I recently discussed pitfalls in attempted causal inference based on reduced‐form regression models. I used as motivation a real‐world example from a paper by Dr. Sneeringer, which interpreted a reduced‐form regression analysis as implying the startling causal conclusion that “doubling of [livestock] production leads to a 7.4% increase in infant mortality.” This conclusion is based on: (A) fitting a reduced‐form regression model to aggregate (e.g., county‐level) data; and (B) (mis)interpreting a regression coefficient in this model as a causal coefficient, without performing any formal statistical tests for potential causation (such as conditional independence, Granger‐Sims, or path analysis tests). Dr. Sneeringer now adds comments that confirm and augment these deficiencies, while advocating methodological errors that, I believe, risk analysts should avoid if they want to reach logically sound, empirically valid, conclusions about cause and effect. She explains that, in addition to (A) and (B) above, she also performed other steps such as (C) manually selecting specific models and variables and (D) assuming (again, without testing) that hand‐picked surrogate variables are valid (e.g., that log‐transformed income is an adequate surrogate for poverty). In her view, these added steps imply that “critiques of A and B are not applicable” to her analysis and that therefore “a causal argument can be made” for “such a strong, robust correlation” as she believes her regression coefficient indicates. However, multiple wrongs do not create a right. Steps (C) and (D) exacerbate the problem of unjustified causal interpretation of regression coefficients, without rendering irrelevant the fact that (A) and (B) do not provide evidence of causality. This reply focuses on whether any statistical techniques can produce the silk purse of a valid causal inference from the sow's ear of a reduced‐form regression analysis of ecological data. We conclude that Dr. Sneeringer's analysis provides no valid indication that air pollution from livestock operations causes any increase in infant mortality rates. More generally, reduced‐form regression modeling of aggregate population data—no matter how it is augmented by fitting multiple models and hand‐selecting variables and transformations—is not adequate for valid causal inference about health effects caused by specific, but unmeasured, exposures.     

Estimating the Risks of Smoking, Air Pollution, and Passive Smoke on Acute Respiratory Conditions

Five years of the annual Health Interview Survey, conducted by the National Center for Health Statistics, are used to estimate the effects of air pollution, smoking, and environmental tobacco smoke on respiratory restrictions in activity for adults, and bed disability for children. After adjusting for several socioeconomic factors, the multiple regression estimates indicate that an independent and statistically significant association exists between these three forms of air pollution and respiratory morbidity. The comparative risks of these exposures are computed and the plausibility of the relative risks is examined by comparing the equivalent doses with actual measurements of exposure taken in the homes of smokers. The results indicate that: (1) smokers will have a 55-75% excess in days with respiratory conditions severe enough to cause reductions in normal activity; (2) a 1 microgram increase in fine particulate matter air pollution is associated with a 3% excess in acute respiratory disease; and (3) a pack-a-day smoker will increase respiratory restricted days for a nonsmoking spouse by 20% and increase the number of bed disability days for young children living in the household by 20%. The results also indicate that the estimates of the effects of secondhand smoking on children are improved when the mother's work status is known and incorporated into the exposure estimate.     

TEI@I方法论及其在外汇汇率预测中的应用

基于TE I@I方法论的理论框架,构建了一个基于TE I@I方法论的外汇汇率预测模型。在此模型中,传统的经济计量模型用于处理外汇汇率的主要趋势,人工神经网络技术用于分析外汇汇率的非线性,而文本挖掘和专家系统用于处理外汇市场中的突现性和不稳定性。最后,基于集成的思想,利用支持向量回归技术对上述3个部分进行非线性集成,从而获得一个更为精确的预测结果。通过实证方法验证了基于TE I@I方法论的外汇汇率预测模型的有效性。     

时间相依利率扩散模型的非参数估计

利率的变动模式会随着时间的推移、经济环境的变化和金融制度的改变而发生变化.时变的扩散模型能更好的描述短期利率的随机行为,文章采用基于核回归的非参数方法,估计中国银行间市场7天回购利率的时间相依CKLS模型.最后比较了时间相依CKLS模型与时齐CKLS模型在波动率预报上的表现,结果表明,时间相依CKLS模型更好的反映了利率实时的变化,提高了预测利率变化的精度.     

Gender Unemployment Dynamics: Evidence from Ten Advanced Economies

The paper investigates gender unemployment dynamics in 10 advances economies applying a recent methodology on widely available Labour Force Surveys data. We calculate the job finding and separation rates for each gender and use them to construct the steady‐state unemployment gap as well as two counterfactual gender unemployment gaps: one generated by differences only in job finding rates and the other by differences only in separation rates. We find that in all countries the gender unemployment gap attributed to differences in the job finding rate is lower than the gap attributed to differences in the separation rate, suggesting that gender differences in the separation rate are the major factor behind the gender unemployment gap.     

Change in Public Concern and Responsive Behaviors Toward Air Pollution Under the Dome

This study evaluates the effect of the documentary Under the Dome on the concern and responsive behaviors of the public regarding air pollution in China, with two surveys conducted before and after watching the documentary. Employing difference-in-differences regression, this study answers two research questions: (1) Does Under the Dome change public concern about air pollution? (2) Does Under the Dome change public behaviors in response to air pollution, including protective behaviors (i.e., wearing face masks) and mitigation behaviors (i.e., reducing car driving)? We find that the information campaign (1) protects against the decline of public concern about air pollution in Beijing and (2) moderates the degree to which people's perceived severity, perceived susceptibility, and sense of self-efficacy influence protective behaviors and moderates the degree to which people's belief in the cooperative behaviors by others influences mitigation behaviors. This study provides evidence that information campaigns of the Under the Dome type are effective in raising public awareness; however, the information campaign did not directly influence public protective and mitigation behaviors.     

Comparing Estimates of the Effects of Air Pollution on Human Mortality Obtained Using Different Regression Methodologies

Studies using regression techniques report their results using a variety of statistics. Evaluation of the consistency of findings, such as in a metaanalysis, requires calculating the statistical estimates of the effect reported in each study in a comparable manner. In this paper, we consider multiple linear regression, multiple Poisson regression, and logistic regression estimates. We present results that are needed to calculate, on a common basis, the slope of the regression function at a specified value, the elasticity function of the regression function at a specified value, the relative risk at a specified value, and the odds ratio at a specified value. We apply these results to studies of the association of daily mortality in an area to the daily air pollution level of ozone and PM₁₀. We calculate the estimated slope of the number of deaths per billion population associated with an increase of 1 ppb of ozone level in studies of daily mortality in three urban areas. These studies, in Los Angeles, New York, and St. Louis, produced very comparable results on a common basis, especially when compared to the coefficients as reported. We also calculated the estimated elasticity function of the daily mortality and daily PM₁₀ level for eight areas and found that the elasticities varied within a factor of roughly two, much less than the variability in the coefficients as reported.     

Acculturation,Ethnicity, and Air Pollution Perceptions

A globalizing world increases immigration between nations, raising the question of how acculturation (or its lack) of immigrants and their descendants to host societies affects risk perceptions. A survey of Paterson, New Jersey, residents tested acculturation's associations with attitudes to air pollution and its management, and knowledge of and self‐reported behaviors concerning air pollution. Linguistic and temporal proxy measures for acculturation were independent variables along with ethnicity, plus controls for gender, age, education, and income in multivariate analyses. About one‐fifth of contrasts between non‐Hispanic whites, non‐Hispanic blacks, English‐interviewed Hispanics, and Spanish‐interviewed Hispanics were statistically significant (Bonferroni‐corrected) and of medium or higher affect size, with most featuring the Spanish‐interviewed Hispanics. Knowledge variables featured the most significant differences. Specifically, Spanish‐interviewed Hispanics reported less concern, familiarity with pollution, recognition of high pollution, and vigorous outdoor activity, and greater belief that government overregulates pollution than English‐interviewed Hispanics (and than the other two groups on most of these variables too). English‐interviewed Hispanics did not differ from non‐Hispanic whites, but did on several variables from non‐Hispanic blacks. Temporal proxies of acculturation among the foreign‐born were far less significant, but concern and familiarity with air pollution increased with time spent in the United States, while belief in overregulation and a positive trend in New Jersey pollution increased with time in the nation of origin. Implications of these acculturation and ethnicity findings for risk perception/communication research and practice are discussed.     

Love of life and death distress: two separate factors

The objectives of the current investigation were threefold: a) to explore the gender differences on love of life (a new construct in the well-being domain) and death distress (death anxiety, death depression, and death obsession); b) to explore the relationship between the scales of these constructs; and c) to examine the factorial structure of these scales. The sample was 245 volunteer Kuwaiti college students (53.5% women). Their mean age was 21.9 (SD = 2.3). They responded to the Love of Life Scale, the Death Anxiety Scale, the Arabic Scale of Death Anxiety, the Death Depression Scale-Revised, and the Death Obsession Scale. Gender differences on love of life were not significant. However, women had significantly higher mean scores for the four death distress scales than did their male counterparts. All the correlations between love of life and the death distress scales were not significant except one pertaining to love of life and death depression (negative) in women. Two oblique factors were extracted: death distress and love of life. It was concluded that these constructs represent two distinct and independent factors. Counselors and clinicians dealing with death distress would find that it is not associated with love of life.     

Poorer Is Riskier

To examine the relationship between economic well being and health status, two economic concepts were explored: the permanent-income hypothesis and the transitory-income hypothesis. A regression analysis of time-series mortality data for the period of 1950-1988 was conducted. The regression results indicated that the total mortality rate is negatively associated with permanent income and positively associated with the transitory income. Results are also reported for the 8 major causes of death in the United States. The implications for risk analysis are discussed.     

Projection of Health Benefits from Ambient Ozone Reduction Related to the Use of Methyl tertiary Butyl Ether (MTBE) in the Reformulated Gasoline Program

To estimate potential public health benefits from ozone (O₃) pollution reduction attributable to the use of methyl tertiary-butyl ether (MTBE) in gasoline, O₃ dose-response estimates from the biomedical literature were combined with model estimates of O₃ reduction. Modeling employed EPA MOBILE5a and Complex models to predict emission changes, industry AQIRP techniques to predict ambient O₃ changes, and the National Exposure Model to predict human exposures. Human health effects considered were lung function decrements and respiratory irritant symptoms (using dose-response functions measured in laboratory and field studies), and increased death rates (using concentration-response functions inferred statistically from public-health data). Other reported health effects, such as lung inflammation, increases in asthma attacks, and hospitalizations, were not addressed because of inadequate dose-response information. Even for the health responses considered, quantitation of improvements due to MTBE use is problematical, because MTBE affects only a small percentage of existing O₃ pollution, and because exposure-response relationships are not well understood for population subgroups most likely to be affected. Nevertheless, it is reasonable to conclude that even small MTBE-associated reductions in peak ambient O₃ levels (1–5 ppb, according to model estimates) should yield considerable public health benefits. Tens of millions of Americans are potentially exposed to O₃ in the concentration range associated with health effects. Even if only a small percentage of them are susceptible, any incremental reduction in O₃ (as with MTBE use) must mitigate or prevent effects for a meaningful number of people. Better quantitative estimates of benefit must await a more detailed understanding of each link in the chain of causation.     

Gender and Race in Beliefs about Outdoor Air Pollution

Universal need for, or reactions to, risk communications should not be assumed; potential differences across demographic groups in environmental risk beliefs, attitudes, and behaviors could affect risk levels or opportunities for risk reduction. This article reports relevant findings from a survey experiment involving 1,100 potential jurors in Philadelphia concerning public responses to outdoor air pollution and air quality information. Flynn et al. (1994) and Finucane et al. (2000) found significant differences in risk ratings for multiple hazards, and in generic risk beliefs, between white men (or a subset) and all others (white women, nonwhite men, and nonwhite women). This study examined whether white men had significantly different responses to air pollution and air pollution information. An opportunity sample of volunteers from those awaiting potential jury duty in city courts (matching census estimates for white versus nonwhite proportions, but more female than the city's adult population and more likely to have children) filled out questionnaires distributed quasi-randomly. On most measures there were no statistically significant differences among white men (N = 192), white women (N = 269), nonwhite men (N = 165), and nonwhite women (N = 272). Nonwhites overall (particularly women) reported more concern about and sensitivity to air pollution than whites, and were more concerned by (even overly sensitive to) air pollution information provided as part of the experiment. Nonwhites also were more likely (within-gender comparisons) to report being active outdoors for at least four hours a day, a measure of potential exposure to air pollution, and to report intentions to reduce such outdoor activity after reading air pollution information. Differences between men and women were less frequent than between whites and nonwhites; the most distinctive group was nonwhite women, followed by white men. Flynn et al. (1994) and Finucane et al. (2000) found a far larger proportion of significant differences, with white men as most distinctive, probably due to use of different measures, study design, and population samples. However, all three studies broadly confirm the existence of gender and race interactions in risk beliefs and attitudes (particularly for white men and nonwhite women) that deserve more attention from researchers.