Abstract
Objectives. We examined the extent to which the stress paradigm linking psychosocial stressors to mental health status has focused disproportionate attention on microlevel social stressors to the detriment of macrolevel stressors. Also, we assessed the effects of the terrorist attacks of September 11, 2001, on subsequent mental health among participants in a Midwestern cohort study.
Methods. Respondents in a 6-wave longitudinal mail survey completed questionnaires before September 11, 2001, and again in 2003 and 2005. Regression analyses focused on measures of negative terrorism-related beliefs and fears, as well as psychological distress and deleterious alcohol use outcomes measured both before and after September 11.
Results. Negative terrorism-related beliefs and fears assessed in 2003 predicted distress and drinking outcomes in 2005 after control for sociodemographic characteristics and pre–September 11 distress and drinking.
Conclusions. The events of September 11 continue to negatively affect the mental health of the American population. Our results support the utility of according greater attention to the effects of such macrolevel social stressors in population studies embracing the stress paradigm.
The stress paradigm guiding research on the effects of psychosocial stressors on mental health outcomes1–4 initially addressed exposure to stressful life experiences involving acute life events, such as the death of significant others, or chronic stressors, such as financial difficulties, as predictors of negative mental health outcomes. An important limitation of stress research has been its narrow focus on micro- or individual-level stressors to the detriment of broader macrolevel social stressors.
Stress researchers3,5,6 reviewing the stressors typically studied in large representative samples have noted the predominant focus on individual-level stressors (e.g., stress caused by life-changing events). Although some studies have also addressed more macrolevel stressors such as economic recessions7 and adverse living conditions,8 Wheaton5 noted that the macro–micro distinction highlights the fact that typical life event or role strain scales have excluded macrolevel traumas such as war stress, nuclear accidents, and natural disasters. For example, Holahan et al.,9 in their longitudinal study of stress, coping, and depression, addressed acute and chronic stressors involving experiences in 8 life domains: spouse, children, extended family, physical health, home, neighborhood, finances, and work. With the exception of neighborhood, all of these constitute microlevel domains.
In Wheaton’s review of the social stress literature, he further emphasized the continuing predominant focus on individual stressors and noted that “we also can and should consider political, military and social events and conditions that act as social stressors.”6(p288) A few researchers have included more macrolevel experiences. For example, the work of Turner et al. has addressed adversities such as “[being] in combat in a war, [living] near a war-zone, [being] present during a political uprising [and being] in a major fire, flood, earthquake or other natural disaster.”10(p232) However, this focus constitutes more the exception than what is typical in the overall literature.
In a different although related perspective on the evolution of the stress paradigm, Link and Phelan argued that the stress literature has gradually shifted from interest in “social conditions as fundamental causes of diseases”11(p85) to intervening mechanisms involving the ways in which individuals cope with these conditions. They concluded that “while the current approach focuses on the individual, it can readily be seen that economic and political forces shape individuals’ exposure to risks.”11(p85) This perspective suggests the importance of historical contexts and changes over time in social conditions, which play etiological roles in detrimental mental health outcomes (e.g., the extent to which social institutions such as the state may be viewed as unable to provide a sense of safety for their citizens).12
The events of September 11, 2001, signaled a major change in the sociopolitical context in the United States, highlighting the salience of political terrorism as a continuing threat to individuals’ sense of safety and well-being. A sizable literature has demonstrated that the September 11 attacks adversely affected the mental health of individuals across the nation13–16 as well as those most directly affected in New York, Washington, DC, and western Pennsylvania.17–19 These empirical studies, conducted in the immediate aftermath of the attacks, demonstrated elevated symptoms of depression, anxiety, and posttraumatic stress disorder (PTSD) and increased alcohol consumption.
Subsequent studies conducted between 2 months20 and 6 months21,22 after September 11 demonstrated lingering feelings of distress and increased use of alcohol and other substances, including cigarettes and marijuana, relative to the period before September 11. It should be noted that research addressing alcohol use outcomes has been much more limited than research addressing manifestations of psychological distress alone. However, one national study, conducted shortly after September 11, showed decreased rather than increased alcohol consumption.12
The extent to which the relatively immediate mental health effects of September 11 revealed in most studies have lingered is just beginning to be addressed. Boscarino et al.23 found that exposure to psychological trauma related to the World Trade Center attack in New York City was associated with increased alcohol consumption 2 years after the attack. Richman et al.24 demonstrated that a substantial percentage of a Midwestern population maintained negative beliefs and fears about their future safety linked to threats of future terrorist attacks in 2003 and that, after they controled for distress and drinking before September 11, these beliefs were significantly associated with distress and problematic drinking. However, a major limitation of that study was that terrorism-related beliefs and fears were measured at the same time point as distress and drinking outcomes, and thus the causal direction of the relationship between terrorism-related fears and mental health was ambiguous.
In this study, we further address the salience of post–September 11 beliefs and fears in terms of mental health outcomes. That is, we examined the extent to which these fears and beliefs, as assessed in 2003, predicted a range of distress and alcohol use outcomes in 2005 after controlling for previous distress and alcohol use. We also examined gender differences given evidence indicating that such post–September 11 effects are more pronounced among women than among men.13,17
METHODS
Sampling and Data Collection
Data were derived from a longitudinal mail survey of employees initially recruited from a Midwestern urban university during the fall of 1996. The sample was stratified into 8 groups according to gender and occupation. Initial wave-1 occupational groups included faculty, graduate student workers or trainees, clerical or secretarial workers, and service or maintenance workers. Employees (2416 men and 2416 women) were sampled from the university payroll database. Dillman’s25 total design method for mail surveys was used in collecting data, along with additional follow-up strategies (i.e supplementary reminder postcards, 2 additional mailings, reminder e-mail messages, and follow-up telephone calls).
The final wave-1 sample comprised 2492 employees (52% response rate). The lower than desired response rate reflected the fact that individuals may have been reluctant to complete questionnaires that were self-administered and contained identifiers for subsequent tracking and highly sensitive material.26 A comparison of the characteristics of the sample with the characteristics of the overall employee population indicated no significant differences in terms of race within each occupational stratum. Gender differences between our sample and the overall population were also very small and non-significant for 2 of the 4 strata (service workers and student trainees). However, men were overrepresented by 8.3 percentage points within the clerical group, and women were overrepresented by 11.3 percentage points in the faculty group.27
One year later (during the fall of 1997), 2038 wave-1 respondents were resurveyed (an 82% retention rate). Three years later (during the fall of 200l), the sample was again resurveyed, producing a sample of 1730 and a retention rate of 70%. Wave-4 data were collected similarly during the fall of 2002, producing a sample of 1654 (and a 67% retention rate). Wave-5 data were collected during the fall of 2003 (a sample of 1453 and a 59.1% retention rate), and finally, wave-6 data were collected during the fall of 2005 (a sample of 1517 and a 62.1% retention rate).
In comparison with dropouts, respondents who completed wave 6 were more likely to be older (mean age = 51 vs 47 years; P < .001), female (56% vs 50%; P < .01), and White (56% vs 45%; P < .001). With respect to mental health characteristics, there were no statistically significant differences between those who did and did not complete wave 6 in levels of depression or hostility and measures of alcohol consumption at wave 1 (quantity of consumption, escapist motives for alcohol use, binge drinking, and drinking to intoxication). However, wave-6 completers had lower anxiety scores than did dropouts (P < .05).
Measures
Terrorism-related stressors were measured with a 12-item scale (score range = 12–60) assessing terrorism-related negative beliefs and fears about the world, other people, and oneself (response options were not true/a little true, somewhat true, and very true/extremely true). (Because a t test showed no significant difference between men and women on this scale, we present only overall results.) Items were specifically linked to September 11 and fears of future terrorist attacks. The scale was a modified version of an instrument developed by Norris.28 Alpha coefficients were 0.82 for women and 0.83 for men.
Measures of mental health focused on symptomatic distress (depression, anxiety, hostility, somatization, and PTSD symptoms linked to September 11 and threats of future terrorist attacks) and alcohol consumption (quantity of consumption, escapist motives for drinking, binge drinking, and drinking to intoxication). Pre–September 11 baseline measures served as controls for all outcomes other than terrorism-related PTSD symptoms and somatization, which were added to the wave-5 questionnaire. For these 2 variables, wave-5 measures served as controls.
Past-week depressive symptomatology was measured with 7 items (score range = 0–21) from the Center for Epidemiological Studies Depression Scale29 that correlate highly with the overall scale.30 Alpha coefficients were 0.87 for women and 0.84 for men. Anxiety during the past week was measured with the 9-item tension–anxiety scale (score range = 0–36) of the Profile of Mood States.31 Alpha coefficients were 0.86 for women and 0.87 for men. The 6-item hostility scale (score range = 0–22) of the Symptom Checklist 90 Revised32 was used to assess hostility during the past week. Alpha coefficients were 0.78 for women and 0.76 for men.
Past-week somatization was measured via the 12-item somatization scale (score range = 0–48) of the Symptom Checklist 90 Revised.32 The alpha coefficient was 0.81 for both women and men. Terrorism-related PTSD symptoms were assessed with an adapted version of the PTSD Checklist–Terror,33 a 17-item instrument that encompasses the criteria for a PTSD diagnosis; scale items were summed to produce a score with a range of 17 to 85. We broadened the instructions for completing the instrument to capture symptoms associated with terrorist-related experiences and fear of terrorism after September 11, as well as fear of foreign wars or the threat of future wars. Alpha coefficients were 0.87 for women and 0.90 for men.
With respect to quantity of alcohol consumption, respondents were asked “When you drank any alcoholic beverage during the last 30 days, how many drinks did you usually have per day?” Escapist drinking motives were assessed via 5 items (to feel less tense, to escape, to cheer up, to forget things, and to forget worries) from the instrument developed by Temple.34 Alpha coefficients were 0.91 for women and 0.90 for men. Binge drinking was measured with a question used by Wilsnack et al.35:
“During the last 12 months, how often did you have 6 or more drinks of wine, beer, or liquor in a single day? (That would be a bottle or more of wine, more than 2 quarts of beer, or a half pint or more of liquor.)”
Drinking to intoxication was also measured with a question used by Wilsnack et al.35: “About how often in the last 12 months did you drink enough to feel drunk, that is, where drinking noticeably affected your thinking, talking, and behavior?”
Data Analyses
Because missing data in longitudinal research are common and potentially problematic,36,37 we used the SAS38 MI procedure to impute missing data and produce complete data sets (including observed and missing values). The missing data were replaced 10 times to produce 10 complete data sets, and the regression procedure (PROC REG) was used to analyze the complete data sets. As a means of generating valid statistical inferences about the parameters of interest, we used the SAS MIANALYZE procedure to combine the results. Results of ordinary least squares regression analyses with deleted cases were compared with results from the imputed data set. Because the analyses involving ordinary least squares with deleted cases produced results equivalent to those involving data imputation, we present the ordinary least squares results for ease of interpretation.
Also, given that many of the outcome variables included in this study were skewed, we used the negative binomial distribution to estimate regression models.39,40 These analyses were carried out through the SAS GENMOD procedure. The negative binomial regression model results were compared with the ordinary least squares regression model results, and the overall results were virtually the same. These analyses further supported our decision to present ordinary least squares results.
We conducted hierarchical multiple regression analyses examining the relationships between terrorism-related negative beliefs and fears measured at wave 5 and distress and alcohol-related outcomes measured at wave 6; these analyses controlled for education, race/ethnicity, age, gender, and baseline distress and drinking (except in the case of PTSD and somatization, for which wave-5 controls were used). Of central interest was the impact of negative beliefs and fears (alone or in interaction with gender) on subsequent distress and drinking after accounting for the variance explained by prior distress and drinking.
RESULTS
Tables 1 ▶ and 2 ▶ present sociodemographic characteristics of the wave-6 sample and mean scores and standard deviations for each of the relevant variables. Participants’ average age was 51 years, and the sample was 56% female, disproportionately White, and skewed toward higher educational attainment.
TABLE 1—
Sample Sociodemographic and Stress-Related Characteristics Among Midwestern US University Employees; Wave 6, 2005
| Wave-6 Respondents | |
| Total, No. | 1517 |
| Age, y, mean (SD) | 51 (11.5) |
| Gender, No. (%) | |
| Women | 847 (55.8) |
| Men | 670 (44.2) |
| Race/ethnicity, No. (%) | |
| African American | 340 (22.7) |
| Asian/Pacific Islander | 192 (12.8) |
| Hispanic | 95 (6.4) |
| White/Other | 870 (58.2) |
| Education, No. (%) | |
| Less than high school, high school, technical or trade school | 186 (12.3) |
| Some college or college | 301 (19.9) |
| Some graduate school or graduate school | 1027 (67.8) |
| Terrorism-related negative beliefs and fears scale score,a mean (SD) | 22.7 (6.8) |
aAs assessed at wave 5.
TABLE 2—
Sample Mental Health Characteristics Among Midwestern US University Employees: Wave 1, 1996, and Wave 6, 2005
| Wave 1, Mean Score (SD) | Wave 6, Mean Score (SD) | |
| Symptomatic distress | ||
| Depression | 3.7 (4.0) | 3.5 (3.9) |
| Anxiety | 7.4 (5.9) | 6.4 (5.3) |
| Hostility | 1.8 (2.8) | 1.5 (2.4) |
| Somatizationa | 4.2 (5.4) | 3.7 (4.8) |
| Posttraumatic stress disordera | 21.2 (6.3) | 20.4 (5.5) |
| Alcohol-related outcomes | ||
| Quantity of consumption | 1.3 (1.4) | 1.7 (1.1) |
| Escape motives for drinking | 6.9 (2.9) | 7.3 (3.7) |
| Binge drinking | 0.6 (1.2) | 0.5 (1.1) |
| Drinking to intoxication | 0.5 (1.0) | 0.6 (1.0) |
aBaseline taken at wave 5.
Table 3 ▶ shows the distribution of negative perceptions of the world and negative self-perceptions resulting from the September 11 attacks, along with fears related to future terrorist attacks. As opposed to items tapping negative interpersonal relationships or feelings of personal failure after September 11, results showed that terrorism’s effects primarily involved perceptions of the world as being a less safe place and as the government being less effective. For example, 30.0% of the participants reported that they felt very or extremely more pessimistic about world peace, and 27.6% reported that they had less faith in the government’s ability to protect them. By contrast, only 1.6% of the participants reported being very or extremely disappointed in themselves during the period following the events of September 11.
TABLE 3—
Ratings on Terrorism-Related Negative Beliefs and Fears Items Among a Sample of Midwestern US University Employees: 2003
| Item | Not at All True/A Little True, No. (%) | Somewhat True, No. (%) | Very True/Extremely True, No. (%) |
| Feel more pessimistic about world peace | 497 (41.3) | 347 (28.8) | 361 (30.0) |
| Feel less faith in government’s ability to protect | 544 (44.6) | 339 (27.8) | 336 (27.6) |
| Feel less safe than before September 11 | 601 (49.3) | 362 (29.7) | 257 (21.1) |
| Remain fearful of potential attacks in future | 631 (51.9) | 305 (25.1) | 280 (23.0) |
| Feel less able to control forces influencing life | 734 (60.3) | 304 (25.0) | 180 (14.8) |
| Feel more pessimistic about own future well-being | 893 (74.4) | 222 (18.5) | 85 (7.1) |
| Disappointed by others’ actions at time of crisis | 1052 (86.5) | 101 (8.3) | 63 (5.2) |
| Like someone less because of actions after September 11 | 1097 (89.8) | 60 (4.9) | 64 (5.2) |
| Disappointed by other person’s lack of help | 1108 (90.8) | 72 (5.9) | 41 (3.4) |
| Believe you should have given more support to others | 1124 (92.2) | 58 (4.8) | 37 (3.0) |
| Feel less confident in abilities to cope with major crises | 1140 (93.5) | 54 (4.4) | 25 (2.1) |
| Disappointed in yourself at time of crisis | 1155 (94.9) | 43 (3.5) | 19 (1.6) |
Three blocks of variables were entered into the hierarchical multiple regression analyses. The first block consisted of the controls: age, racial/ethnic group, educational level, distress (depression, anxiety, hostility), and alcohol use (drinking quantity, escapist motives for drinking, binge drinking, drinking to intoxication) measures at wave 1 and distress measures (somatization and PTSD) at wave 5. The second block included gender and terrorism-related negative beliefs and fears, and the third block included the interaction term of gender and terrorism-related negative beliefs and fears.
Table 4 ▶ presents the significant distress outcomes predicted by terrorism-related negative beliefs and fears and the extent to which, after controlling for sociodemographic variables and previous distress levels, these beliefs and fears predicted distress. Results showed that negative beliefs and fears were predictive of significantly increased levels of depression (P < .01), anxiety (P < .01), hostility (P < .001), and PTSD (P < .01) but not somatization. None of the interactions between gender and terrorism-related negative beliefs significantly predicted distress outcomes.
TABLE 4—
Hierarchical Linear Regression Models Predicting Distress and Drinking-Related Outcomes at Wave 6 From Terrorism-Related Negative Beliefs and Fears at Wave 5: Midwestern US University Employees
| Depression (n = 1050) | Anxiety (n = 1038) | Hostility (n = 1081) | Posttraumatic Stress Disorder (n = 1081) | Escapist Motives for Drinking (n = 1057) | Drinking to Intoxication (n =859) | |||||||
| b (SE) | δR2 | b (SE) | δR2 | b (SE) | δR2 | b (SE) | δR2 | b (SE) | δR2 | b (SE) | δR2 | |
| Step 1 | .144*** | .218*** | .129*** | .310*** | .195*** | .284*** | ||||||
| Age | −0.001 (0.010) | −0.042** (0.013) | −0.021*** (0.006) | 0.017 (0.012) | −0.018* (0.009) | −0.007** (0.002) | ||||||
| Race (1=White) | 0.312 (0.259) | 0.747* (0.326) | −0.007 (0.152) | 0.296 (0.312) | 0.321 (0.224) | 0.062 (0.066) | ||||||
| Education | −0.129* (0.052) | −0.042 (0.065) | 0.012 (0.030) | −0.011 (0.061) | −0.082 (0.044) | 0.007 (0.013) | ||||||
| Baseline outcomesa | 0.384*** (0.029) | 0.383*** (0.025 | 0.273*** (0.024) | 0.488*** (0.023) | 0.532*** (0.034) | 0.465*** (0.028) | ||||||
| Step 2 | .026*** | .016*** | .019*** | .004* | .008** | .002 | ||||||
| Gender (1=women) | 0.511* (0.225) | 0.613* (0.286) | −0.122 (0.134) | 0.067 (0.275) | 0.220 (0.199) | −0.033 (0.055) | ||||||
| Negative beliefs | 0.094*** (0.017) | 0.098*** (0.022) | 0.051*** (0.010) | 0.068** (0.023) | 0.050** (0.015) | 0.008 (0.004) | ||||||
| Step 3 | .000 | .000 | .001 | .000 | .000 | .003* | ||||||
| Negative beliefs by gender (1=women) | 0.044 (0.033) | 0.011 (0.042) | −0.023 (0.019) | −0.049 (0.040) | −0.021 (0.029) | −0.018* (0.008) | ||||||
| Total adjusted R2 | .170 | .234 | .149 | .314 | .203 | .289 | ||||||
Note. R2 values are adjusted.
aFor depression, anxiety, and hostility, baseline measures were taken at wave 1; for posttraumatic stress disorder, baseline measure were taken at wave 5; for drinking-related outcomes, baseline measures were taken at wave 1.
*P < .05; **P < .01; ***P < .001.
Table 4 ▶ also shows the effects of terrorism-related negative beliefs and fears at wave 5 on escapist motives for drinking at wave 6. Negative beliefs and fears were predictive of significantly increased escapist motives for drinking (P < .01). By contrast, negative beliefs and fears were not significantly predictive of quantity of alcohol consumption or binge drinking.
Finally, the statistically significant interaction between gender and terrorism-related negative beliefs and fears at wave 5 predicted drinking to intoxication at wave 6 (P < .05). As men’s terrorism-related negative beliefs increased, their propensity to drink to intoxication increased. Among women, increases in terrorism-related negative beliefs only minimally altered their propensity to drink to intoxication.
DISCUSSION
We argue that the stress paradigm has paid inadequate attention to macrolevel social stressors. Our results demonstrate the continuing significance—here, with respect to subsequent distress and problematic alcohol use—of the terrorist events of September 11, 2001, 4 years after these events led to political terrorism being defined by the government and other sectors of society as a salient macrolevel social stressor for the American population. We found that in 2003 our participants continued to fear future terrorist attacks and to believe that the government could not protect them. We also found that in 2005 these fears and beliefs had negative effects on symptomatic distress and escapist motives for drinking among both men and women as well as negative effects on drinking to intoxication among men (after we controled for previous distress and drinking). Thus, our data suggest that political terrorism is a macrolevel stressor of major public health significance.
Our findings should be interpreted within the context of its methodological strengths and limitations. The central strength entailed our ability to address the mental health consequences of exposure to the terrorist attacks of September 11 within the context of an ongoing longitudinal study that included pre–September 11 assessments of mental health status. This is in contrast to the majority of studies of the effects of exposure to the Sep-tember 11 attacks, which were initiated after September 11, 2001, and thus could not take into account individuals’ previous mental health status or could do so only within the context of the biases inherent in the use of retrospective measures.
Limitations
The study’s limitations included the less-than-ideal initial response rate and the differing rates of attrition typically associated with long-term longitudinal research. In particular, by the wave-6 time point, the sample was disproportionately White, middle aged, and highly educated. Thus, future research is needed to address similar issues in the context of a sample representative of the larger population in terms of age, social class, and racial/ethnic composition. However, the sociodemographic biases of our sample were somewhat offset by the fact that those who completed wave 6 did not differ from the original wave-1 sample with respect to most of the outcome variables assessed.
Moreover, additional analyses were conducted to examine whether sociodemographic variables (in addition to gender) interacted with negative terrorism-related beliefs to create differences in prediction of outcomes. Because these analyses produced nonsignificant results, they were not clearly suggestive of strong biases linked to sociodemographically based sample attrition. Nonetheless, future studies of more-diverse population groups may reveal stronger relationships between negative terrorism-related beliefs and mental health outcomes.
Another limitation is that the measure assessing negative terrorism-related beliefs and fears may have tapped other period effects, including the war in Iraq, economic conditions, and the general policies of the Bush administration, that have covaried with terrorism-related issues. Finally, it may be useful in future studies to disaggregate the overall negative terrorism-related beliefs measure as a means of addressing the relative salience of different subcomponents with respect to mental health assessment.
Another interesting issue for future investigation involves the relative salience of terrorism-related threats experienced by the Midwestern US population assessed here as opposed to the continuous and unpredictable exposure of those in other parts of the world to a variety of real dangers and hardships such as genocide, extended droughts, and massive political displacements. Although a major Midwestern landmark (the Sears Tower in Chicago) was an intended target of the September 11 perpetrators, this clearly did not represent the same level of personal danger or hardship as the types of macrolevel social stressors just described. Future research could also address the salience of terrorism-related fears and beliefs in conjunction with both protective factors (e.g., social support and religiosity) and vulnerability factors (e.g., genes that interact with deleterious environmental experiences in predicting outcomes associated with alcohol use and psychopathology).41,42
Conclusions
Our findings regarding the salience of gender with respect to the effects of terrorism-related beliefs and fears on distress and drinking outcomes are interesting in that they are discrepant with earlier results from this data set regarding the more immediate aftereffects of the September 11 terrorist attacks on distress and alcohol use. These earlier findings showed that women, but not men, manifested elevated levels of alcohol consumption and anxiety during the 6 months after September 11.13 By contrast, our results showed that 4 years after September 11, terrorism-related fears and beliefs (assessed in 2003) predicted distress similarly in women and men, whereas men but not women exhibited changes in alcohol consumption. Moreover, in contrast to women’s elevated quantity of consumption after September 11, men’s later reactions involved an increase in an indicator of more-clearly pathological levels of drinking, that of drinking to intoxication.
It has been speculated that changes in gender-linked mental health reactions to the September 11 terrorist attacks may reflect gender-differentiated styles of emotional feeling states in which women more quickly react to powerful threats to their security and men initially are more likely to deny or intellectualize fearful states.24 Our findings regarding the links between terrorism-related beliefs and fears and men’s increased levels of drinking to intoxication suggest that men’s reactions to politically threatening experiences may be more pathological over time. Future research is necessary to further explore the meaning of gender-differentiated reactions over time to terrorist events and threats.
Finally, our results indicate the significance of exposure to political terrorism as a macro-level social stressor. It might be useful in future research to develop an instrument that covers the broad array of potential macrolevel stressors and addresses the salience of different stressors for mental health outcomes in broad population surveys. Such an instrument might focus on exposure to disasters associated with natural causes (e.g., hurricanes, tornadoes, earthquakes, fires) as well as other causes (e.g., governmental inaction related to ensuring the safety of the food supply or lack of response to actual or possible infectious disease outbreaks). Moreover, in line with calls for epidemiological studies to include multilevel analyses,43 future research should encompass both macrolevel stressors across varied social contexts and individual-level characteristics as they relate to mental health and other outcomes.
Acknowledgments
This study was funded by the National Institute on Alcohol Abuse and Alcoholism (grant R01AA009989).
We thank the University of Illinois at Chicago Survey Research Laboratory for collection of the data and Sally Freels for her helpful statistical advice.
Human Participant Protection This research was approved by the institutional review board of the University of Illinois at Chicago. Respondents received a consent information document with each questionnaire, and informed consent was assumed if respondents returned completed questionnaires.
Peer Reviewed
Contributors J. A. Richman originated the study and assumed primary responsibility for writing the article. L. Cloninger conducted the data analyses and provided methodological expertise. K. M. Rospenda contributed to study conceptualization and interpretation of findings. All of the authors reviewed drafts of the article.
References
- 1.Dohrenwend BS, Dohrenwend BP. Overview and prospects for research on stressful life events. In: Dohrenwend BS, Dohrenwend BP, eds. Stressful Life Events: Their Nature and Events. New York, NY: John Wiley & Sons Inc; 1974:313–331.
- 2.Pearlin LI. The sociological study of stress. J Health Soc Behav. 1989;30:241–256. [PubMed] [Google Scholar]
- 3.Thoits P. Stress, coping and social support processes: where are we? What next? J Health Soc Behav. 1995;35(extra issue):53–79. [PubMed] [Google Scholar]
- 4.Turner RJ, Wheaton B, Lloyd DA. The epidemiology of social stress. Am Sociol Rev. 1995;60:104–125. [Google Scholar]
- 5.Wheaton B. Sampling the stress universe. In: Avison WR, Gotlib IH, eds. Stress and Mental Health: Contemporary Issues and Prospects for the Future. New York, NY: Plenum Press; 1994:77–114.
- 6.Wheaton B. The nature of stressors. In: Horwitz AV, Scheid TL, eds. A Handbook for the Study of Mental Health: Social Context, Theories, and Systems. New York, NY: Cambridge University Press; 1999:176–197.
- 7.Brenner MH. Mental Illness and the Economy. Cambridge, Mass: Harvard University Press; 1974.
- 8.Aneshensel CS, Sucoff C. The neighborhood context of adolescent mental health. J Health Soc Behav. 1996;37:293–310. [PubMed] [Google Scholar]
- 9.Holahan CJ, Moos RH, Holahan CK, Brennan PL, Schutte KK. Stress generation, avoidance coping, and depressive symptoms: a 10-year model. J Consult Clin Psychol. 2005;73:658–666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Turner HA, Turner RJ. Understanding variations in exposure to social stress. Health. 2005;9:209–240. [DOI] [PubMed] [Google Scholar]
- 11.Link BG, Phelan J. Social conditions as fundamental causes of disease. J Health Soc Behav. 1995;30:80–94. [PubMed] [Google Scholar]
- 12.Knudsen HK, Roman PM, Johnson JA, Ducharme LJ. A changed America? The effects of September 11th on depressive symptoms and alcohol consumption. J Health Soc Behav. 2005;46:260–273. [DOI] [PubMed] [Google Scholar]
- 13.Richman JA, Wisler JS, Flaherty JA, Fendrich M, Rospenda KM. Effects on alcohol use and anxiety of the September 11, 2001, attacks and chronic work stressors: a longitudinal cohort study. Am J Public Health. 2004;94:2010–2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Schlenger WE, Caddell JM, Ebert L, Jordan BK, Rourke KM. Psychological reactions to terrorist attacks: findings from the national study of Americans’ reactions to September 11th. JAMA. 2002;288:581–588. [DOI] [PubMed] [Google Scholar]
- 15.Schuster MA, Stein BD, Jaycox LH, et al. A national survey of stress reactions after the September 11, 2001, terrorist attacks. N Engl J Med. 2001;345:1507–1512. [DOI] [PubMed] [Google Scholar]
- 16.Silver RC, Holman EA, McIntosh DN, Poulin M, Gil-Rivas V. A nationwide longitudinal study of psychological responses to September 11. JAMA. 2002; 288:1235–1244. [DOI] [PubMed] [Google Scholar]
- 17.Cardenas J, Williams K, Wilson JP, Fanouraki G, Singh A. PTSD, major depressive symptoms, and substance abuse following September 11, 2001, in a mid-western university population. Int J Emerg Ment Health. 2003;5:15–28. [PubMed] [Google Scholar]
- 18.Centers for Disease Control and Prevention. Psychological and emotional effects of the September 11 attacks on the World Trade Center—Connecticut, New Jersey, and New York, 2001. JAMA. 2002;288:1467–1468. [PubMed] [Google Scholar]
- 19.Simeon D, Greenberg J, Knutelska M. Schmeidler J, Hollander E. Peritraumatic reactions associated with the World Trade Center disaster. Am J Psychiatry. 2003;160:1702–1704. [DOI] [PubMed] [Google Scholar]
- 20.Stein BD, Elliott MN, Jaycox LH, et al. A national longitudinal study of the psychological consequences of the September 11, 2001, terrorist attacks: reaction, impairment, and help-seeking. Psychiatry. 2004;67: 105–117. [DOI] [PubMed] [Google Scholar]
- 21.DeLisi LE, Maurizio A, Yost M, et al. A survey of New Yorkers after the September 11, 2001, terrorist attacks. Am J Psychiatry. 2003;160:780–783. [DOI] [PubMed] [Google Scholar]
- 22.Vlahov D, Galea S, Ahern J, Resnick H, Kilpatrick D. Sustained increased consumption of cigarettes, alcohol, and marijuana among Manhattan residents after September 11, 2001. Am J Public Health. 2004;94:253–254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Boscarino JA, Adams RE, Galea S. Alcohol use in New York after the terrorist attacks: a study of the effects of psychological trauma on drinking behavior. Addict Behav. 2006;31:606–621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Richman JA, Shannon C, Rospenda KM, Flaherty JA, Fendrich M. The relationship between terrorism and distress and drinking: two years after September 11, 2001. Subst Use Misuse. In press. [DOI] [PMC free article] [PubMed]
- 25.Dillman DA. Mail and Telephone Surveys: The Total Design Method. New York, NY: John Wiley & Sons Inc; 1978.
- 26.Sudman S, Bradburn N. Improving mailed questionnaire design. In: Lockhart DC, ed. Making Effective Use of Mailed Questionnaires. San Francisco, Calif: Jossey-Bass Publishers; 1984:33–47.
- 27.Richman JA, Rospenda KM, Nawyn SJ, et al. Sexual harassment and generalized workplace abuse among university employees: prevalence and mental health correlates. Am J Public Health. 1999;89:358–363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Norris FH. 9/11 instrument. Available at: http://obssr.od.nih.gov/Content/About_OBSSR/Activities/attack.htm. Accessed November 3, 2007.
- 29.Radloff LS. The CES-D Scale: a self-report depression scale for research in the general population. Appl Psychol Meas. 1977;1:385–401. [Google Scholar]
- 30.Mirowsky J, Ross CE. 1990. Control of defense? Depression and the sense of control over good and bad outcomes. J Health Soc Behav. 1990;31:71–86. [PubMed] [Google Scholar]
- 31.McNair DM, Lorr M, Droppleman L. Profile of Mood States. San Diego, Calif: Educational and Industrial Testing Service; 1981.
- 32.Derogatis LR. SCL-90-R: Administration, Scoring and Procedures Manual–II. Baltimore, Md: Clinical Psychometric Research; 1983.
- 33.PTSD Checklist–Terror. Available at: http://obssr.od.nih.gov/Documents/About_OBSSR/Activities/module3.pdf. Accessed November 3, 2007.
- 34.Temple M. Trends in college drinking in California: 1979–1984. J Stud Alcohol. 1986;47:274–282. [DOI] [PubMed] [Google Scholar]
- 35.Wilsnack SC, Klassen AD, Schur BE, Wilsnack RW. Predicting onset and chronicity of women’s problem drinking: a five-year longitudinal analysis. Am J Public Health. 1991;81:305–318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Collins LM. Analysis of longitudinal data: the integration of theoretical model, temporal design, and statistical model. Annu Rev Psychol. 2006;57:505–528. [DOI] [PubMed] [Google Scholar]
- 37.Allison PD. Missing Data. Thousand Oaks, Calif: Sage Publications; 2001.
- 38.SAS/STAT User’s Guide, Version 8. Cary, NC: SAS Institute Inc; 1999.
- 39.Pedan A. Analysis of count data using the SAS system. Available at: http://www2.sas.com/proceedings/sugi26/p247-26.pdf. Accessed July 7, 2006.
- 40.Long JS. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, Calif: Sage Publications; 1997.
- 41.Caspi A, Sugden K, Moffitt TE, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science. 2003;301:386–389. [DOI] [PubMed] [Google Scholar]
- 42.Kaufman J, Yang BZ, Douglas-Palumberi H, et al. Genetic and environmental predictors of early alcohol use. Biol Psychiatry. 2007;61:1228–1234. [DOI] [PubMed] [Google Scholar]
- 43.Schwartz S, Susser E, Susser M. A future for epidemiology? Annu Rev Public Health. 1999;20:15–33. [DOI] [PubMed] [Google Scholar]