PERSONAL RESPONSIBILITY AND VOLUNTEERING AFTER A NATURAL DISASTER: THE CASE OF HURRICANE KATRINA

Abstract

This study examines prosocial behavior and volunteerism in the context of Hurricane Katrina. Using interviews from East Baton Rouge Parish residents, results demonstrate that self-efficacy, education, religious attendance, and organizational membership exert significant, positive effects on feeling personally responsible for helping victims. Education, presence of children in the home, organizational membership, and the interaction between religious attendance and personal responsibility exert significant, positive effects on total hours spent volunteering at shelters for victims. These findings suggest that there is not a one-to-one correspondence between the characteristics of the typical volunteer and those who volunteer in a natural disaster.

I would like to thank Dr. Jeanne S. Hurlbert, Dr. Duane A. Gill, and the annonymous reviewers for their helpful comments.

Notes

¹We estimate the response rate as 65 percent of screened eligible respondents.

²Because the sampling unit was the household and the unit of analysis was the individual, interviewers selected randomly among adults by asking to speak with the person in the household who had the last (most recent) birthday.

³We asked respondents, “How much of an impact do you think you could have if you donated your time to helping the victims of Hurricane Katrina?,” coding responses as no impact at all (1), a small impact (2), a moderate impact (3), and a big impact (4).

⁴We asked respondents to report the highest year of education completed, coding responses as less than 1 year (0), high school graduate (12), first year of college or technical school (13), second year of college or technical school (14), third year of college or technical school (15), Bachelor's degree (16), one year of graduate or professional school (17), two years of graduate or professional school or Master's degree (18), three years of graduate or professional school or law degree (19), and four years of graduate or professional school, or Ph.D. or M.D. degree (20).

⁵We asked respondents to place themselves in the first appropriate category: under $5,000, under $10,000, under $15,000, under $25,000, under $35,000, under $50,000, under $75,000, under $100,000, over $100,000. I then recoded income to thousands of dollars, assigning the midpoint for each category. Forty-six respondents failed to report their family income. I used the following equation to impute income for those respondents: − 38.509 + (2.492 × education) + (− 6.175 × female) + (7.944 × work) + (− .06290 × age) + (11.915 × white) + (6.876 × health) + (1.987 × conservatism) + (15.103 × own home) + (19.516 × married).

⁶Respondents were asked to report the number of children under the age of 18 presently living in their home. This variable was recoded as whether (1) or not (0) respondents had children under the age of 18 living in their home.

⁷Respondents were asked to report the year in which they were born. To compute age, I subtracted that year from 2005.

⁸Previous research on the skewed distribution of volunteer hours recommends taking the natural log of hours volunteered, computed as the natural log of the hours volunteered plus .01, to avoid taking the log of zero (Thoits and Hewitt 2001).

∗p < .05;

∗∗p < .01; one-tailed tests.

∗p < .05;

∗∗p < .01; one-tailed tests.