2001/02 NSF SGER Project:

The Impact of The Sept. 11 Terrorist Attacks on Factors that Influence the Decision to Stay, Remodel or Move Among American Metropolitan Households

PI: Hazel A. Morrow-Jones

Co-PI: Elena G. Irwin

Collaborator: Brian Roe

Project Overview

In late summer 2001, the Principal Investigators (PI’s) conducted a mail survey of a random sample of homeowners in Franklin County, Ohio, the central county of the Columbus metropolitan area. The survey asked questions about housing/neighborhood satisfaction and desires, mobility and remodeling plans, and demographic characteristics of the household. After September 11th, SGER funds were requested to send out the same survey (with a few additional questions regarding attitudes about and expectations resulting from the terrorist attacks) to a new random sample of homeowners from the same geographic area in order to study the possible effects of the terrorist attack on people’s views of their homes, neighborhoods and communities in a geographic region that was not directly affected by the attacks on September 11th. The goal of this follow-up survey was to gather data that would permit an examination of the possible differences between the answers given immediately before the attack and those within a month or two after the attack.

The post-September 11th survey was completed and mailed in November 2001 to a random sample of 1,940 households in Franklin County, Ohio. The final version of the collated data became available for analysis in February 2002 with a total of 803 returned surveys (41% response rate). Analysis to date has included univariate comparisons between pre- and post- attack variables and multivariate analyses using discrete choice conjoint methods to analyze respondent's preferences for combinations of housing features. This SGER study looks at the immediate impact of the terrorist attacks on people's satisfaction and plans.

Research Findings To-Date

Our initial analysis of the data has focused on identifying differences in the responses to specific questions regarding home and neighborhood satisfaction and potential plans for moving. The major findings from this univariate analysis of the pre- and post-event survey responses are:

• We find evidence that social ties within the neighborhood are stronger among the post-September 11th group. Specifically, the post-event respondents indicated significantly higher levels of involvement with their neighbors in terms of talking with them and doing favors for them on a regular basis.
• We find somewhat lower desires to move for either house or neighborhood reasons in the post-attack group and find that respondents in the post-event group reports a significantly longer time period before they expect a move.
• Overall satisfaction levels with neighborhood density are not significantly different between the pre- and post-event groups. In separating out the groups based on whether the household resides in a rural, suburban, or urban neighborhood, we find that households living in rural and suburban neighborhoods have a significantly higher satisfaction with neighborhood density than do those living in an urban setting. We find no significant differences among these stated satisfaction levels in the pre- vs. post-event survey responses.
• In addition, although higher levels of satisfaction with both the home and other attributes of the neighborhood are present among those in the post-September 11th group, these differences are not significantly different across the pre- and post-event samples.

We also analyzed the responses of the post-September 11th group to selected questions regarding their perceptions of terrorist attacks. The major findings from this analysis are:

• About one quarter of the post-event respondents were concerned about a possible terrorist attack in Franklin County.
• In comparing the responses of those who expressed moderate or strong concerns over a terrorist attack in Franklin County with the responses of those who did not have this concern, we find that households that have stronger concerns regarding a possible attack are, on average, significantly less likely to consider moving to a downtown or near downtown residential location and are significantly much more likely to want to move to a rural area. In contrast, no significant differences arise across the two groups' stated beliefs regarding their likelihood of moving or the possible timing of a move at any point in the future.
• Households with children are significantly more likely to state that the terrorist attacks have made it more likely that they will move and specifically, they are significantly more likely to state that they are likely to move to a rural area.
• Women are significantly more likely than men to express concern over a possible future terrorist attack in Franklin County.
• Respondents over the age of 55 are significantly more likely to state that they are more likely to move to a rural area as a result of the terrorist attacks.

In order to further examine the impact of September 11th on future housing choices, we undertook a multivariate conjoint analysis using a subset of the survey questions in which individuals stated their choice over a hypothetical housing choice in which houses varied by specific combinations of housing attributes including commuting distance, neighborhood median income, neighborhood density, presence of agricultural land, presence of permanent agricultural land, presence of parks, school quality, and neighborhood safety. The results show that a number of expected factors were significant in influencing the housing choices of the pre-September 11 sample, including: less commuting, higher local income among neighboring residents, less density, more agriculture land (including permanent crop land), parks, better schools, and increased safety in the neighborhood all significantly and positively influenced the choice of a house. Surprisingly, the same analysis of the post-September 11th sample responses indicates that only one of these variables is statistically significant in explaining housing choices after the attacks. The only variable found to be statistically significant is neighborhood density, which is found to have a negative and very significant effect on housing choice. This finding suggests that a major shift in housing preferences occurred immediately after the September 11th events: concerns over neighborhood density seem to have swamped all of the other usual considerations in determining the desirability of a residential location.

Based on these preliminary results, we conclude that the most immediate impact of the terrorist attacks has been to influence the potential location of a household's future move. Specifically, we find evidence that households are more likely to avoid more urban and densely populated neighborhoods and are more likely to choose more rural, lower density neighborhoods. In other words, the attacks of September 11th appear to have greatly increased the importance of low density as a desirable neighborhood characteristic. In addition, we find evidence that suggests that the terrorist attacks have had a positive effect on social ties within neighborhoods and that respondents are somewhat happier with their current situations and are potentially more cautious with regard to moving to a new home and neighborhood in the post-September 11th climate. However, given that we only have data for the immediate post-event time period, the duration and permanency of these effects remains unknown.

Significance of Research

Our work to-date suggests several results that have important implications for research on urban and regional spatial structure in the fields of planning, economics, and geography. First, several findings point to significant differences in the spatial behavior of the pre- and post-event groups. The post-September 11th group expressed significantly stronger preferences for low-density, more rural areas as a residential location. Clearly if these impacts persist, then pressures for low-density suburban and exurban development will be even higher than they have been in the past, which has important implications for urban and regional spatial structure. Our evidence suggests that such threats, either perceived or real, may accelerate the decentralization process even further, thus augmenting the growth pressure in outer suburban, exurban, and perhaps even rural areas and further depopulating urban and inner suburban areas. This underscores the need for regional governance approaches to managing responses to terrorist threats and more generally, to managing the process of regional growth.

Second, the results indicate that the September 11th terrorist attacks have had a somewhat stronger effect than other forms of disaster considered in most of the literature on the behavioral effects of natural and technical hazards. In particular, the higher satisfaction with home and neighborhood in the wake of the attacks and lower desire to move among a population that is far removed from the actual location of the attack are effects that have not been found in previous studies.

Lastly, our results underscore the importance of using more sophisticated multivariate approaches to uncover subtle, but potentially important effects. Using simple univariate analysis, we failed to find a significant difference in the stated responses concerning satisfaction with neighborhood density levels. However, using multivariate conjoint analysis, we found a clear and strong preference for lower density neighborhoods in the post-attack sample.