by
Christine M. Rodrigue, Ph.D.,
Eugenie Rovai, Ph.D.,
and
Susan E. Place, Ph.D.
presented to the:
Southern California Environment and History Conference
Nature's Workshop: Environmental Change in 20th Century Southern California
California State University, Northridge
19 September 1997
INTRODUCTION
In the early morning of January 17, 1994, a magnitude 6.7 earthquake shook the Los Angeles area. Several buildings and freeways collapsed, killing some 71 people and injuring thousands. Although loss of life was moderate, damage to property was enormous (current estimates are from $40 to $42 billion). The January 17 earthquake is now considered the costliest natural disaster in US history.
Disasters befall rural as well as urban areas. On April 25-26, 1992, three powerful earthquakes (magnitudes 7.1, 6.6, and 6.7) struck the rural Humboldt County area on the northwest coast of California. With their epicenters offshore, south of Petrolia, these earthquakes created a great deal of damage to the small towns of the area, including Ferndale, Fortuna, Petrolia, Rio Dell, and Scotia. Luckily, no-one was killed by the quakes themselves. Rovai has made a case study of this earthquake, and her findings will be used in this study of the Los Angeles earthquake by way of comparison.
It is generally recognized that such disasters have differential impact on various segments of the affected population, both because of the underlying physical processes and the social geography of the area (Mortenson 1994a,1994b; White and Haas 1975; Whittow 1979). This turned out to be true even in the more rural Humboldt County case, despite the narrower range of social variation seen there. Los Angeles, however, is an extremely culturally and economically diverse city, with strong spatial segregation of its various racial, ethnic, and national groups and socio-economic classes. Geographers would expect that quake damage, because of its discrete spatial patterns, would have differential impacts on the Los Angeles population. Our research finds substantial social differentiation in risk exposure and in access to emergency relief, recovery, and reconstruction across the social space of Los Angeles. This differentiation reflects the patterns of segregation in the city and suggests biases in media coverage.
SOCIAL GEOGRAPHY OF LOS ANGELES
Because Los Angeles is such a profoundly ghettoized city, residential segregation has led to extreme crowding in areas of recent immigration and to tremendous sprawl as "white flight" creates the "edge city" phenomenon described by Garreau (1991). This process accelerated during the 1980's as a result of de-industrialization of old established industrial areas (also, not surprisingly, traditionally the site of non-white residential zones) and a rapid influx of immigrants from Mexico, Central America, and Asia (Davis 1991). The demographics of the Los Angeles metropolitan area changed dramatically during the 1980's (Wisner 1994). Massive white flight and overcrowding in old racial and ethnic neighborhoods, in the context of a kaleidoscope of different political jurisdictions, translates into nightmarish problems when a disaster strikes (Wisner 1994). Emergency response and subsequent recovery are all compromised by the large number of different political jurisdictions that need to be coördinated. This is made doubly difficult by the budgetary problems caused by the declining tax base in the older ethnic communities, the resistance to paying taxes by the denizens of the affluent edge cities, and political fragmentation. The mean-spirited, anti-civic consciousness that first manifested with Proposition 13 in 1978 continues unabated, and indeed has become more deeply entrenched during the current economic crisis/restructuring (Soja 1989), as seen in the anti-immigrant sentiment in the state.
DEVELOPING A CONCEPTUAL FRAMEWORK: MEDIA COVERAGE, MENTAL MAPS, AND RESPONSE TO DISASTER
This paper reports on aspects of the research the Center for Hazards Research has conducted on the Los Angeles earthquake of January 17, 1994. This research traces the perceptual and behavioral linkages among the events of January 17, the representation of these events in local media, local residents' mental maps of earthquake damages, and the spatial and temporal allocation of emergency response, recovery, and reconstruction activities. In Manufacturing Consent, Edward S. Herman and Noam Chomsky (1988) provide a model for predicting the behavior of media, which we have applied to this case. They identify several filters operating to bias media selection of newsworthy items from the chaos of daily events. These filters include, first, the intense capital concentration in the media, which limits critical public debate on issues involving parent corporations and encourages sensational coverage likely to increase circulation and the bottom line (see also Bagdikian 1992 and Lee and Solomon 1991). Second, dependence on advertising revenue constrains serious and critical discussion of anything that would upset the income flow of the advertisers and also promotes the interests of the prosperous target markets of the advertisers (see also Steinem 1990). Third, media fear pressure from well-off and well-organized readers and politicians. The end result of such filters is a bias in media coverage favoring the interests and concerns of affluent people and de-emphasizing the interests and concerns of poorer people.
Given these filters, it can be expected that media reportage will tend to emphasize more upscale areas in a chaotic natural disaster situation. Northridge, the site of the three most sensational building failures and the horrific loss of sixteen lives in one of these, appropriately commanded a great deal of media attention. The naming of the Reseda-epicentered quake as the "Northridge" quake, however, may possibly reveal demographic bias on the part of media: Northridge is a much more upscale community than Reseda, which is a downwardly transitional blue-collar and modest white-collar community. For example, per capita income in Northridge is US$24,122s while in Reseda it is US$15,142. For the state of California, that figure is US$16,409 (US Census 1990). Another odd incident was the failure of the media to cover the hardest-hit Zip code in the city, 90016, in the Crenshaw District, which, though 20 miles from the epicenter, had the greatest concentration of destroyed and heavily damaged buildings in the City. This Zip code far surpassed any in the Northridge area.
That this disparity between the geographies of damage and of news coverage, of epicenters and media naming, has significant consequences and is more than a understandable one-incident aberration is suggested in Eugenie Rovai's (1993) work on the California North Coast earthquake of 1992. There, the epicentral community was Petrolia, but the national media dubbed it the Ferndale Quake, after a fashionable community famous for its trendy restored Victorians. There was virtually no coverage of Rio Dell, a downwardly transitional old Italian blue-collar community nearby, which experienced identical dollar damages. In a poor community, this translates into more extensive and dramatic infrastructural damage than in a more prosperous community, because of the depressed property values in the former. It is, therefore, particularly ironic that the devastation of Rio Dell was an invisible disaster outside the local area, due to biased media coverage, as in the case of the Crenshaw District of Los Angeles.
Similarly, Lee and Solomon (1991) and Smith (1992) have critiqued press coverage of the 1989 Loma Prieta quake, which emphasized the suffering of the affluent Marina district of San Francisco sixty miles from the epicenter and little covered the distress in the largely African-American Oakland equidistant from the epicenter. Conspicuously ignored was catastrophic damage in the largely Latino Watsonville and in the countercultural college town, Santa Cruz, both much closer to the epicenter. Some of this has to do with the spatial as well as social orientation of the media: the major regional papers are in San Francisco rather than in the smaller communities outside the Bay Area.
The significance of such disparities, should there prove a consistent upscale bias in reporting, is that emergency, recovery, and reconstruction activities are allocated to affected areas on the basis of disaster management personnel's mental maps of damage (Kiernan 1995). These mental maps may be strongly influenced by media coverage, with the possible result that better-off communities may secure thereby the lion's share of disaster relief. While everyone in coastal, desert, and Sierran California is at some risk to the earthquake hazard (Hornbeck 1983: 29), uneven performance of reconstruction can mitigate vulnerability for more affluent communities and exacerbate vulnerability for the more marginalized (Blaikie et al 1994: Ch. 8). Ferndale today is nearly completely recovered; Rio Dell has not recovered. With regard to the Loma Prieta earthquake, Charlene Shaffer, director of the Watsonville Chamber of Commerce and Agriculture, said: "It's distressing to see hard-hit communities near the quake's epicenter ... being overshadowed by the Bay area in the media's coverage ... Donations, which could make the difference between economic survival and devastation to a small community like ours, are pouring into San Francisco instead..." (cited in Lee and Solomon 1991).
Within a community, there are also substantial differences in vulnerability to disasters. The poorest people are the most vulnerable, especially those in the informal sector and the secondary labor market (Blaikie et al. 1994; Wisner 1993). In Los Angeles, recent immigrants are frequently relegated to the informal sector and, thus, marginalized. Their economic situations, already highly precarious, make them extremely fragile in a disaster, such as this earthquake. An anecdote from La Opinión illustrates the obstacles faced by such people in obtaining relief and rebuilding their lives following a disaster. A Central American woman, who had been working as a live-in domestic/nanny with a family in the San Fernando Valley in exchange for room and board and a small amount of money, found herself homeless and jobless following the January 17 earthquake. The family with whom she had lived left their damaged home to move in with relatives, where there was no room for the domestic. She was unable to obtain aid, because she had no proof of having lost home or employment as a result of the earthquake ("Cierre de albergues..." 1994).
In order more fully to address the concepts raised by structural hazards theory in the context of the Los Angeles and Humboldt County earthquakes, it is necessary to differentiate between risk and vulnerability, which is rarely been done in hazards literature (Rodrigue, 1993). Risk is actual and direct exposure to the destructive aspects of a natural event: losing one's home, assets, mementos, and, quite possibly, one's life, health, or loved ones in an earthquake. Vulnerability could usefully be defined as low capacity to evade, withstand, or recover from a disastrous event through personal resources or societal mechanisms of risk mitigation. The first is a statistical concept; the second is a social, economic, political, and, sometimes, a cultural one. All households in most of California indeed are at risk to earthquake, but they are not equally vulnerable due to differential access to attention and assistance.
At the household level, obviously, households with high incomes and net worths have personal resources (if not always sufficient motivation) to learn about earthquake hazard and mitigate their exposure through occupation of better-built homes and purchase of earthquake insurance. After an earthquake, their greater education levels and sophistication in dealing with bureaucracy and paperwork are likelier to yield more timely and effective assistance from a variety of governmental and private sources. In the L.A. area, with its large immigrant population, there may also be fear of governmental agencies on the part of both legal and illegal immigrants in light of the climate of hostility generated during recent political campaigns. Further exacerbating differences in vulnerability, local and national media tend to focus their coverage on the plight of the better-off, rather than of a more representative cross-section of victims in a disaster (Davis, 1995a; Place, 1995; Place and Rodrigue, 1994; Rodrigue, 1995, 1994; Rovai and Rodrigue, 1994; Rovai, 1995, 1993).
Evaluated here are the following hypotheses. First, wealthier communities garnered more media attention in this earthquake. Second, communities receiving excess media attention are further along in the process of recovery than poorer communities.
DATA AND METHODS
With these ideas from the structural approach to hazards and the propaganda model of the media, we used Los Angeles City Department of Building and Safety data, literature content analyses of the Los Angeles Times and La Opinión, and telephone surveys, to accomplish the following objectives:
To accomplish the first objective, we simply counted place references in earthquake-related front page articles in the Los Angeles Times and all such articles in the much smaller La Opinión. This yielded geographies of newspaper coverage. We then compared these geographies with the numbers of earthquake damaged buildings in each community. Simple linear regression generated a model of how the geography of damage "should" have been reflected in the media geographies. It turned out that the actual pattern of damages accounted for only 34 percent of the variation in the media coverage pattern. There is, thus, a significant signal from damage to coverage, but it is weak.
The communities with truly extreme residuals, the grossly overcovered and undercovered communities, were then subjected to further analysis of their socio-economic characteristics to determine if there was a bias in over- and under-coverage. Overcovered communities were 61 percent non-Hispanic white, while undercovered communities were only 22 percent non-Hispanic white. Furthermore, per capita income in the overcovered communities was $26,069, whereas in the undercovered communities, it was only $14,145. This pattern of skewing echoes that found in comparing the media-overcovered community of Ferndale with that of the media-ignored community of Rio Dell in Humboldt County.
To examine the third objective, the place name counts in the two papers were compared with one another and with the underlying pattern of damages, again using the simple linear regression method. There was no significant difference between the two papers' place name counts: the L.A. Times pattern accounts for 94 percent of the variation in La Opinión coverage. This is a bit odd, given that the two papers and their supporting advertisers target different market segments, who might be assumed to have somewhat different interests. Perhaps it isn't so odd, given the pattern of media concentration in the City -- the L.A. Times owns stock in La Opinión's parent corporation.
For the fourth objective, a random telephone survey of Los Angeles area residents was conducted in February of 1995, in order to create a composite map of Angelenos' mental maps of the disaster. The map of communities identified by Angelenos as the most hard-hit conformed almost perfectly with the maps generated from media coverage: media coverage accounted for fully 95 percent of the variation in residents' mental maps of the disaster, and the actual pattern of damages accounted only for 35 percent of the variation in the mental maps. It seems that people living in the middle of a disaster, even those right in the hardest-hit areas, rely on the media for their overall impressions of the disaster rather than on their own experiences, recognizing that they may only have visited a small fraction of the affected area.
The last objective was assessing the variations in disaster emergency response, recovery, and reconstruction activites among the socially, economically, and ethnically diverse communities of the greater Los Angeles area. One of the ways to do this is to classify activities as pertaining to response, recovery, or reconstruction and then construct timelines of these three phases. Data on these activities are obtained from media and from interviews of residents, businesses, and agencies. Rovai was easily able to construct such timelines for the Humboldt County area, but it proved impossible to do this in Los Angeles. One of the major reasons for this was the failure of media to cover poorer and minority communities.
Instead, Rodrigue purchased different editions of the L.A. Building and Safety database, which tracks the status of all inspected buildings in the City of Los Angeles. The more than 100,000 buildings in the database are categorized as red-tagged or condemned, yellow-tagged or limited entry (unsafe but repairable), and green-tagged or safe to occupy. Buildings are added to the database as inspection continues, subtracted when bulldozed, or moved from yellow to green tag status when repaired and reinspected. By comparing different editions of the database by Zip code, it became possible to construct a crude recovery rate for each community.
Very disturbingly, there proved to be an association between media attention, itself skewed by ethnicity and race and by income, and rates of recovery. Among the overcovered communities, there was a crude recovery rate of -42 percent of red-tagged buildings being bulldozed and yellow-tagged buildings being repaired, reinspected, and reoccupied. That is, of the 835 red-tagged and yellow-tagged buildings in these communities in April, 1994, there were only 485 were still red or yellow tagged by August, 1994. In the undercovered communities, there were many more damaged buildings: 3,066 in April. By August, there were still 2,030 of these still standing damaged, for a crude recovery rate of only -34 percent, which is significantly lower than the rate for the overcovered communities. Interestingly enough, this tendency to slower recovery in media-ignored areas affected two well-to-do white areas that also wound up undercovered: Woodland Hills and Chatsworth. That two prosperous white areas were undercovered and are lagging in recovery suggests the importance of media above and beyond wealth and race in the recovery from disaster.
CONCLUSION
This work has troubling ramifications. First, media narratives of disaster are not particularly reliable in their representation of damages in socio-economically and ethnically diverse areas, which echoes the expectations of many media critics. Second, the socio-economic structure, together with and reinforced by media, clearly affect the abilities of communities to recover from disaster: poorer communities recover more slowly than more prosperous (especially whiter) ones. Third, emergency and disaster management personnel must develop alternative means of assessing the hardest-hit areas in the critical first hours after a disaster, rather than rely on local papers, Associated Press, or other media sources. Given that poorer and minority areas are apt to be undercovered, emergency response and recovery personnel coming in from outside the region could obtain maps of Census data to identify potentially vulnerable areas and independently go to such areas to assess needs there. Fourth, our work also demonstrates the value of local knowledge in providing a counterweight to biased media images, and emergency response personnel might make a special effort to identify community leaders in minority and poorer communities to make use of their special local knowledges. Fifth, our results underscore the necessity for reporters and editors to confront their own, perhaps unconscious, biases and ameliorate their effects in their representations of disasters or any other news item.
A few avenues of further research are in order. It is not yet clear to us what the specific mechanism is that links socioeconomic status, ethnicity, and media biases to the disparate rates of recovery seen in this and the Humboldt County studies. Do the slow recovery patterns in more marginalized areas result from bias in media attention as it affects emergency response and recovery personnel? Do they instead reflect biases on the part of such personnel independently of the media? Do they have nothing to do with biases in media and response personnel but some other factor having to do with securing emergency assistance? Do they perhaps show that poorer and non-white households are less likely to qualify for assistance for reasons independent of any of the above effects?
REFERENCES
Table 1. Population and Housing Characteristics
___________________________________________________________________________
Income
Median household $25,875 $19,331
Per capita $13,504 $ 9,559
Educational Attainment
0 to 12th grade, no diploma 11% 37%
College or professional degree 39% 9%
Occupational Employment Structure
Managerial and professional specialty 32% 5%
Operators, fabricators, and laborers 10% 29%
Industrial Employment Structure
Agriculture, forestry, and fisheries 8% 3%
Manufacturing, durable goods 5% 33%
Professional and related services 27% 12%
Unemployment: Labor force not employed 3% 18%
Public Assistance
Population receiving public assistance 8% 25%
Aggregated public assistance $219,062 $1,700,000
Median Home Value $111,700 $67,100
___________________________________________________________________________
Source: U.S. Census, 1990
Table 2. Correlation and Simple Linear Regression of Los Angeles Times Place Name Mentions on Damaged Buildings (26 April 1994) by Community
___________________________________________________________________________
X Y
2
r r t b a
Total number of LA Times front 0.581 0.338 4.167 0.067 -5.212
damaged build- page article
ings (April 26) place mentions P=.00
in 1st month
___________________________________________________________________________
Table 3. The Ethnicity, Income, and Recovery Rates of Communities Overcovered and Undercovered by the Los Angeles Times
___________________________________________________________________________
Residuals of Los Angeles
Times place mentions on | # of | % non- | per | % rate |
April counts of damaged | commun- | Hisp. | capita | of |
buildings | ities | white | income | recovery |
| | | | |
Overcovered communities | | | | |
(residuals >= +5.000) | 8 | 61.2 | $26,069 | -41.9 |
| | | | |
Undercovered communities | | | | |
(residuals <= -5.000) | 9 | 21.7 | $14,145 | -33.8 |
| | | | |
| | | | |
| | Z=244.560 | no std | Z=4.326 |
| | | dev | |
| | P= 0.000 | data | P=0.000 |
___________________________________________________________________________