Thus there is an association between the vulnerability variables and consulting a doctor. However, this behaviour can be seen as a confirmation or reflection of the vulnerability measures rather than as a contributory factor. We would not expect doctor visits to exacerbate the patients’ health and stress problems, although unsatisfactory consultations could do so.
6.3.2 Flood event characteristics Depth, number of rooms flooded and contamination of floodwaters
There were weak but significant correlations between maximum depth of flooding in main rooms (bathroom, bedroom, kitchen, and living room), the number of main rooms flooded, and the four vulnerability variables. However, what is perhaps surprising is that the relationship between depth and extent of flooding is not stronger. It is clear from the data that individuals flooded to the same depth and extent respond very differently to the experience (Table 6.7).
There were differences in the proportions with high scores on the current and worst time GHQ12 according to the maximum depth of main room flooding (Figure 6.15). However the main distinctions were between those with no or low (less than 10 centimetres) flooding and those flooded to greater depths. It is also surprising that a proportion of those with no main rooms flooded who of course will have had flood waters in other parts of their dwelling such as a hall or basement also recorded high GHQ12 scores.
Table 6.7: Correlations between depth and number of rooms and vulnerability variables:
Figure 6.15: Percentage with high current and worst time GHQ12 scores by maximum depth of main room flooding: Intangibles Survey
Regarding pollution in floodwaters, 77% of respondents said that the floodwaters contained sewage or other pollution. Respondents who believed the flood waters to be polluted were significantly more likely to have high (4+) GHQ12 scores, particularly GHQ12 worst scores (68% compared with 41% for those who did not consider the flood waters to be polluted). The proportions scoring four or more on the current GHQ12 were also significantly different for those who thought the flood waters contaminated and those who did not (27% compared with 12%). This would seem to indicate that the presence of sewage or other pollution was one of the characteristics of the flood that had a significant effect on the mental health of respondents not only at the time of the flood but also in the long term. Respondents who reported pollution in the flood waters were also more likely to score higher on all four vulnerability variables (Figures 6.16 and 6.17).
Thus, as predicted in the model, certain flood characteristics, chiefly the depth, number of rooms flooded and the presence of sewage and other pollution in the flood waters have an impact on the vulnerability of respondents and their households but the relationship between flood characteristics and vulnerability is perhaps not a strong as might be expected.
Figure 6.16: Mean scores on the stress of the flood event and overall severity of the flood by pollution of the floodwaters: Intangibles Survey
Stress: t test: p<0.001
Overall severity: t test ; p<0.001
Figure 6.17: Mean current and worst GHQ12 Likert scores by pollution of the floodwaters:
Worst GHQ12: t test; p<0.001
Current GHQ: t test : p<0.001
6.3.3 Flood warnings, rate of onset and duration of flood
Only 23% (229) of respondents received a warning before the flood event. There were no significant differences between people that had received a warning and those who had not in terms of GHQ12 (worst and current), overall severity, and subjective stress scores. People that had not received a warning were not more likely to score 4+ in the GHQ12 (current or worst). No correlations were found between any of the variables and the length of warning. The warnings are intended to provide property owners with the opportunity to protect their property and to reduce the risk to life, but they are also intended to reduce the stress of a flood by allowing people time to prepare mentally for the flood. However, this did not appear to be the case on the evidence of the bi-variate analysis.
Other flood event characteristics that were examined in bi-variate analyses did not appear to have an effect upon the vulnerability measures. The rate of rise of floodwaters had no effect on the subjective scores, the GHQ12 scores or the likelihood of scoring 4 or more. There were also no correlations between the effects of stress, GHQ12 scores and duration of the flood. The duration of the flood was not significantly longer for respondents with high GHQ12 scores. There was only a weak significant correlation between the duration of the flood and the overall severity of the flood (r=0.11; p<0.01).
6.3.4 Tenure and housing characteristics Tenure
Respondents that lived in rented accommodation showed greater vulnerability than those who owned or were buying their property on mortgage (Figures 6.18 and 6.19). Fewer respondents who owned their property or were buying on mortgage scored 4 or more in the GHQ12, (19% versus 27% of those who lived in rented homes, chi-square; p < 0.05).
Vulnerable properties are those situated on the ground floor or basement and where there are no upper floors to seek refuge from the floodwaters or to move furniture to. These include bungalows, ground floor and basement apartments and mobile homes. Those who lived in vulnerable properties had significantly higher vulnerability scores, except on the GHQ12 current (Figures 6.20 and 6.21).
Figure 6.18: Mean current and worst GHQ12 Likert scores by tenure:
*Current GHQ12: t test; p<0.05
** Worst GHQ2: t test; p < 0.01
Figure 6.19: Mean scores on the stress of the flood event and overall severity of the flood by tenure: Intangibles Survey
*** Stress: t test; p < 0.001
* Overall severity: t test; p < 0.05
Figure 6.20: Mean current and worst GHQ12 Likert scores by whether or not respondents were living in vulnerable property: Intangibles Survey
* GHQ Worst: t test: p < 0.05
GHQ Current: Not significant
Figure 6.21: Mean scores on the stress of the flood event and overall severity of the flood by whether or not respondents were living in vulnerable property: Intangibles Survey
** Stress: t test; p < 0.01
*** Overall severity: t test; p < 0.001
Area house prices and length of residence
The respondents were drawn from 30 locations across England and Wales. A house price rating (scale 1 to 5, 1 = high and 5 = low) was included as a surrogate variable to reflect locational characteristics (see Section 3.3.4). House prices were then compared to the national averages and assigned a simple rating value on a scale 1 to 5 (1 equates to areas where house prices are more than 1.4 times the national average, whilst a 5 equates to areas where house prices are up to 60% of the national average) (RPA, FHRC, 2004).
Area house price rating was weakly correlated with stress (Correlation 0.20, p <0.001, n = 972) and overall severity (Correlation 0.12, p<0.001, n = 973), i.e. the lower the area house price the higher the vulnerability of the household on these variables.
There was no correlation between the length of residence and any of the vulnerability variables. However, those who had high GHQ12 worst scores had on average lived in the area almost three years less than those who scored lower than 4 (or who did not respond to the GHQ12 questionnaire): 16 years versus 19, ( t test; p <0.01).
6.3.5 Other factors in the aftermath of flooding
Other factors are associated with the aftermath of the flood event and may also increase the severity of the effects of the flood or help the recovery. These factors include: having to leave the home, length of evacuation, length of disruption, problems with builders and insurers, worry about future flooding, and level of resources. Resources can be material or personal such as income, education, having insurance or social capital.
Problems with builders and insurers
Respondents were asked to rate the impact that problems with builders and insurers had on their household on a scale 1 = no effect to 1- very serious effect. There were some among the flooded who did not have to call upon builders or insurers after the flood. Thus, there was significant non response to these questions mainly because they did not apply to these respondents. A total of 88 respondents 9% did not reply on insurers and loss adjustors; for builders, the figures were 144 or 15%. These are excluded from the correlations shown in Table 6.8. Having problems with insurers and builders is not directly related to the flood event characteristics, it is something that happens after the event during the recovery period. There were significant correlations with the four vulnerability variables which indicate that these were factors that exacerbated the impact of the flood event itself and affected the resilience and vulnerability of the respondents.
Table 6.8 shows the Pearson correlations between the four dependent variables and the effect of problems with builders and insurers on the household:
Table 6.8: Correlations between vulnerability variables and problems with builders and insurers: Intangibles Survey
Effects of problems with insurers (scale 1-10)
Effect of problems with builders (scale 1-10)
Stress of the flood event
0.323, p < 0.001, n = 891
0.226, p < 0.001, n = 835
0.355, p < 0.001, n= 889
0.306, p < 0.001, n = 832
Current GHQ12 (scale 0-36)
0.238, p < 0.001, n = 733
0.175, p < 0.001, n = 696
Worst GHQ12 (scale 0-36)
0.337, p < 0.001, n = 731
0.309, p < 0.001, n = 693
Worry about future flooding
There were positive correlations between ‘worry’ about flooding in the next year and effects of stress, overall severity and GHQ12 scores. ‘Worry about future flooding’ generally was one of the ‘subjective effects’ that respondents were asked to rate on a scale from 1 (no effect) to 10 (extremely serious effect). There were significant correlations between this variable and the four vulnerability variables (Table 6.9). Worry about future flooding over the short and longer terms was quite strongly associated with vulnerability. However, worry is probably best seen as a component factor in the health and stress effects of flooding or indeed as a consequence of the health and stress effects experienced in a recent or worst flood event rather than as an explanatory factor for vulnerability.
Table 6.9: Correlations between vulnerability variables and worry: Intangibles Survey
“How worried are you about the possibility of your property being flooded during the next 12 months?”
Material and personal or household resources include having insurance against flooding, income levels, car ownership and education. The time need for recovery and time taken off work can also be considered as a resource. The use of these resources to recover from the flood prevents people from using them somewhere else.
Most households were insured against flooding. 85% of homes had building/structure insurance, 80% had ‘new for old’ contents insurance and a further 13% had other types of contents insurance. Respondents who had insurance were not less likely to obtain higher scores in the GHQ12 both current and worst time. There were also no significant differences in subjective stress, overall severity and mean GHQ12 scores between insured and uninsured households. Although it would be expected that having insurance against flooding would increase the resilience of the household, most respondents were insured so this may hide the effect on vulnerability. However, as shown above, the way in which insurers and loss adjustors handle insurance claims was a significant factor for the insured.
Respondents were asked whether their home or contents had been damaged in the flood. Not surprisingly since all the respondents had had flood waters in their homes, 96% of respondents said that their homes and/or contents suffered from flood damage. Respondents with insurance cover were asked to estimate the damages that were paid out by insurance, whether building and structure, contents or both. A significant proportion of those with damages and insurance were unable to give an estimate either because they did not know, could not remember or were unwilling to divulge the information. Respondents were able to answer in terms of costs to buildings and structure and contents damage separately, together or both. Thus the answers to these questions have been given by slightly different groups of respondents depending on whether they were able to give separate information on buildings and structure and content or not. The results shown in Table 6.10 are very rough estimates and must be treated with considerable caution because it is not possible to establish whether those who gave information on insured costs differed from those who did not in terms of the damages incurred.
Respondents were also asked to estimate the damages incurred that were not covered by insurance. Not all those who suffered damages of this kind were able to give an estimate and those who did so were only able to provide very approximate figures in many cases. Over a third (38%) suffered costs that were not covered by insurance. These respondents may have been uninsured, under-insured or may have been unable to convince their insurers of the value of their damaged property or costs, or may have lost patience with negotiating with them and decided to cover costs themselves. Most of those who incurred uninsured costs had insurance. As many as 80% of those with uninsured costs, compared with 88% for those not incurring such costs, had some buildings and structure insurance. For ‘new for old’ contents insurance the percentages were 67% as compared with 89%. However, those incurring uninsured costs were significantly less likely to have both types of insurance cover (chi-square; p<0.01 and p<0.001 respectively).
It should be noted that in addition, 88% of respondents also lost irreplaceable items such as photographs or personal papers to which it was impossible to attach a value as a consequence of the flood.
Table 6.10: Insured and uninsured damages incurred: Intangibles Survey
Non-response on damages of total the sample including those uninsured and those without damages
Buildings and structure
Non-response on uninsured damages of those who reported uninsured damages (375)
Buildings and structure
Tables 6.11 and 6.12 show the expected significant relationships between the extent of main room flooding and the maximum depth of main room flooding and the level of insured damages incurred. However, the small number of cases on which some of the means are based, and the large standard deviations (not shown), should be noted.
Table 6.11: Insured damaged incurred by the number of main rooms flooded: Intangibles Survey