Natural Hazards and Job Choice: Hazard exposure and concern

We examined the cumulative and individual hazard risks of participants’ current locations to assess their current level of hazard risk exposure. The hazard risk exposure for most participants (75%) was in the low to moderate range. Slightly more geoscientists than non-geoscientists lived in areas with high to very high hazard risk (22% vs 17%). However, a Chi-squared test yielded no statistically significant difference between the geoscientist, non-geoscientist and whole population cohorts.

Chi-squared test
chi2 = 1.564
P-value = 0.9916

By individual hazards, risk exposure was greatest for severe weather, slides, and floods among participants. Most participants lived in high to very high-risk areas (76%-77%) for severe weather, while only 20%-21% lived in moderate-risk areas. For slide hazards (i.e., landslides, mudslides, debris flows, avalanches, etc.), most participants lived in moderate-risk areas (63%), with slightly higher percentages for geoscientists (67%) than non-geoscientists (58%). However, the percentage of those living in high to very high-risk slide areas was the similar for all cohorts (17%-19%). Floods were also a common hazard, with most participants lived in moderate-risk areas (42%). The percentage of participants living in high to very high-risk areas was similar across cohorts (27%-28%), and interestingly, slightly more non-geoscientists lived in very low-risk areas (7%) than geoscientists (4%).

Hazard exposure to earthquakes, wildfires and volcanic hazards was lower among participants. Most participants lived in low-risk earthquake areas (43%), while high to very high-risk exposure was the same across cohorts (20%). Slightly more non-geoscientists lived in very low-risk earthquake areas (30%) than geoscientists (23%).

For wildfire hazards, half of participants lived in very low-risk areas (52%), with more non-geoscientists living in very low-risk areas (57%) than geoscientists (49%). Additionally, more geoscientists lived in high to very high-risk areas (19%) than non-geoscientists (13%).

Exposure to volcanic hazards was rare, with 94% of participants living in areas with no risk of volcanic hazards. However, more non-geoscientists (5%) lived in areas with high to very high-risk than geoscientists (3%). Chi-squared tests yielded no statistically significant differences between the geoscientist, non-geoscientist and whole population cohorts for each individual hazard risk exposure.

Chi-squared test

Prior Experience with Hazards

We examined the severity of hazard impacts experienced by participants, both as the cumulative sum of impacts as well as impacts from individual hazards. In examining the overall experience of hazard impacts, most participants reported experiencing low levels of impact from hazards (44% - 45%), while one-third reported medium levels of impact. Slightly more geoscientists (7%) than non-geoscientists (5%) reported high severity impacts from hazards. A Chi-squared test yielded no statistically significant difference between the geoscientist, non-geoscientist and whole population cohorts.

Chi-squared test
chi2 = 0.2773
P-value = 0.9996

Most participants reported experiencing no impacts from volcanic hazards (95%-98%), slides (78%-82%), (i.e., landslides, mudslides, debris flows, avalanches, etc.), earthquakes (61%-66%), and wildfires (60%-65%), while just under half of participants reported the same for floods (47%-49%), and just under a third (29%-31%) reported the same for severe weather impacts. Slightly more geoscientists than non-geoscientists reported impacts from wildfires, floods, slides, and volcanoes than non-geoscientists, and slightly more non-geoscientists reported impacts from earthquakes and severe weather than geoscientists. However, the percentages between cohorts were similar. Chi-squared tests yielded no statistically significant differences between the geoscientist, non-geoscientist and whole population cohorts for impacts from each individual hazard.

Chi-squared test

Concern with Hazards at Current Location

We examined participants reported concern for hazards at their current location of residence, both as a cumulative concern for all hazards as well as concern for individual hazards. Most participants were slightly concerned about hazards where they currently lived (61% geoscientists, 52% non-geoscientists), with just over a quarter reporting that they were somewhat concerned (28% geoscientists, 26% non-geoscientists. Slightly more non-geoscientists (5%) reported being “Very” or “Extremely” concerned about hazards where they currently lived than geoscientists (3%). A Chi-squared test yielded no statistically significant difference between the geoscientist, non-geoscientist and whole population cohorts.

Chi-squared test
chi2 = 4.6082
P-value = 0.7985

Most participants reported no concern regarding volcanic hazards (86%-91%%), slides (68%-71%), (i.e., landslides, mudslides, debris flows, avalanches, etc.), earthquakes (62%-65%), and wildfires (52%-55%), while about half of participants reported being “Slightly” or “Somewhat” concerned about floods (45%-52%) and severe weather (47%-50%). Severe weather drew the most concern from participants, with 24% of non-geoscientists and 30% of geoscientists reporting being “Very” or “Extremely” concerned about these hazards. Chi-squared tests yielded no statistically significant differences between the geoscientist, non-geoscientist and whole population cohorts for concern about each individual hazard.

Chi-squared test

Factors Influencing the Choice of Participant’s Current Residence and Current Occupation

We examined the factors influencing the choice of participant’s current residence and occupation to better understand what was most important for these choices, and where hazard risk ranked among other factors. Hazard risk and crime risk were ranked not important by most participants, with more geoscientists (58% and 52% respectively) than non- geoscientists (48% and 40% respectively) indicating that these did not factor into their choice of current location.

Factors that were “Very” or “Extremely” important to participants in their choice were location (39%-46%), distance from social networks (33%-36%), community resources (30%-33%), income (27%-34%), climate & weather (24%), crime risk (13%-24%), other reasons not listed (16%-18%, and hazard risk (8%-14%). Of these very to extremely important factors, geoscientists had the lowest percentages for hazard risk, crime risk, income, and distance from social networks. Factors ranked as “Slightly” or “Somewhat” important included climate (43%-46%), community resources (36%-43%), income (38%), hazard risk (34%-38%), crime risk (35%-36%), location (28%-37%), distance from social networks (29%-31%), and other reasons not listed (12%-16%). For these slightly to somewhat important factors, geoscientists had the lowest percentage for all reasons.

Chi-squared tests yielded no statistically significant differences between the geoscientist, non-geoscientist and whole population cohorts for the importance of factors influencing the choice of participant’s current location.

Chi-squared test

Among factors influencing the choice of participant’s current occupation, job characteristics were “Very” to “Extremely” important by most participants (77% geoscientists, 64% non-geoscientists), followed by location (42% geoscientists, 40% non-geoscientists), and then income (35% geoscientists, 41% non-geoscientists). Just over a third of participants indicated that location and income were “Slightly” to “Somewhat” important (36% geoscientists, 41% non-geoscientists, and 37% geoscientists, 38% non-geoscientists, respectively). Other reasons influencing the choice of current occupation were “Not important” to two-thirds of participants. Chi-squared tests yielded no statistically significant differences between the geoscientist, non-geoscientist and whole population cohorts for the importance of factors influencing the choice of participant’s current occupation.

Chi-squared test

Correlation between Hazard Concerns, Experience with Hazards, and Current Hazard Risk

We calculated the Spearman correlation between concern over hazards at participant’s current location and the severity of impacts they experienced from hazards, hazard concern and the hazard risk at their current location, and severity of impacts experienced and the current hazard risk at their current location.

In examining the correlation between concern and severity of impacts experienced, we wanted to understand if individuals who experienced greater impacts were more concerned that those who experienced lesser impacts from hazards. Correlations were moderate to strong across all hazards and groups, with non-geoscientists having higher correlations for most hazards than geoscientists. The highest alignment between concern and hazard experience related to severe weather, floods, and wildfires.

We next wanted to understand if individuals who were more concerned about hazards in their current location lived in areas with high hazard risk. Correlations between concern and hazard risk at an individual’s current location were strong across all cohorts for earthquakes and volcanoes. Notable differences were seen between geoscientists and non-geoscientists for severe weather indicating higher concern and more awareness of severe weather hazards among geoscientists. In addition, there was a weak correlation for all cohorts for floods and slide hazards, suggesting that individuals may underestimate or be unaware of the risks of these hazards where they live.

We also investigated if individuals who experienced high severity impacts from hazards lived in high hazard risk areas. Correlations between impact severity and hazard risk were weak across cohorts and hazard types, suggesting that past impacts from hazards may not have been a factor in choosing the current location of residence.

We conducted Fisher z-tests to see if there was any statistically significant difference between cohorts for these three sets of correlations (i.e., concern and impact severity, concern and hazard risk, impact severity and hazard risk). The only statistically significant difference found was between geoscientists and non-geoscientists for hazard concern and impact severity (p = 0.0101) related to volcanoes, suggesting that geoscientists may be more likely to have concern tied to personal experiences of hazard impacts. Marginal statistical significance existed between geoscientists and non-geoscientists over concern and impact severity related to flood (p = 0.0657) which suggests that geoscientists may have a stronger connection between concern and experience with flood hazards than non-geoscientists. Across all hazards and all group comparisons, no other p-values fell below conventional significance thresholds, indicating that concerns and experience of hazard impacts were broadly similar across groups.

We conducted Fisher z-tests to see if there was any statistically significant difference between cohorts for correlations between the importance of hazard risk in choosing their current location of residence, the concern for hazards, prior hazard impacts, and the current level of hazard risk at their current location of residence. Overall, participants who were concerned about hazards were slightly more likely to report that they considered hazard risk when choosing where to live — but the relationship was modest and similar across all groups (ρ ≈ 0.32–0.34), with no statistical difference. Non-geoscientists showed a stronger correlation between hazard impact severity and the importance of hazard risk in their location choice than geoscientists (ρ = 0.433 non-geoscientists vs 0.252 geoscientists), and this difference was statistically significant (p = 0.0493). Correlations were weak and not significant between the importance of hazard risk in choosing the current location of residence and the actual level of hazard risk for that location, suggesting that people who live in high-risk areas did not report hazard risk as an important factor in their location choice.

Willingness to Pay for Protection from Hazards

We asked a subset of participants how much they would be willing to spend to avoid harm or loss from each type of hazard, and received 359 valid responses (205 geoscientists and 154 non-geoscientists). While there was no significant differences between geoscientist and non-geoscientist cohorts, a higher percentage of geoscientists across all hazard types were willing to spend more for protection from hazards. However, across all cohorts, the general upper limit of spending for hazards was at the $10,000 level, with more individuals being willing to spend more on protection from severe weather and floods than on other hazards.

The majority of respondents indicated they would be willing to pay less than $1,000 for protection from volcanoes, slides, and earthquakes. Half of respondents noted the same for wildfires. Floods and severe weather saw the largest percentages of investment by respondents. For floods, 41% of respondents noted they would be willing to spend less than $1,000 and 35% indicated they would be willing to spend between $1,000 and $10,000. Protection from severe weather garnered higher spending amounts with 33% of respondents willing to spend less than $1,000 and 39% willing to spend between $1,000 and $10,000.

Chi-squared test

For each type of hazard, we also examined if willingness to pay for protection from the hazard was influenced by having a geoscience degree, the current cost of living, prior experience with hazard impacts, and the concern for that hazard at one’s current location. Across all hazards, perceived risk (concern) and prior experience were consistent drivers for an individual’s willingness to pay for protection.

In particular, for flood hazards, the actual level of hazard risk (current_flood) had a significant negative effect on willingness to pay, indicating that those living in flood-prone areas reported lower willingness to pay. In addition, cost of living had a marginal positive effect but weak effect on willingness to pay, suggesting that people in higher-cost areas may be more inclined to invest in flood prevention. These results may be indicative of effects of U.S. flood insurance programs, whereby living in flood-prone area decreases willingness to pay due to subsidized insurance costs, or there may be an lower concern due to coverage from these programs and/or federal disaster aid.

Volcanoes are another hazards to note as their impacts and actual risk levels among participants are relatively low and rare. While concern for volcanic hazards is a strong driver as it is for all other hazards, for this specific hazard, geoscientists tend to be more willing to pay for protection from the impacts of this hazard. The marginal effect of having a geoscience degree for volcanoes may reflect a disciplinary awareness signal: geoscientists may recognize the long-tail risks and mitigation gaps associated with volcanoes better than non-specialists, leading to higher stated willingness to pay.

Ordered Logistic Regression Summary: Earthquake Prevention Cost

Variance Inflation Factors

Ordered Logistic Regression Summary: Wildfire Prevention Cost

Variance Inflation Factors

Ordered Logistic Regression Summary: Flood Prevention Cost

Variance Inflation Factors

Ordered Logistic Regression Summary: Severe Weather Prevention Cost

Variance Inflation Factors

Ordered Logistic Regression Summary: Slide Prevention Cost

Variance Inflation Factors

Ordered Logistic Regression Summary: Volcano Prevention Cost

Variance Inflation Factors

We plotted the mean willingness to pay against prior experience with hazards (impact severity) and perceived risk (concern for hazards) to visualize the combinations of experience and perceived risk that resulted in higher and lower willingness to pay for protection from a specific hazard. The charts illustrate how concern is a driving factor for willingness to pay, and how for some hazards, prior experience amplifies willingness to pay.

Location, Hazard Risk, and Cost of Living

As an aside, we also explored how cumulative hazard risk and cost of living vary across different location types (inland, coastal watershed, and coastal shoreline areas) as reported by survey participants. We mapped the county FIPS codes for each participant to the NOAA Coastal Counties list. Coastal watershed counties are US counties on or close to the U.S. coastline, while coastal shoreline counties are directly adjacent to the open ocean, major estuaries, and the Great Lakes. Coastal shoreline counties generally experience a higher prevalence and severity of impacts from coastal hazards than coastal watershed or inland counties. Each point on the chart below represents a respondent’s current location, with cost of living as a proxy for minimum income needed, cumulative hazard risk score, and the type of location.

Coastal shoreline areas are consistently high in both hazard risk and cost of living. This supports the notion that shoreline areas — such as large coastal cities — often involve trade-offs between economic opportunity and environmental exposure. Inland areas show lower cost of living and diverse hazard levels. Inland respondents tend to reside in lower-cost locations, though hazard risk varies considerably from low to moderately high. This range suggests inland risks (e.g., tornadoes, wildfires) are present but may be less tied to economic concentration. Coastal areas fall between shoreline and inland types in both cost of living and hazard risk. These may include suburban or rural coastal counties with moderate exposure and cost — suggesting a middle-ground trade-off.

Individuals living in shoreline areas are more likely to face both higher hazard exposure and higher living costs, possibly reflecting economic or lifestyle incentives that outweigh risk concerns. In contrast, inland areas tend to be more affordable, though hazard risk is not absent — reinforcing that hazard exposure is not limited to the coast.

Continue reading

• About the game
• Methodology
• About the participants
• Hazard exposure and concern
• Factors influencing job choice and relocation
• Reasons for current location, occupation, and final job choice
• Alignment of attitudes and actions
• Predicting job choice outcomes
• Conclusions