Societal Risk Perception in Present Day Russia

ABSTRACT

The study analyzed new information regarding the way in which present day Russians (living in Moscow or Tula) perceive societal risks. The main concerns of Russians in year 2003 were essentially the same ones that were identified in previous studies conducted on Western samples: violence, sex, and addiction-type hazards received the highest risk ratings. Women's mean risk judgments were systematically higher than men's mean judgments, and older participants' mean judgments were higher than younger participants' mean judgments regarding domestic hazards. Technically trained people perceived higher societal risk than people with training in the humanities, but the difference between people with technical training and people trained in the humanities was more accentuated among younger people than among the elderly. Also, men living in Tula perceived themselves as less exposed to risks of violence than women living in Tula and both men and women living in Moscow.

Key Words:

• risk perception
• health
• environment
• Russia
• gender
• education
• media

INTRODUCTION

“People respond to hazards according to their perceptions of the risks they pose. What they perceive, why they perceive it that way, and how they will subsequently behave is a matter of great import to industries and governments trying to assess and implement new technologies” (Peters and Slovic 1996, pp 1427–1428). The aim of the present work was (a) to analyze new information regarding the way in which present day Russian people perceive societal risks, (b) to examine the way in which broad demographic characteristics—age, gender, education, and city of residence—were associated with risk perception, and (c) to compare these risk perceptions with the ones that have been recently evidenced in a typical Western country (Bouyer et al. 2001).

SOCIETAL RISK PERCEPTION

Studies on societal risk perception by Slovic et al. (1985, see also Slovic 1987.) have inspired numerous similar studies in a variety of countries (see Boholm 1998 and Mullet et al. 2004, for recent reviews). In these studies, participants—usually, college students—were instructed to rate the overall severity of hazards such as nuclear weapons, vaccinations, smoking, and herbicides on scales ranging from Not at all risky to Extremely risky. These studies have shown that risk ratings may substantially vary from one country to another. As an example, the level of risk associated with the item “Nuclear Power Plants” was rated 31 (out of 100) in a Hungarian sample (Englander et al. 1986), 34 in a Swedish sample (Nyland 1993), 47 in a Norwegian sample (Teigen et al. 1988), an Australian sample (Finucane and Maybery 1996), and a Finnish sample (Mullet et al. 2004), 49 in a Polish sample (Goszcynska et al. 1991), 62 in an Italian sample (Savadori et al. 1998), 66 in a Macanese sample (Neto and Mullet 2001) and a Brazilian sample (Nyland 1993), 67 in a Portuguese sample (Neto and Mullet 1999), 68 in a Hong Kong sample (Keown 1989), 69 in a French sample (Karpowicz-Lazreg and Mullet 1993), 72 in a U.S. sample (Slovic et al. 1985) and a Korean sample (Cha 2000), 76 in an Egyptian sample (Ahmed et al. 2006), 78 in a Burkina Faso sample (Koné and Mullet 1994), and 84 in a Spanish sample (Muñoz Sastre et al. 2006).

The effect of the perceiver's personal characteristics on risk perception have also been extensively studied. Among these characteristics, three have been systematically examined: age, gender, and educational background. Hermand et al. (1999) compared mean risk magnitude judgments expressed by participants aged 8 to 75. Between elderly people and adults, very few notable differences were detected. Overall mean estimations were very close, and linear associations between younger and older adults' ratings were very high (always higher than 0.90). Karpowicz-Lazreg and Mullet (1993) showed that between women's and men's ratings, and between participants with scientific educational background and participants with artistic educational background, the linear associations were also very high (see also Gustafson 1998; Sjöberg 2000). The only marked differences between men and women concerned the risks associated with domestic activities and public transportation (see also Barke et al. 1997; and Slovic 1997), and the only differences between artists and scientists concerned health risks. Among the other personal dimensions considered one can cite income (e.g., Flynn et al. 1994; Palmer 2003), religious orientation (e.g., Sjöberg and af Wahlberg 2002), political preferences (e.g., Sjöberg 2000), risk attitude (e.g., Vollrath et al. 1999; Sjöberg 2003), values (e.g., Sjöberg 2000), social trust (e.g., Vicklund 2003; Siegrist et al. 2000), anxiety (e.g., Bouyer et al. 2001; Källmén 2000), self-efficacy (e.g., Kouabenan 1998; Källmén 2000), locus of control (e.g., Källmén 2000), worldviews (e.g., Bouyer et al. 2001; Brenot et al. 1998), and classical personality factors (e.g., Sjöberg 2003). Typically, the part of the variance in risk assessment explained by all these personal factors has proven to be low (Sjöberg 2003).

The only data gathered to date on Russian samples were published by Mechitov and Rebrick in 1990 (in other words, before the many transformations that have affected Russia since then). In addition, the sample of participants in their study was composed of only 25 people (who were research assistants), and the set of selected items was limited to 50 (and did not include important items such as Nuclear Power and Industrial Pollution). Mechitov and Rebrik (1990) essentially showed that before the fall of the Wall, their participants' risk ratings were considerably lower than the ones registered in studies conducted in Western countries such as the United States, France, or Italy. These ratings were, however, close to the ones observed in a study conducted in Hungary (Englander et al. 1986) on a larger and more diversified sample.

Since the fall of the Wall, the economical and sociopolitical situation in Russia has changed considerably; namely, new products, new activities, and new technologies have been introduced. The Internet, today accessible to many Russians, is a new and practical source of information on every subject, including accidents (which were seldom reported in the former USSR). It is therefore conceivable that all of the societal changes, which have produced much uncertainty, may have had an impact on the way Russians assess hazards. Thus, conducting a new survey on risk perception in Russia seemed timely.

We expected that the main concerns of the Russians in 2003 would be the same as the ones that were identified in previous studies conducted among Western samples (e.g., Bouyer et al. 2001; Savadori et al. 1998). In particular, we expected that the Russians were particularly concerned by warfare, terrorism, and crime (e.g., nuclear weapons), by the propagation and use of illicit drugs (e.g., heroin), and by environmental issues (e.g., nuclear power plants).

Although gender and age have been shown not to be strong determinants in risk perception, they were systematically examined. Indeed, in present day Russian society, unemployment, poverty, and violence are common ingredients of everyday life(Williams et al. 2001). In such uncertain contexts, as common sense indicates, females tend to feel more vulnerable than do males. As a result, they should perceive life as riskier, especially as regards physical violence and outdoor activities in general (Williams et al. 2001).

Although age has also been shown not to be a strong determinant in risk perception, age differences were also systematically examined because older people tend to feel generally less adaptable than younger people. As a result, they should perceive life as riskier than do younger people (Williams et al. 2001).

Finally, although educational background has been shown not to be a strong determinant in risk perception, educational background differences were examined because important technical changes have occurred in Russia. It is fair to state that in the previous system, the production system was largely obsolete, and the people employed in the factories had only limited knowledge about the more advanced Western technologies. Therefore, after the transformation, people with technical training may have possibly found themselves more exposed to the dramatic changes that occurred in the production system than the people trained in the humanities who were therefore less aware of changes and perhaps able to more easily adapt.

METHOD

Participants

The sample comprised 800 Russian participants who all agreed to participate in the study, either in Moscow or in Tula (a city of 800,000 inhabitants, located 250 km south of Moscow). The sample comprised eight subgroups: (a) 100 young participants living in Moscow, aged 18–25 and trained in humanities (e.g., foreign languages, history, sociology, economics), (b) 100 young participants living in Moscow, aged 18–25 and with scientific/technical training (e.g., physics, biology, civil engineering, mechanics), (c) 100 older participants living in Moscow, aged 40–55 and trained in humanities, (d) 100 older participants living in Moscow, aged 40–55 and with scientific/technical training, (e) 100 young participants living in Tula, aged 18–25 and trained in humanities, (f) 100 young participants living in Tula, aged 18–25 and with scientific/technical training, (g) 100 older participants living in Tula, aged 40–55 and trained in humanities, and finally (h) 100 older participants living in Tula, aged 40–55 and with scientific/technical training. Each subgroup comprised 50 females and 50 males. No member of ethnic minorities was included in the sample. The younger participants were recruited in the campus of the universities of Moscow and Tula. The older participants were recruited through professional associations. The 50 participants in each subgroup were the first 50 persons who agreed to participate in the study. All participants were non-paid volunteers.

Material

The questionnaire comprised 79 items borrowed from Slovic, Fischhoff, and Lichtenstein's risk questionnaire, 28 items borrowed from Bouyer et al. (2001), and 3 additional items introduced in order to reflect the present day living conditions in Russia. The Russian version of the series of 79 items was composed based on the original English version, whereas the Russian version of the 28 items was composed based on the original French version. In designing the Russian version, the author followed the guidelines proposed in the literature on cross-cultural methodology (Brislin 2000): independent/blind/back translation, educated translation, and small-scale pretests (Brislin 2000, p. 798). A Russian translation was executed by a trilingual translator (French, English, and Russian) and sent to another translator for back translation. The trilingual translator also verified, subsequently, that the final Russian version was consistent with the original English and French versions.

Procedure

The data were gathered in 2003. Each participant responded individually. Before responding, participants were informed that risk referred to the risk of serious illness, serious injury, or death (as in the original study by Slovic et al.). The exact wording was: “Please rate the degree of risk for health and environment you associate with each of the substances, activities, technologies, and situations listed below.” They were then invited to ask questions about terms they were unfamiliar with and to fill in the questionnaire. When an item was not sufficiently clear, participants were instructed to not respond. Responses were given on an 11-point scale labeled from “No risk” to “Extremely Severe risk.” No time limit was imposed. It took approximately 40 minutes to complete the questionnaire.

Responses were recorded in two successive steps. First, participants were invited to circle, with a black pencil, the number (0 to 10) that best reflected their view. Then, they were instructed to change their responses using a red pencil (only this second set of ratings was taken into account in the analyses.) In particular, they were asked to try to clearly differentiate between risks. The main goal of this two-step procedure was to ensure that participants used the entire range of the response scales (see Hermand et al. 1999).

RESULTS

The Main Concerns

The mean judgments and standard deviations registered for the 110 items are shown in Table 1. Not surprisingly, participants were mainly concerned about organized violence (nuclear weapons, warfare, contract killings, kidnapping of children, rapes, criminality, and terrorism), illnesses that are very difficult to treat (AIDS or other severe sexually transmitted illnesses, open-heart surgery), and addictions (heroin, marijuana). They were much less concerned with substances used on a daily life basis (cosmetics, hair dying), with sport activities (jogging, swimming, bicycle), and with common, well accepted technologies (air conditioning, home appliances).

Table 1 Means and standard deviations for the 100 items. Results of the exploratory factor analysis.

	Russia
			Factors
Items	M	SD	I	II	III	France M
Abandoned Animals	4.28	2.86	.37		.36
AIDS	8.53	2.36		.71
Air Conditioning (in Public Areas)	2.52	2.34	.38		.46	2.12
Alcoholic Beverages	5.41	2.72			.38	5.38
Amphetamines	5.02	2.71			.36	7.09
Anesthetics	4.05	2.43			.51	5.17
Antibiotics	4.38	2.55			.57	3.21
Asbestos	5.27	3.17			.58	7.97
Aspirin	3.15	2.45	.36		.49	2.22
Attacks in the Subway	6.85	3.02		.70
Barbiturates	5.08	2.73			.31	6.17
Bathing	2.56	2.01	.64
Bicycle	2.14	2.18	.58			2.87
Boxing	4.75	3.13	.39			4.32
Bridges	3.33	2.32	.63		.30	2.91
Caffeine	3.36	2.48			.55	3.43
Cellular Telephones	3.78	2.76	.39		.55
Charters	4.94	2.78	.51	.35		4.23
Chemical Disinfectants	4.41	2.51	.42	.38		6.31
Chemical Solvents	4.95	2.59	.44	.38
Chemical Fertilizers	5.02	2.61	.38	.39	.34	6.55
Childbirth	5.21	3.01		.31		3.03
Christmas Tree Lights	2.57	2.15	.52		.42	1.18
Commercial Aviation	4.85	2.90	.44			4.75
Conscription	6.06	3.19		.19
Contraceptives	4.48	3.03			.44	2.33
Contract Killings	8.06	3.30		.56
Cosmetics	2.59	2.24	.32		.53	2.07
Crime	7.73	2.37		.58		8.29
Dams	3.13	2.15	.52			3.70
Diving	4.00	2.81	.63	.33		4.48
DNA Research	2.78	2.77	.49			6.10
Domestic Animals	2.28	2.06	.53		.37
Dynamite	7.74	2.78		.56	.30	7.14
Electro Domestic Tools	3.20	2.42	.55			3.47
Equipments in Bad State	6.11	2.51		.49
Explosions at Home	7.87	2.93		.60
Exported Food that has Passed its Use-By Date	6.65	2.70		.47	.40
Family Violence	6.94	3.15		.68
Firefighting	6.30	2.75	.34	.52		5.14
Fireworks	3.50	2.52	.62
Food Coloring	4.86	2.66			.55	2.91
Food Preservatives	5.00	2.86	.50	.34		3.17
Gas Furnaces	3.60	2.32	.49			5.13
Hair Dying	2.99	2.53	.30		.52	1.73
Handguns	6.91	2.87		.60		8.50
Hashish	6.99	3.30		.54	.41
Herbicides	6.10	2.78		.35	.50	5.64
Heroin	8.39	2.93		.63		8.58
Home Appliances	2.71	2.30	.42			2.14
Homosexual Relationships	5.82	3.66		.45	.39	6.63
Hunting	4.12	2.68	.38	.36	.30	4.89
Hydroelectric Power Plants	3.72	2.66	.40			4.09
Immigration	4.17	2.75	.32
Industrial Pollution	6.93	2.56		.49	.34
Internal Terrorism	8.69	2.22		.64
Jogging	1.48	1.88	.50			1.69
Kidnapping of Children	8.17	2.60		.71
Lasers	3.86	2.73	.43		.38	4.31
Lifts	3.16	2.32	.54			2.24
Liquid Natural Gas	4.75	2.57	.34		.34	4.69
Marijuana	7.03	3.37		.57	.37	6.24
Microwave Ovens	3.72	2.63	.36		.57	3.35
Morphine	6.60	3.12		.46	.39	6.73
Motor Vehicles	4.60	2.31	.59			6.35
Motorcycles	5.18	2.68	.43	.40		4.93
Mountain Climbing	5.81	2.90	.55	.45		6.06
Mushroom Hunting (and Consumption)	4.09	2.37	.46			3.33
National Defense	4.68	2.92
Natural Catastrophes	7.79	2.63		.66
Nerve Gas	5.71	2.79	.33	.51
Non-Nuclear Power Plants	4.35	2.66	.49	.30		4.42
Non-Protected Sexual Relationships	7.80	2.65		.39
Nuclear Weapons	9.16	1.90		.58		9.14
Nuclear Power Plants	6.92	2.98	.32	.37		7.24
Open-heart Surgery	7.46	2.55		.61		6.39
Passive Smoking	5.83	3.09		.35	.42
Personal Car	4.23	2.61	.54
Pesticides	5.83	2.68		.47	.46	6.18
Police Work	5.70	2.69	.30	.46
Power Mowers	2.34	2.26	.55			2.54
Pregnancy	3.55	2.92	.30			3.07
Prescribed Drugs	3.30	2.25	.32		.45	3.29
Private Aviation	5.00	2.62		.31	.54
Public Transportation	3.66	2.37	.55
Public Demonstrations	4.01	2.52	.41			5.54
Radiotherapy	5.10	2.63		.40	.32	4.15
Railroads	3.47	2.26	.63			4.03
Recreational Boating	2.83	2.32	.63			3.23
Saccharine	3.79	2.67			.47	2.67
Satellites	2.98	2.74	.60			2.77
Skateboard	4.21	2.68	.59	.36		2.61
Skiing	4.03	2.66	.53			3.89
Smoking	6.42	3.05		.38	.36	7.08
Soccer	2.60	2.17	.59			2.24
Sodium Nitrite	5.88	2.57		.45	.44	5.80
Space Exploration	3.35	2.93	.51			4.42
Sunbathing	3.77	2.63	.42		.47	5.09
Surfing	4.00	2.77	.56	.33		3.20
Surgery	5.97	2.48	.44	.43		4.71
Swimming Pools	1.66	1.99	.44			3.14
Tall Buildings (Skyscrapers)	3.25	2.57	.53			3.64
Terrorism	8.70	2.21		.67		7.99
Tourism	2.79	2.15	.68
Tranquilizers	6.16	2.66		.52	.34	4.88
Ultra Violets Rays	4.52	2.63	.36		.42
Vaccinations	3.45	2.61	.34			2.56
Warfare	9.04	2.16		.53		7.47
X-Rays	5.35	2.76			.38	5.90
Explained Variance			15.99	14.07	10.28
Part of Explained Variance			.15	.13	.09
Note: Figures higher than .50 have been put in boldface.

Structure of Risk Perception

Also shown in Table 1 are the results of an exploratory factor analysis conducted on the whole set of items. The extraction method was the principal components method. Based on the scree test, as well as on more advanced techniques (Zwick and Velicer 1986), a three-factor solution was chosen, and VARIMAX rotations were applied.

The first factor was called Outdoor Activities. It explained 15% of the variance. It was loaded by items such as recreational boating, skiing, surfing, skating, jogging, diving, bathing, soccer, bicycle, charters, motor vehicles, railroads, bridges, and tourism. A score was computed by simply averaging the values observed for the ten items with the highest loading on this factor. The mean value was 3.43. The second factor was called Violence, Sex and Addictions. It explained 13% of the variance. It was loaded by a set of items corresponding to violent acts and violent technologies (e.g., nuclear weapons, contract killings, terrorism), and by a set of items corresponding to hallucinogenic substances and potentially risky sex (e.g., heroin, smoking, homosexual relationships). The mean value was 7.84; that is, much higher than the mean value observed for the fist factor. The third factor was called Domestic hazards. It explained 9% of the variance. It was loaded by a whole set of items corresponding to everyday hazards and accepted domestic technologies: Hair dying, microwave ovens, cellular telephones, antibiotics, food coloring, asbestos, home appliances, and cosmetics. The mean value was 4.00; that is, intermediate between the two other values but closer to the one found for the first factor.

Association with Gender, Age, Education, and Location

An ANOVA was conducted on the three mean scores with a Hazard Factor × Gender × Age × Education × Location, 3 × 2 × 2 × 2 × 2 design. Hazard factor was a within-subject factor and the other factors were between-subjects factors. Owing to the great number of comparisons, the significance level was set at p < .001.

The effect of the Hazard factor was significant, F(2, 1568) = 2896, p < .001. The set of items forming the Violence, Sex and Addiction factor was perceived as more risky than the set of items forming the Outdoor Activities factor.

The effect of gender was significant, F(1, 784) = 41.02, p < .001. Women (5.39) perceived the hazards as riskier than men did (4.84). The effect of age was significant, F(1, 784) = 28.73, p < .001. Older participants (5.34) perceived the hazards as riskier than did younger participants (4.84). In addition, the Hazard × Age interaction was significant, F(2, 1568) = 6.97, p < .001. The effect of age was more pronounced (4.38 versus 3.62) as regards the domestic-type hazards than as regards the other hazards.

The effect of type of education was significant, F(1, 784) = 15.94, p < .001. Participants with technical training (5.27) tended to perceive the hazards as riskier than participants trained in the humanities did (4.91). The Hazard × Education interaction was significant, F(2, 1568) = 11.02, p < .001. The difference in risk perception observed as a function of education was mostly concentrated on the domestic-type hazards. Domestic hazards were perceived as riskier among participants with technical training (4.36) than among participants trained in the humanities (3.65). As for the two other hazard factors, no such difference was observed.

The Age × Education interaction was significant, F(1, 784) = 25.66, p < .001. The effect of type of education was only present among the younger participants. The younger participants trained in the humanities perceived the hazards as less risky (4.42) than the younger participants with technical training (5.26). Among the older participants, there was no difference as a function of type of training (5.39 and 5.29). In addition, the Hazard × Age × Education interaction was not significant; that is, the younger participants trained in the humanities perceived the hazards as less risky than the younger participants with technical training, irrespective of the kind of hazard considered.

The location factor was not significant but the Hazard x Location interaction was, F(2, 1568) = 15.84, p < .001. The participants living in Moscow perceived the violence, sex, and addiction-type hazards (8.05) and the domestic hazards (4.20) as riskier than the participants living in Tula did (7.63 and 3.80, respectively).

The Hazard × Gender × Location interaction was significant, F(2, 1568) = 7.97, p < .001. The gender difference was higher as regards the perception of the violence, sex, and addiction-type hazards among participants living in Tula than in any other condition. Men living in Tula (7.06) perceived these hazards as less risky than women living in Tula did (8.21). The Hazard × Age × Location interaction was also significant, F(2, 1568) = 9.14, p < .001. The age difference was stronger as regards the perception of the violence, sex, and addiction-type hazards, and as regards the domestic hazards among the participants living in Moscow than among participants living in Tula. Conversely, the age difference was weaker as regards the perception of outdoor activities-type of hazards among participants living in Moscow than among participants living in Tula.

Comparison with Data Gathered in France

Comparable data were gathered in France in 1999 using a questionnaire that has, partly, inspired the one used in the present study. The mean risk values observed in France are also shown in Table 1. At first glance, these vales were very similar to the ones observed in present day Russia. A series of t tests were performed between the two series of values. For three items the difference between Russia and France was higher than 2 points, the hazard being perceived riskier among the French than among the Russian. The corresponding items were DNA research, asbestos, and amphetamines. For an additional set of 11 hazards, the difference was higher than 1 point. The corresponding items were chemical disinfectants, swimming pools, gas furnaces, public demonstrations, chemical fertilizers, handguns, power mowers, sunbathing, anesthetics, barbiturates, and space exploration.

For two items, the difference between France and Russia was higher than 2 points, the hazard being perceived riskier among the Russian than among the French. The corresponding items were Childbirth and Contraceptives. For an additional set of 12 hazards, the difference was higher than 1 point. The corresponding items were fire fighting, surgery, saccharine, antibiotics, tranquilizers, hair dying, open-heart-surgery, skateboard, Christmas tree lights, warfare, food preservatives, and food coloring. All these differences were significant, p < .001.

DISCUSSION

The present study examined recent data on risk perception gathered from several Russian samples. The main concerns of the Russians in 2003 were expected to be the ones that were identified in the many previous studies conducted among Western samples. This is what was observed. Risk judgments showed a 3-dimension structure that was easy to interpret. Some elements of this structure were highly reminiscent of factors previously found in Bouyer et al. (2001) and in Hermand et al. (2003). Two of these factors were highly similar in content. The Domestic Hazards evidenced in the present study was close to the Common Individual Hazards evidenced in Bouyer et al. (2001). Also in both studies, the Outdoor Activities factor had the same content. Finally, the Violence, Sex and Addiction factor that was evidenced in the present study may be considered as including the Deviance, Sex and Addictions factor, the Urban Violence factor, and the Weapons factor found in Bouyer et al. (2001). The level of risk associated with the Violence, Sex and Addiction factor was much higher than the level of risk associated with both other factors. A comparison with data gathered in France on the same series of hazards showed that in 50 cases (out of 79 comparable items), the difference did not exceed one point, and in only five cases, the difference exceeded two points. In particular, the French cohort's ratings were notably higher as regards three hazards (DNA research, asbestos, and amphetamines), the risky nature of which has recently attracted much attention in this country, and notably lower as regards two hazards with which the French are either well familiarized (contraceptives) or that tend to be well mastered in this country (childbirth). Overall, these results were consistent with findings by Sjöberg et al. (1996), who showed that risk perception in Bulgaria and Romania has considerably increased as a function of the many changes that occurred in these countries.

It was also expected that gender differences should be present, with, in general, women assessing risks higher than men. This is what was observed: women's mean judgments were systematically higher than men's mean judgments. It was expected that age differences should be present, with, in general, older people assessing risks higher than younger people. In fact, older participants' mean judgments were practically always higher than younger participants' mean judgments. It was, finally, expected that systematic educational differences should be present, with technically trained people perceiving higher societal risk than people with training in humanities. This is what was observed as regards domestic hazards.

Two other findings deserve consideration. First, the difference between people with technical training and people trained in the humanities was shown to be more accentuated among younger people than among the elderly. This may be explained by the fact that younger participants trained in the humanities tend to see more opportunities in the changes occurring in the Russian society than both older participants (who are generally less adaptable than younger participants) and participants with technical training (whose training tends to become obsolete with time).

Second, some differences between males and females were shown to be more accentuated in Tula than in Moscow. These differences concerned the Violence, Sex and Addiction factors. Men living in Tula perceived themselves as less exposed to these risks than women living in Tula and both men and women living in Moscow. This result was reminiscent of findings by Flynn et al. (1994), who showed that for many technologies and many substances, the difference in risk perception between affluent people and less affluent people was mediated by gender. More precisely, affluent males perceive these hazards as less risky than non-affluent males, while affluent females and non-affluent females tend to perceive these hazards in the same way as non-affluent males (see also Palmer 2003). In other words, Flynn et al. (1994) evidenced a robust Gender × Socioeconomic Level interaction as regards to risk perception. Although our findings did not measure any Gender × Socioeconomic interaction (mainly because we were not in a position to accurately measure participants' socioeconomic level), it did evidence a Gender × Location interaction. In Moscow, violence, robbery, and addictions tend to affect both males and females more or less in the same way, and, as a result, risk perception as regards these hazards should be more or less equivalent for both males and females. By contrast, in Tula, a smaller city, males may feel more secure than males living in Moscow (because of lower criminality level); however, females may not feel more secure than females living in Moscow, partly because of endemic males' violence toward females (at least in urban areas).

Theoretical and Political Implications

From a theoretical viewpoint, one of the main findings of the study was that, although in general, gender, age, and educational level do not tend to alter risk perception, in specific cases where economic and societal changes have been fast and strong (e.g., in Russia for the last 15 years) these factors can have a non-negligible impact. This impact was in the expected direction: the participants who, due to their gender or their age or their educational background were more affected by the changes, were also the ones whose risk perception levels were higher. Furthermore, these factors interacted in a mostly meaningful way, which was reminiscent of previous findings.

From a practical viewpoint, the present study clearly showed that present day Russia and the Western world share very similar views as regards risk perception. As a result, joint environmental policies and projects can be considered and discussed between Russians and Europeans without running the risk of complete misunderstanding from both parties.

Also, as the Russian population is now clearly sensitized to many health and environmental concerns, it will become more difficult for any future Russian government not to take into account these concerns when formulating health, security, and environmental policies.

ACKNOWLEDGMENT

This work was supported by the Laboratoire Ethique and Travail (Ecole Pratique des Hautes Etudes) and by the UMR Travail and Cognition (Mirail University). Thanks are extended to Félix Neto for his many helpful suggestions.