Values and social representations of HIV/AIDS in Central and Eastern Europe:

A multi-method investigation in five nations  

Parts (First section of report):  1 2 3 4

3.2. Results

3.2.1. Knowledge of HIV amongst our respondents

Overall, respondents scored relatively highly on the knowledge questions about the transmission of HIV,

although for some items there was undue concern over transmission as well as some ignorance about methods

of protection from transmission. Thus almost half of the total sample (45%) thought that sharing

knives/plates/ forks with someone with HIV was ‘moderately’ or ‘highly’ risky, and 55% thought being

coughed or sneezed on by someone with AIDS was moderately or highly risky. 45% of nurses and 40% of

doctors thought kissing someone with AIDS was ‘moderately’ or ‘highly’ risky, 56% of nurses thought

sharing plates was ‘moderately’ or ‘highly’ risky, and 66% of nurses (and 29% of doctors) thought ‘being

coughed on or sneezed at’ was ‘moderately or highly risky’ (46% or nurses and 29% of doctors rated this as

 ‘very risky’). 30% of nurses thought using a diaphragm or having a vasectomy were ‘moderately effective’

at preventing HIV infection.

There were significant cultural and occupational variations in levels of knowledge about the manner in which

 HIV could be spread as well as a significant country x occupation interaction. The most knowledgeable

nationality were the Estonians (adjusted mean = 38 out of 45, a knowledge score significantly higher than

that scored by any other country). Least knowledgeable were the Russians and Georgians (adjusted Ms of 32.3

and 33.4). As might have been anticipated, the most knowledgeable occupational group were the doctors

(adjusted mean of 36) but interestingly least knowledgeable were the nurses (adjusted mean of 32.8: significantly

 worse than the doctors). The most knowledgeable group overall were the Estonian doctors (39.6), least were

 the Georgian nurses (29.6). All 26 Polish nurses thought that being coughed on by someone with AIDS was

 ‘very high risk’, and all but one of the 23 Polish doctors.

3.2.2. Sexual behaviour across the sample

Approximately a third of the sample had asked their partner if they had been tested and knew the HIV status

of their partner. Nurses were most likely to report they knew this (47% claimed to know the HIV status of their

 partner compared to only 25% of the business people). Only 10% of Georgians, but 46% of Russians, claimed

they had asked their partner about their HIV status. 49% of Georgians claimed they did know the status of their

 partner, compared to only 14% of the Hungarians (x² (4) = 34.88 p< .001).

44% of Russians, but only 11% of Georgians, had had an HIV test, data that reflects the interview responses

reported in stage 1 in our earlier data collection (where Russians were most likely to report that they had been

tested and Georgians the least). Most of the respondents (91%) did not feel themselves at risk of getting AIDS,

but only 63% (and 48% of Hungarians) said ‘it did not worry them’. 91% did not feel they had been exposed to

HIV, and only 4% reported having sex with a high risk person (8% of business people ticked this box). Finally,

almost half (43%) did feel it was possible that they could contract HIV, but they did not worry about it ‘every

day’. Doctors (53%) and Russians and Estonians (61% and 56%) were most concerned about this.

72% of the sexually active sample reported that they had had only one regular sexual partner over the previous

 six months (10 % of the sample reported they were not sexually active over the past six months). 16% of the

complete sample reported more than one sexual partner a month, 7% more than one partner a week. Russians

reported the most sexual partners overall, with 48% of Russian business people - and 43% doctors - reporting

more than one sexual partner over the past six months (see figure C). Almost a third of Russian business people

 reported more than one sexual partner a month, almost 20% more than one a week. Similarly, 30% of Russian

doctors report more than one sexual partner a month, and 20% more than one a week. In contrast, 90% of sexually

 active Poles reported only sexual partner during the previous six months.

Those highest on risk (multiple partners over the past six months and poor record of condom use) were the

Russian respondents, with Poles the least (F (4, 450) = 5.48 p< .001). Business people were the highest risk group

(F (2, 450) = 5.85 p< .003). There were no nation x occupation interactions. Overall, men were significantly more

likely to engage in multiple partnerships than women (x² (1) = 20.46, p< .001)

17% of the sample had had a previous STD. An STD history was reported more by Russian, Georgian and

Estonian respondents (25-27%) more than Hungarian (5%) or Polish (2%) respondents (x² (4) = 42.3 p<.001)

(F (4, 463) = 5.80 p< .001). Business people (23%) or doctors (15%) rather than nurses (4%) (x² (2) = 16.7 p<.001)

 (F (2, 463) = 8.27 p< .001) were more likely to report STDs. There were no nation x occupation interactions. Men

were more likely to have had a sexual disease than women (x² = 26.13 p< .001).

In analyses of variance, condom usage was also dependent on nationality (F (4,450) = 4.53 p< .001) but not on

occupation (F (2, 450) = .34) or an interaction of nationality and occupation. Condom usage was highest in Poland

 (M = 2.34 on the four point scale), significantly greater than in Georgia (M = 1.83) or Estonia (M = 1.78).

3.2.3. Values and sexual behaviour

We next compared rankings of Schwartz and Bardi’s values data for these five nations with our own findings.

 Data was standardised using a procedure described by Schwartz (1999): mean scores were first computed for

each sample and the sample mean was subtracted from the scale mid-point. These difference scores were then

added to the score for each value type. Table 5 gives mean scores and rank orders for our cultural-level values.

Our findings here provided moderate support for the ranking of values across the two data sets. As we

anticipated (Hypothesis 1), Georgians and Poles were the most conservative respondents and Russians the

most Hierarchical (Hypothesis 2). In our sample, egalitarianism was again highest amongst Georgians and

Estonians, as anticipated from Schwartz & Bardi’s (1997) data (Hypothesis 3). However, the Poles were not

the highest scorers on Mastery as hypothesised (Hypothesis 4) : here it was the Russians who were

unexpectedly high scorers here. Finally, Estonians were anticipated to be the highest on Harmony scores

(Hypothesis 5). Our findings demonstrate a great deal of similarity in scores across cultures here, but it was

the Hungarians (2^nd in Schwartz’s data) and Poles (3^rd in Schwartz’s data) who were highest on harmony.

Team member Dr Realo then conducted an analysis of the structure of the collectivism measure used in our

analysis (Realo, 2000). A principal components analysis of all 21 items was followed by a varimax rotation

using all subjects as if they came from one sample. Three clearly distinguishable types of collectivism,

focusing on relations with family, peers and society, were found. The three-factor solution accounted for

46.1% of the total variance and provided for the best possible simple structure: each item loaded above .30

on appropriate factors whereas only five items had secondary loadings (none above .34) on another factor.

This provided us with good support for the validity of this scale as used in this context. A one-way ANOVA

 revealed significant differences between samples on all three types of collectivism (F (4, 494) = 11.3 for family

subscale, 51.7 for peer scale and 20.0 for society subscale, p< .001 for each)(see figure D). The Georgian sample

was most collectivist overall, especially with respect to peers and society, whilst the Estonians were significantly

 less collectivistic in regard to family and peers compared to all other samples.

We then conducted a further analysis of the relationship between individual-level values scores and sexual

behaviour. Table 6 provides an overall correlation matrix for all the major variables assessed in this study.

 Individual-level scores were obtained using a standardisation procedure described by Schwartz (1999). In this

 procedure mean scores of all 10 values are calculated for each country to produce 5 mean scores. Also

calculated were global mean scores across all participants across all 29 values. From each value type score for

 each country we subtracted the local mean score and added global mean scores for all participants to produce

 standardised scores (Schwartz, 1999). Above we hypothesised that those high on Openness to Change

(Hypothesis 6a) and high on Hedonism (Hypothesis 7) would be expected to be more likely to have more sexual

encounters and have had a greater exposure to sexual disease. A simple t-test analysis confirmed our expectations

here: those highest on risk (who had multiple partners over the previous six months and only rarely used a

condom) were high on Hedonism (t (457) = -4.82 p< .001) and high on Openness to Change (low on traditionalism)

 (t (446) = -4.04 p< .001). ‘High risk’ individuals were also high on Self-enhancement (achievement and power)

versus Self-transcendence (universalism and benevolence)(t (453) = 6.06 p< .001). Sexual disease was also

higher amongst those scoring high on these three dimensions, with sexual disease prevalence greater amongst

 those high on Hedonism (t (484) = 2.31 p< .02), Openness to Change (t (473) = 5.22 p< .001) and Achievement

and Power (t (480) = -4.56 p< .001).

Collectivism and Openness to Change were highly correlated, as we might anticipate (r (490) = .54 p<.001).

Collectivists were more conservative in their sexual behaviour, reporting less multiple partnerships

(t (470) = 2.26 p< .02). They were less likely to report multiple partnerships with low condom use (t (440) = 1.00 p< .05)

and were less likely to report any record of previous sexual disease (t (465) = -2.84 p< .01) (Hypothesis 8a)

(see Figure E). Also as anticipated (Hypothesis 8b), the strongest correlations between collectivism and sexual

behaviour were between the sub-scale scores on familism and sexual behaviour, with those high on familism

less likely to report multiple partners (t (487) = 3.65 p< .001), multiple partners with low condom use

(t (455) = 2.32 p< .02) or to have previously had a sexual disease (t (482) = -3.89 p< .001). In addition, those who

 reported that they have had sex with someone who could possibly have put them in risk for AIDS (n = 21) or

who said that it is possible that they will catch HIV one day but they do not worry about it everyday (n = 217)

were significantly less family-collectivistic (M= 3.30, SD = .83 and M= 3.57, SD = .79, respectively) than those

who felt that they were not vulnerable to catching the AIDS virus (M = 3.69, SD = .76 and M = 3.75, SD = .74)

with respective F-values (1, 195) = 6.6 and 6.4, p < .05.

As predicted (hypothesis 9) fatalists were significantly more likely to report multiple partnerships

(t (489) = -2.23 p< .03) and to belong to the high risk group with multiple partners and low condom use

(t (457) = -2.56 p< .01). Those high on optimistic bias were less likely to perceive themselves at risk

(r (499) = .12 p< .01), supporting hypothesis 10. In addition, those low on optimistic bias were more likely to participate

 in risky sexual activities (t (489) = -5.57 p< .001; for multiple partnerships; t (457) = -4.22 p< .001 for multiple

partnerships with low condom usage; t (484) = 4.17 p< .001 for sexual disease history). Trust levels were a

significant correlate of very risky sexual activity only (t (458) = 2.29 p< .02), with those high on trust less likely

 to belong to this riskiest of groups (Goodwin, Nizharadze, et al, 2000).

We finally conducted a series of multivariate analyses in which we examined the impact of a combination of

demographic factors, alongside scores on Schwartz’s values, knowledge of HIV, trust, fatalism and concern

about life events, on sexual behaviour and knowledge. An ANCOVA on condom usage found usage highest

in Poland (M = 2.34) and lowest in Georgia and Estonia (Ms of 1.83 in both countries)(F (4, 421) = 2.48 p .04).

Post-hoc Scheffe comparisons showed significant mean differences between Poland and Georgia and Estonia

 (mean difference .52 p< .04) and there was a country x occupation interaction, with condom use highest amongst

Polish nurses and lowest amongst Georgian nurses. Openness to Change was also significantly correlated with

 condom usage, with those open to change more likely to use condoms (F (1,421) =4.78 p< .03). There were no

other significant influences on condom use for the variables analysed. We then conducted a series of

hierarchical logistic regressions, using the forward stepwise method (Likehood Ratio) for variable selection.

For multiple partnership over the previous six months the inclusion of country of participation, sex, and scores

on Hedonism, Openness to Change Self-transcendence, unrealistic optimism and fatalism allowed us to accurately

 classify 79.1% of cases (x² (2) = 20.5 p<.001). For the high risk group who experienced multiple partnerships and

had low use of condoms analysis led to the inclusion of sex, the interaction between country and occupation,

and scores on Hedonism, Self-transcendence, Openness to Change and fatalism and unrealistic optimism.

This allowed for accurate classification of 84.0% of cases (x² (2) = 16.74 p<.001). For the prediction of a sexual

disease ever, sex, Openness to Change, Hedonism and Self-transcendence, along with unrealistic optimism,

allowed us to classify 83.7% of cases (x² (1) = 7.95 p<.001). Finally, knowledge of the partner’s status was

 predicted accurately in 72.1% of cases by Openness to Change, Self-transcendence, Hedonism and fatalism

 (x² (1)= 9.01 p< .002). These findings are summarised in table 7.

3.3. Representations of HIV: Using matrix data

We used a modified version of the free associations produced by our 500 respondents to formulate a matrix

task that could then be subjected to individual-level multi-dimensional scaling (Carroll & Chang, 1970). Here

 we focused on the twelve words most frequently mentioned in the earlier task: AIDS (the stimulus word);

Africa; Blood; Casual sexual activity; Condoms; Death; Disease; Drugs; Fear and fright; Homosexuals;

Prostitutes and Sex.

Respondents were given twelve cards (with one word on each) and asked to rank order the level of similarity

of each to each other, forming a 12 x 12 matrix for each respondent. These 500 matrices were then analysed

alongside our questionnaire.

3.4.Results from the matrix analyses

Analysis here was primarily conducted by Dr Kwiatkowska with the aid of Professor Wierczorkowska –

Nejtardt of the Polish Academy of Sciences (Kwiatkowska, 2000; Kwiatkowska et al, 2000a). A two-dimensional

 solution was seen as optimal given the relatively low number of key terms employed (Stress for two dimensions

 for the total data set =.263 R² = .726). In the shared matrix Africa is the most important term for the classification

of the dimensions (R² = .822). Mean importance for all participants was.420 for dimension 1 and .306 for dimension

2. Matrix data for the complete data set and for each country separately is provided in Appendix 5.

Dimension 1 of the overall solution we termed "Immoral, deadly sex". Homosexuals (loading = 1.16), condoms

(1.14) sex (.92), casual sex (.91) and prostitution (.88) can be contrasted with Africa (-2.00), death (- 1.11) and

disease (- .82). Dimension 2 we labelled the ‘Risk’ dimension and contrasts Drugs (-1.54), blood (-1.41) and fear

 (-.94) with Africa (1.40), sex (1.20) and condoms (1.10). At the cultural-level, the highest country loading on

dimension 1 was Poland (.69) followed by Hungary (.67) Estonia (.65) Georgia (.63) and Russia (.57). Highest

on dimension 2 was Estonia (.57) followed by Georgia .(55); Russia (.54); Hungary (.54) and Poland (.52).

We then examined the relationship between agreement with these dimensions and knowledge about HIV and

sexual behaviour. We therefore correlated knowledge and sexual behaviour scores with the dimension scores

and created ‘flattened subject weights’ which can be treated as ‘regular’ independent variables (Doise et al,

1993). Those who scored high on dimension 1 also demonstrated higher levels of knowledge about

transmission of HIV (r (465) = .10 p< .03) and had higher rates of optimism about life events (r (502) = -.10 p< .03).

 They were less likely to have had an STD (t (475) = -3.03 p< .003), multiple partnership (t (490) = 2.51 p< .01)

or multiple partners with low condom use (t (458) = 2.56 p< .01). There were no significant relationships between

 dimension 2 scores and knowledge about HIV transmission or sexual behaviour. Flatweight scores were

 related to sexual behaviour in the same manner as dimension 1 scores: those with high flatweight scores

 were less likely to have had an STD (t (485) = -2.78 p< .01), and were less likely to be in the moderately

high-risk group (t (490) = 2.08 p< .04) or the highest risk group (multiple partnerships and low condom

use: t (458) = 2.19 p< .03).

3.5. Discussion of stage 2

This stage of the analysis provided us with the most detailed quantitative data concerning both the

values and social representations associated with HIV and the sexual behaviour and knowledge of our

sample. Whilst our respondents overall were relatively knowledgeable about the dangers of HIV transmission

 they had a strong tendency to exaggerate the risks associated with risk-free activities such as being

coughed or sneezed on by someone with AIDS, with this misunderstanding of risk particularly notable

amongst the nurses in our sample. These results were strongly differentiated across the participating

nations in our study with Estonians the most knowledgeable about HIV risks and Russians and Georgians

 the least (Goodwin, 2000; Goodwin, Kwiatkowska et al, 2000).

In terms of actual sexual behaviours, sexual activity was highest amongst our Russian respondents, with

very high rates of multiple partnerships amongst the Russian business people and doctors. In contrast,

our Polish sample reported very few multiple partnerships. Perhaps unsurprisingly, therefore, sexual

disease was more frequently reported by the Russian, Georgian and Estonian respondents, with sexual

 diseases most frequently reported by male respondents.

Our findings suggest that values are reliable and consistent indicators of actual sexual behaviour,

whilst social representations, at least when transformed into individual level indicators, might also

provide additional predictors of sexual behaviour. Our data showed that three hierarchically related,

yet clearly distinguishable, types of collectivism – previously evident in studies of Estonian, Russian,

and North American populations – were also evident in our sample. People scoring high on family-related

collectivism were more conservative in their sexual behavior—they were less likely to report multiple

partnerships both in the last one and six months; they were more likely to know the HIV status of their

partner and they reported having not previously had a sexual disease. At the same time, people who

appeared to be more vulnerable to contracting HIV were likely to score low on family-collectivism scale.

When examining Schwartz’s values, findings here were generally as we had hypothesised, with the one

apparently counter-intuitive finding that using a condom was more likely amongst those more ‘open to

change’. This may be because less conservative, high sensation seekers are also more likely to carry

condoms, possibly because they are more likely to engage in sexual intercourse overall (Donohew et al,

2000).

One enduring problem is the way in which individuals learn the particular images and representations

that they associate with sexual disease. In the next stage of our study we turned to consider the manner

 in which different media images in different countries might be related to the levels of knowledge of HIV

 and images of sexual disease revealed in the first two stages of this work.

Section 4: Media analysis

4.1. Introduction to the analysis

For this stage of the research, each of the research participant team were asked to analyse the most

 frequently read newspapers purchased by each of our two respondents groups for a minimum of

two months. Most of our researchers exceeded these minimal demands, providing a comprehensive

data bank of newspapers for this period.

In Russia, eight newspapers were analysed for the period October 1st - December 1^st 1999. This sample

 comprised three newspapers from St Petersburg (two daily and one thrice weekly); one daily Moscow

paper; three Russian national papers (two daily and one weekly) and one bi-weekly medical paper. 383

articles were analysed in total from Kommersant; Komsomolsky Pravda; St Petersbursky Verdomosti;

‘Smena’; Dalavoy Petersburg; Medicinska Gazeta; Argumenti i Facti and Moskovksy Komsomolets.

In Poland 338 articles were examined between the period of 4th July and 17th December 1999 taken from

 two daily newspapers and two weekly papers. In Hungary, two daily newspapers were analysed

between 1^st October and 30^th November 1999 (102 articles). In Georgia, two daily newspapers were

analysed between September 23^rd and December 31^st 1999, a total of 169 articles. Finally, in Estonia

four newspapers were analysed between September 1^st and October 31^st 1999 (one weekly three daily),

although one, Äripäev (Business Daily) did not mention AIDS at all during the analytic period (a total

of 113 articles were examined).

4.2. Hypotheses

On the basis of our previous findings in stages 1 and 2 of the research we formulated a number of

broad hypotheses to be examined in our media analysis. First, we suggested that the higher the levels

of infection within a society, the greater the numbers of articles on HIV/AIDS in the media of that nation

 (hypothesis a). We secondly hypothesised, on the basis of our findings in stage 1 of the research, that

AIDS would most likely to be seen as originating in "the West" in the Polish and Russian newspapers

 (hypothesis b). We suggested that those topics that were most frequently mentioned in the media

should reflect the main source of infection in each country (hypothesis c). In Poland, our free

associations data suggested that Africa, casual sex, drugs and prostitution should be mentioned more

frequently (hypothesis d) whilst in Georgia the misfortune of AIDS sufferers should receive greater

discussion (hypothesis e). Finally, we anticipated that, following on from our first data collection,

there would be a greater mention of ‘miracle’ cures in the Russian newspaper media than in the other

 nations surveyed (hypothesis f).

4.3. Evaluation of the hypotheses

A full tabulation of our analyses is provided in table 8. Overall, we obtained mixed support for these

hypotheses. We first hypothesised (hypothesis a) that the higher the levels of infection in a society,

the greater the coverage of HIV/AIDS in the media of that country. As anticipated, there was considerably

 greater coverage of HIV/AIDS in the country with the greatest infection – Russia (29% of those articles

examined discussed AIDS). Perhaps more significantly, coverage was primarily in the form of feature articles

 in Russia (69% of the articles were in this format), a far more detailed coverage than in any of the other

four nations. The use of images to illustrate stories associated with HIV/AIDS was evident in approximately

 a third of the stories in Estonia, Poland and Russia but was not evident in Georgia or Hungary.

We then examined the extent to which AIDS was most closely linked to the West across the five countries

(hypothesis b). Thirteen of the 112 Russian articles that mentioned HIV linked AIDS to the West, but

incidence was very low for the other countries. As noted in section 1.2 (above) infection is most strongly

related to drug use in Russia, whereas infection is most prominent amongst homosexuals in Hungary and

prostitutes in Poland. In our media analysis, casual sex was cited as a source of infection in two of the

sixteen Polish articles that mentioned HIV (as predicted in hypotheses c and d), whereas homosexuality

was more likely to be discussed in Estonia (although this was in the context of a review of plays about

HIV in the USA). Drugs were discussed as a major source of infection in Russia, but were also discussed

in Estonia and Georgia, where the risk of HIV transmission through drug infection is still low but is likely

 to be rising. Prostitution was not widely discussed in the Polish media and was proportionally mentioned

more in Hungary than in any other culture (although the numbers here are very small across the sample).

AIDS was cited as an ‘African’ problem most frequently in the Polish media, as anticipated (19% of articles

 mentioned this, supporting hypothesis d).

On the basis of stage 1 of our research we also anticipated that there would more likely to be a compassionate

 discussion of the plight of AIDS sufferers in Georgia more than in any other nation (hypothesis e).

Compassion, however, was most evident in the Polish and Estonian media. Finally, 11% of Russian articles,

and 12% of Polish articles, discussed the possibility of a cure for HIV/AIDS, at least partially supporting

 hypothesis f. The nature of these Russian discussions are considered in more detail below.

detail below.

4. Individual country patterns

Each team member in each country then produced a brief report concerning issues of interest in the

media portrayal of HIV in their own country. These brief reports are summarised below and in

Appendices 4A-4D. More details are available from the researchers in each of the five nations.

Georgia. In Georgia, the newspaper Alia was most preferred by the medics whilst Resonanzi was

most frequently read by our business respondents. Overall, in the context of a difficult economic

situation and continuing conflict with Abhazia, AIDS was not a prominent topic in the Georgian

media. Nevertheless, there was a large difference in the coverage of AIDS between the two papers.

 In Alia, only four articles mentioned HIV. In all cases this was in the entertainment sections or in

‘passing’ stories concerned with other events (for example the Russian controversial politician

 Zihrinosky mentioned that AIDS was a "Western problem"). Resonanzi presented far more detail,

especially after the parliamentary campaign. The most interesting section was ‘readers letters’,

where a whole series of articles discussed HIV issues and sexual education. An anonymous free

hotline number allowed for an open discussion of sexual disease and the paper co-operates with

an NGO Aids Centre financed by Medicin san Frontiers. A brief report prepared by Dr Nizharadze

is presented in Appendix 4A.

Russia. In Russia, the most frequently read newspaper was Argumenti Facti, with 34% of respondents

in both groups reading this. This is a weekly newspaper, and in two months eleven articles mentioned

 HIV/AIDS. In six of these AIDS was the main topic. In three of these articles figures about the

spread of HIV were contrasted with those of Hepatitis C, with some suggestion that the latter is a more

 serious threat to the country as a whole (one journalist claimed that although AIDS was the plague of

the 20th century Hepatitis C was the plague of the 21st). Overall there was no discussion of sexual

education in this paper, and compared to other papers HIV was not seen as a big problem.

The second most popular newspaper amongst both Russian groups was Moskovsci Komsomolets, a

daily newspaper read by 22% of the businessmen and 15% of the medics. In the two months in which

this newspaper was purchased nine articles concerned AIDS, although in three of these AIDS is only

mentioned in passing. This paper also does not mention sexual education or prevention programmes,

but prefers instead to present a rather pessimistic portrayal of the AIDS epidemic. In one particularly

depressing article, Russia is described as getting closer to the "end of the world", with AIDS one

problem amongst many facing this country.

The third paper analysed, Kommersant, is a daily newspaper read only by the business people in our

sample (22% of our respondents). Over the two months of the analysis eleven articles mentioned AIDS,

although in five of these the mention was ‘in passing’. There are no articles about safe sex or education,

but instead a cynical appraisal of the position of the Russian Government, which the writers claim is not

 interested in helping in this situation. Gays are mentioned in five articles as a high-risk group that spread

 AIDS but drugs are only mentioned once. One short article discusses a man with HIV who infected 60

women whilst another has a photo of an ill African man. In this context two "specialists" claim that Russia

could soon be "like Africa".

The fourth paper read was Delavoy Petersburg (read by 12% businessmen but none of the medics).

This is a thrice weekly business paper. This newspaper has a special health page but only in one case was

 AIDS mentioned (in 25 editions), in an article about medical injuries. The fifth paper, Komsomolsky Pravda,

 is a youth newspaper read by 11% of the medics in our sample (only 2% of the businessmen). 31 articles

 mention AIDS, although twelve of them only cite AIDS in passing. Particularly interesting here are the

 eight articles concerning a particular patient treated by Armenicum, a treatment financed by this

 newspaper. This story lasted a total of nine months (from February – October 1999) and a total of about

30 articles appeared about this patient during this period in this newspaper. Following early positive

articles about the results of this treatment it soon became apparent that the HIV situation was deteriorating

for this patient, but there was still a considerable amount of optimism expressed by this newspaper, and

the patient was encouraged to repeat his treatment. Other articles concerned drug addicts (seven articles)

with two large articles about prostitution and two about safe sex and condoms. Overall this paper presents

a rather optimistic view, with some expectation for the development of a lasting treatment for HI/AIDS.

The sixth paper analysed was St Petersburski Verdomosti, read by 4% businessmen and 8% of the medics.

Eleven articles mentioned AIDS during this period, with three of them only mentioning AIDS in passing.

The main topic here is drug addiction and AIDS. One interesting view expressed is that St Petersburg is

safer than Moscow for HIV infection, but there is a general recognition that all are at risks of AIDS infection.

Smerna was read by 6% medics and no businessmen, with fourteen articles mentioning AIDS. This is

predominantly a youth newspaper, and more than half of the articles connected AIDS and drug addiction.

This was the only newspaper actually giving information about places for testing in St Petersburg. Contrary

to the other St Petersburg paper discussed above, there is far less optimism about the situation in this city.

The last paper analysed in Russia was the professional medical paper Medicinska Gazeta. This is a twice

weekly newspaper read by 23% of the medics but none of the business people. 22 articles in this period

mentioned AIDS in 17 issues. Half of the articles scientific articles were about how AIDS spread from

monkeys and new theories about the origin of AIDS. Only this paper specifically noted that all regions of

Russia have AIDS and talks about new medicines developed in Russia (although it gave a rather negative

interpretation of the Armenicum phenomenon). On 1st December (‘AIDS day’) there were also pessimistic

articles about AIDS and children, although the newspaper was relatively neutral in its coverage of the AIDS

 epidemic overall.

In sum, our detailed analysis of this range of newspapers in Russia allowed us to demonstrate the wide diversity

 of coverage of HIV/AIDS in Russia (the full materials for this analysis are lengthy and are available from the

project co-ordinator of Ms Kozlova). In particular, it became evident that ‘vested interests’ (in, for example,

promoting an individual drug treatment) were likely to be important contributors to the debate about

HIV/AIDS in this country. Remarkably, given the wide coverage of HIV/AIDS in this media as a whole, there

is very little direct information given about the necessary methods needed to avoid infection or the places

available for AIDS testing.

Estonia: Here an analysis of three newspapers produced 21 articles in total during the period of September

and October 1999. Coverage was even across the period, with an average of 2.3 articles per week. Most

articles were written by journalists, with frequent reference to the AIDS situation overseas (for example

two images from the Estonian daily paper showed Magic Johnson, and seven articles were ‘cultural articles’

discussing plays, books and CDs referring to HIV). The Estonian Express discussed the ‘post Estonian

AIDS generation’, a generation knowledgeable and open about HIV and safe sex. Overall, AIDS in Estonia

was portrayed as "just one of those things that happen and something we have to live with". A more detailed

 report on this data prepared by Andu Rämmer and Ahto Külvet,, who assisted Dr Realo in this part of the

project, is provided in Appendix 4B.

Hungary: Magyar Nepszabadsag is the most popular newspaper in Hungary, claiming to have two million

readers. Nine articles mentioned HIV, with four just mentioning HIV in passing. None of these nine articles

directly deal with the Hungarian situation, but focus on problems elsewhere. The other paper examined

(Magyar Herlap) is more concerned with politics, and the four articles that discuss HIV do not discuss the

 Hungarian situation, but focus on foreign countries and the research being conducted there. Overall, HIV

in Hungary is not a big issue, with the matter discussed in a generally impersonal and neutral way. A more

detailed report is provided in Appendix 4C.

Poland. Two daily papers were analysed in Poland, one a local paper, the other a more widely read paper

highly popular and influential across the country. In addition, two weekly magazines were examined, with

analysis conducted from July – December 1999. A more detailed summary report prepared by Dr

Kwiatkowska is provided in Appendix 4D.

One particularly notable trend in Poland is for newspaper articles to be written by respected professionals.

Thus two articles were written by professors (one of psychiatry the other of pharmacy/medicine). These

professors created an air of remoteness and their views can be controversial: in the article by the professor

of psychiatry, in the Science page of a leading newspaper, HIV infection is mentioned as emerging from a

chain of events: young people smoke marijuana, lose control, have multiple partners and unprotected sex.

Much of the emphasis was on the threat from ‘outside’ rather than dangers within Poland: for example one

article discusses the ‘messengers of Death’ from Africa whilst others focus on the dangers from the Former

Soviet Union, implying that foreigners are the main source of danger. Where problems such as drug addiction

 are mentioned they are presented in a very vague way. However, those infected are not always portrayed as

belonging to risky groups but are instead sympathetically portrayed, as innocent victims. Here it is more

likely to be the infecting partners who are described as ‘degenerates’.

4.5. Discussion of stage 3

The media analysis provides us with important insights into the complexity of HIV coverage across these

Central and Eastern European nations. Rather than furnishing readers with a uniform story we end up with

 a series of approaches designed not only to target particular readerships (compare, for example, the

coverage of HIV/AIDS in the medical and business newspapers within Russia) but also, and perhaps

 more significantly, to further the interests of particular newspaper groups. This was most evident in the

 sponsorship of treatment through the AIDS-drug Armenicum by Komsomolsky Pravda. The optimism

demonstrated by this youth newspaper might have important, and rather deleterious, implications for the

 perceptions of HIV and risky sexual behaviours amongst an increasingly high-risk group, a possibility

we discuss in our final data collection in Russia (below).

Overall, discussions of HIV/AIDS were most prominent in the country where this epidemic is having the

 greatest impact – Russia. In contrast, in the other countries in this region, HIV was no longer seen as a

‘hot topic’, and took its place alongside other lifestyle risks and health and economic concerns.

Coverage in these countries was also often very much ‘in passing’, and framed with reference to other

events and personalities outside of the country of publication (for example, the coverage of ‘Magic

Johnson’ in the Estonian press). From a positive perspective, the relatively ‘neutral’ tone of the coverage

 of the epidemic was encouraging – there was little evidence of direct condemnation of life-styles or

particular groups, and although casual sexual activity was discussed as a factor underlying HIV most

frequently in the Polish media there was also considerable compassion expressed towards those who

were infected in this country. However, on a more negative note, there was very little direct information

in any of the newspapers about where anonymous testing could be conducted or even on the manner

of transmission of HIV/AIDS, and it was only in the comparatively ‘low risk’ nation of Georgia that

details were provided of ‘hotline’ numbers for those seeking further information about the epidemic.

Clearly, therefore, the newspapers in these countries saw the propagation of information about HIV/AIDS

as only a very minor part of their reportage (Nizharadze et al, 2000).

Section 5: Stage 4: Local analyses

In this final stage of our work we utilised the particular research interests, methods and skills of the

investigators in each country to conduct a series of small ‘satellite’ studies. These studies were specifically

 aimed at exploring particular facets of HIV/AIDS, and the wider societal environment surrounding the epidemic.

5.1. Streetchildren in Russia

In our first ‘satellite’ analysis, Ms Kozlova analysed 100 adolescents, fifty from state schools (twenty-one

boys) and fifty from shelters (twenty-six boys). The average age of the schoolchildren was 15.5 years

 (ranging from 15-16 years) and 14 years for the shelter children (ranging from 13-17 years). Ten of the

shelter children were injecting heroin at the time of this analysis and attended a rehabilitation centre for

drug addicts.

Ms Kozlova first asked all the children five questions derived from stage 1 of our analyses: What do you

 know about AIDS/what is AIDS? How did you hear about AIDS? How does AIDS spread amongst people?

 What kinds of person gets AIDS? and What can people do in order not to catch AIDS? All were interviewed

 alone in their rooms in the shelter or in an empty classroom at the school, and Ms Kozlova took notes during

 each interview. Responses were categorised by Ms Kozlova, with respondents permitted to give multiple

answers. Full responses are presented in table 9A.

Children gave a mixture of answers to the first question. 19% boys and 13% of girls in the shelters thought

that AIDS can now be treated and that appropriate medical cures have now been found. 50% of all children

from both samples knew AIDS sufferers die. Four schoolchildren knew little about AIDS and wanted to know

 more, and three girls from 24 in the shelter admitted they knew little about the epidemic. Rather worryingly,

one boy from a shelter said a friend of his had AIDS "but now he is OK".

In response to the question about where they learned about HIV/AIDS, the most popular answer was "from

TV programmes". Only six of the 100 children heard about AIDS in the school environment. Question 3

asked about the spread of AIDS. 70% of schoolchildren noted drugs as a route of infection but only 27%

of the shelter boys and 37% shelter girls mentioned drugs in this context. Nine children from shelters thought

 HIV could be transmitted through kisses, compared to only one schoolchild, and only shelter children

thought that shaking hands, going to hairdresser’s etc., a risk factor.

For the fourth question the children discussed the characteristics of high risk groups. 76% of schoolboys

and 30% of shelter boys claimed that drug addicts can get AIDS whilst the popularity of this response was

equal amongst the girls for the two samples. All children knew prostitutes were at risk but only three mentioned

the risks to their clients. Fourteen school children noted that those having casual sex can contract HIV/AIDS

 but only five from the shelter stated this.

When asked about means of AIDS prevention the most popular answer was to use condoms (noted by 85%

of school children and shelter boys but by only half of the shelter girls). 20% of all the respondents said it

was necessary to have one tested partner to be safe: only four children from shelters (but twelve from the

school sample) noted the necessity of avoiding drugs.

Ms Kozlova then gave the same children the knowledge questionnaire used in our stage two analysis (see

table 9B). Here there was evidence of a great deal of misinformation amongst our respondents. Seventeen

shelter children viewed eating somewhere where cook has AIDS as highly risky and twenty-two of them

thought kissing someone with AIDS a highly risky activity. Fifteen shelter children rated sharing plates as

 high risk, and six shelter girls thought using public toilets high risk. Fifteen shelter children thought

 being coughed or sneezed on was high risk and four that public saunas were also high risk. Only shelter

 respondents thought touching someone with AIDS was high risk (five children indicated this). These shelter

 children thus showed a strong tendency to exaggerate risk: whilst they knew what was risky they did not

know what was low risk. Although we need to allow for the fact that the shelter sample was slightly younger

 than the school sample, our findings from this small study are clear: overall, shelter children were more

ignorant than the school children about HIV and associated risky behaviours (t (98) = 3.74 p< .001; item

Ms 1.64 vs. 1.33).

5.2. Gender stereotypes and sexual behaviour in Poland

Dr Kwiatkowska (Kwiatkowska et al, 2000b) examined the relationship between gender stereotypes and

sexual behaviour and knowledge, using a gender stereotypes scale previously developed for work in

Poland (Kwiatkowska, 1999) and the sexual behaviour and knowledge items employed in the second

stage of our study. The 100 business and medical respondents who completed our stage two of the

research completed 60 items assessing popular views on Polish women and men. Respondents

indicated their responses on 9-point Likert scales (ranging form very much agree to very much disagree).

Following a factor analysis, five major stereotypes emerged from these responses. For female stereotypes,

 the concepts of "female as mothers" and "females as ladies" (elegant, respectable etc.) explained most

of the variance. The predominant male stereotypes were "father as authoritarian and head of family";

"male as selfish and egoistic" and "male as loser". Overall, a high score on all these stereotypes

 indicated a lower knowledge of AIDS and the means of preventing infection. Those who held a

strong belief in "females as ladies" were less likely to use condoms (x² (3) = 11.56 p< .009), whilst

those who emphasised the "female as mother" (women as caring and empathetic) were less likely

to see sharing needles as high risk (x² (3) = 9.77 p< .002). For the male stereotypes, those who saw

men as "losers" exaggerated the risks of harmless activities such as going to public baths and eating

where the cook has AIDS, whilst those who scored higher on the "father" stereotype scored low on

avoidance of AIDS (r (99) = -21. p<.04), indicating they were less knowledgeable about the steps

required to avoid infection. Further details of this data are available from Dr Kwiatkowska in Bialystok.

5.3. Hungary: gender stereotypes

Working in tandem with Dr Kwiatkowska Dr Nguyen-Luu examined the prevalence of particular

gender stereotypes amongst students and teachers in Hungary. Participants in her analysis were

138 Hungarians aged between 18-19 and 30-40: 72 secondary school teachers (46 females aged 30-41,

mean age 35.5, SD 3.09) and 58 students (43 females, aged 18-22, mean age 19.53, SD 1.01).

Participants were all students on psychological training programmes that took place at ELTE in

Budapest.

Dr Nguyen-Luu examined gender stereotypes from two perspectives: the ideal man and woman and

the participant’s own self concept as a man and a woman. For this, she used a list of 22 items that

came from pilot studies of how people think about men/women in Hungary (e.g. men as brave,

strong; females as gentle, submissive). Respondents rank ordered eleven male and eleven female

characteristics for both the ideal man and woman and for their relevance to their own self concept.

As hypothesised, Dr Nguyen-Luu found significant differences between the genders, with men

more likely to see their ideal woman as more feminine and traditional than the non-traditional traits

emphasised by women, whilst the ideal man was less differentiated across the sexes. In terms of

self-descriptions, men predominantly thought of themselves as more sensitive, emotional and kind

than the male stereotype. Women viewed themselves in a more complex way: on the one hand females

 saw themselves as less stereotypical than the typical image (they saw themselves as more ambitious,

brave and able to get their way) whilst they were keen to maintain a more "feminine" side that expressed

‘emotional richness’ (sensitivity, emotionality and loyalty). This strong sense of loyalty and sensitivity

shown across the sexes may be related to the strong emotional links sought by our respondents in their

sexual relationship, as well as the loyalty that emerged from the patterns of sexual behaviour observed

in our earlier data collection in Hungary. Needless to say further research is needed to explore this link

 between stereotypes and sexual behaviour in greater detail.

5.4. National stereotypes and HIV in Georgia and the changing nature of fatalism.

In our fourth ‘satellite study’ Dr Nizharadze examined the extent to which non-Georgians were

specifically blamed for the AIDS epidemic in Georgia. 40 medical and business respondents responded

to an item directly asking about blame for the HIV epidemic, but, contrary to our expectations, nationality

 was not mentioned as a factor in the spread of the epidemic - even amongst those sympathetic to nationalist

 groups in Georgia.

Levels of fatalism were also compared by Dr Nizharadze across sample periods using data from earlier

projects sponsored by the EU (INTAS) and collected from 96 managers (in 1995) and 38 managers

(in 1998). Our sample demonstrated a notable drop in fatalism over time: in 1995 managers scored an

average 5.92/10 on this scale but this dropped to 4.58 in 1998 and 3.99 in our data. Given that fatalists

in Georgia were more likely to have more sexual relationships this may have important implications for

future changes in sexual attitudes and behaviours amongst this group.

5.5. Personality and sexual behaviour (Estonia)

In our final ‘satellite’ study we aimed to study the relationship between sexual behaviour, personality

 and knowledge. The sample here was 127 Estonians (75 women) ranging in ages from 19-52 (mean age

 27.3; SD 6.7). Respondents included both undergraduate students (40% of the sample) graduate

students (9% of the sample) and open university students (51% of the sample). Many of the latter

were mature students working and attending classes part time.

For this study Dr Realo employed the Estonian version of the Norman Adjective Checklist (1963),

which consists of 20 bipolar adjectives measuring five major domains of personality: Neuroticism,

Extroversion, Openness to Experience, Conscientiousness and Agreeableness. She also included

items from the second stage of our study to assess sexual behaviour (number of partners over the

 past week/ one month/ six months), use of condoms (frequency of use over last six months) and

perceived vulnerability to contracting the HIV virus (five items, as described above). Finally she

 included items 2-11 of the knowledge scale described in stage two of our study and used by Ms

Kozlova (above).

First, Dr Realo examined the general trend in sexual partnerships across the sample. 13% of her

sample were not sexually active over the past month, and the majority of those sexually active

(62%) had had just one partner over the previous six months. Only one person reported more

than one sexual partner in the previous week. Personality traits failed to show a strong correlation

 with sexual behaviour, although there was a correlation between extraversion and number of

partners over the last month (r = .21 p< .05: comparison of high versus low risk groups, as defined

in section 2 above, t (117) = -3.9 p< .001 Ms 3.1 vs. 7.2). Finally. Dr Realo compared knowledge

scores amongst these Estonian students with those of the medics and business people in our

main data collection. There was no overall statistical difference here between the groups, at least

partially supporting the external validity of our previous data collection in this country.

Section 6: Overall Conclusions

6.1. Summary

A multi-stage project was conducted to examine the impact of values and social representations

of HIV/AIDS on sexual behaviour in five Central and Eastern countries. Participants completed

interviews and free associations, questionnaires and matrix tasks whilst a media analysis and

satellite studies aimed at further exploring the cultural environment in which our findings could

be located. Our results indicated significant cultural and group differences in representations of

HIV and those seen as at-risk from infection. Values and beliefs were significant predictors of

reported sexual behaviours, although knowledge of risk factors was a poor predictor of such

behaviours. Media analyses suggested that HIV/AIDS was not seen as a major issue for discussion

in any nation in our study but Russia, where coverage was influenced by the nature of the

newspaper as well as the particular interests of a newspaper in promoting a drug treatment. Finally,

the satellite studies indicated important directions for future study, demonstrating an intriguing

relationship between gender stereotypes and sexual behaviour as well as important misperceptions

about HIV/AIDS amongst a highly vulnerable adolescent shelter