Monday, July 27, 2020

Georgia has a vaccine misinformation problem

[Note: This article was co-published with OC Media. The article was written by Dustin Gilbreath, Deputy Research Director at CRRC Georgia. The views presented in the article are the views of the author alone and do not represent the views of the Embassy of the Netherlands, CRRC Georgia, or any related entity.]

While the majority of Georgians believe that vaccines are a net positive for society, a majority also express skepticism about their safety and effectiveness with only 42% interested in receiving an COVID-19 vaccine if it became available. 

Many experts believe that to fully remove the restrictions which have emerged because of the COVID-19 crisis, a vaccine is needed. While vaccines are only expected in the medium term, if and when they are available, Georgia may face large challenges with implementing a large scale vaccination program. 

Rather than money or logistics being the primary barriers to vaccination, misinformation might be. In other contexts, anti-vaccine sentiment has led to the re-emergence of diseases that had long been under control. The newly released COVID-19 Monitor data, which CRRC Georgia collected with the support of the Embassy of the Netherlands in Tbilisi, suggests that large shares of the public are misinformed about vaccines. Furthermore, the more negative or uncertain people’s attitudes are towards vaccines, the lower is their chances of wanting to be vaccinated if a COVID-19 vaccine was available.

The survey asked respondents a set of 11 questions about vaccines ranging from whether vaccines cause autism to if vaccines are effective at preventing the diseases they are supposed to. The results suggest there are high levels of uncertainty and misinformation about vaccines in Georgia. 

One in five people (19%) agree with the statement that vaccines cause autism.  A further 52% are uncertain.

One in five people (21%) believe that infant immune systems cannot handle as many vaccines as doctors give them. Another third (35%) are uncertain.

One in five (21%) also believe that if they vaccinate their child, it may create serious problems and a quarter (24%) are uncertain.

Although less than half the public believe these factually inaccurate statements, the shares are relatively high. For example, in the United States, anti-vaccination sentiment is considered both a public health and security risk. Yet, in the US, half as many people (10%) believe that vaccines cause autism and a slightly lower share (46%) were uncertain, according to a January 2020 Gallup survey.

Public sentiment is not entirely negative. Most people (74%) think that vaccines are necessary to protect the health of young people and that they do a good job at preventing the diseases they are intended to prevent (72%). 

However, most people express at least some scepticism or uncertainty towards vaccines. The chart below presents an index of attitudes towards vaccines. Respondents were given 1 point if they reported a pro-vaccine attitude and 0 points if they expressed either uncertainty or a negative attitude towards vaccines. Roughly equal shares of the public have attitudes that tend to be more positive than negative/uncertain and more negative/uncertain than positive. 

Attitudes towards vaccines are reflected in people’s interest in getting a vaccine. If a vaccine was available six months from now 42% would be interested in getting it, 43% would not want the vaccine, and the remainder were either uncertain or refused to answer the question. 

Those that did not want to get the vaccine reported they would not want the vaccine most frequently, because it would not be tested thoroughly enough (40%). 

However, data collected a week later suggest that similar shares would want (38%) and not want (43%) the vaccine if it was available two years from now rather than six months, when presumably the vaccine would be better tested. 

Aside from the lack of testing, scepticism towards vaccines in a variety of forms was also frequently mentioned among those that did not want to get a vaccine. One in seven (14%) reported vaccines cause larger health problems for those who take them, and one in nine (11%) reported that vaccines are not effective at treating disease.

Note: The data on the above chart do not sum to 100 as respondents were allowed to name more than one response.

There is a strong correlation between people’s attitudes towards vaccines and whether or not they would want to get a vaccine if one was available six months from now. The chart below shows the adjusted probability of wanting a vaccine if one were available by the attitude index shown above. Controlling for age, educational attainment, settlement type, and whether or not there were children in the respondent’s household, the results suggest that people who have entirely uncertain or negative attitudes have a 10% chance of wanting a vaccine. By comparison, a person with fully positive attitudes has an 87% chance of wanting a vaccine.

Controlling for attitudes towards vaccines, a number of other factors are associated with whether or not someone would want a vaccine if one were available. Women are 25 percentage points less likely than men to want a vaccine, all else equal. People in Tbilisi are 11 percentage points less likely to want a vaccine than people in other urban areas and 13 percentage points less likely than people in rural areas.

The above data clearly shows that Georgia has a vaccine misinformation problem. This matters for both public health in general as well as for the eventual defeat of COVID-19.

Friday, July 24, 2020

Covid 19 and the Georgian Orthodox Church

[Note: This article was written by Dustin Gilbreath, Deputy Research Director at CRRC Georgia and published originally on The views presented in the article are the views of the author alone and do not represent the views of the Embassy of the Netherlands, CRRC Georgia, or any related entity.]

One of the most controversial points in the Covid 19 crisis to date has been the Georgian Orthodox Church’s response. Of particular contention were the Church’s refusal to sanitize the communion spoon or to use a replacement and the decision to leave churches open for the Easter liturgy. In an apparent attempt to discourage church attendance, the government banned car use unexpectedly on April 17th in the days before Easter, with the ban continuing until April 27th. 

While there was much controversy over the Church during the crisis, how many people actually attended church during Easter and what did the public think of the Church’s response?

The results of the newly released Covid 19 Monitor survey, which CRRC Georgia carried out with the support of the Embassy of the Netherlands in Tbilisi, suggest that church attendance was less than a tenth of its past. Further, the public tended toward disapproving the communal spoon policy. At the same time, data on people’s views of how the church handled the crisis are ambiguous.

The study asked respondents whether they attended Easter Liturgy this year as well as last. 
Only 4% of Orthodox Christians reported attending church on Easter this year. This compares to 44% who reported that they went to church on Easter last year. Comparing the two (4% this year /44% last year) suggests attendance at Easter liturgy was 9% the year prior. This is likely a partial explanation for why Georgia did not experience a spike in Covid 19 cases following Easter.

Further analysis of the data suggests that younger people were a bit more likely to go to church (7% of 18-34 year olds and 5% of 35-54 year olds) than older people (1% people over 55).

When it comes to approval of the use of a communal spoon, 33% of Orthodox Christians approved of the communal spoon policy, 43% disapproved, and 21% were uncertain. Older people (55+) are more disapproving of the policy (55+: 51%) than younger people (18-34 and 35-54: 39%). 

While relatively few people attended church this year and the public tended toward disapproving of the communal spoon policy, has this impacted trust in the church or the public’s perceptions of how well the church is performing? On these points, the data does not provide a clear answer.

On the 2019 Caucasus Barometer survey, 71% of Georgian Orthodox Christians reported trusting the religious institution they belong to. Similarly, 69% of Georgia’s Orthodox population reported the same on the Covid 19 Monitor survey. However, only 28% reported fully trusting the church in the Covid 19 Monitor survey compared with 35% in the Caucasus Barometer survey. This suggests that the degree of trust in the church has declined. 

The Covid 19 Monitor Survey asked about the Church’s performance using the same question asked in the November/December 2019 NDI survey. The results suggest that since December, there has been an increase in positive assessments of the Church’s performance from 50% to 66%. However, this is in a context where performance assessments increased for all institutions that were asked about on both surveys.

The public tended towards not supporting the Church’s communal spoon policy, and most people who went to Church for Easter in the past did not this year. The Church’s approval ratings remain high, and have even increased since the start of the crisis. Yet, this follows a broader pattern of increased institutional performance assessments in the country more broadly, and the degree of trust that Orthodox Christians have in the church appears to have declined.

Thursday, July 23, 2020

Covid-19 Monitor: New Report on Georgian Public Opinion on the Covid-19 Crisis

Today, CRRC Georgia released a report on public opinion in Georgia on the Covid-19 crisis. The report includes data collected between late April and early June, 2020. During this time, CRRC Georgia conducted weekly public opinion surveys and presented the results on a weekly basis to policy makers, local NGOs and the international community. 

The study covered a wide range of issues from disinformation to food insecurity and employment. Some of the key findings include:
  • The public’s assessments of the performance of a wide range of institutions became significantly more positive during this period;
  • The vast majority of the public approved of the nearly all of the policies that the government implemented during the crisis;
  • Between a third and a quarter of the population lost a job during the crisis;
  • The majority of households experienced at least some level of food insecurity during the crisis;
  • Median household incomes roughly halved during the crisis;
  • People tended towards favoring opening up the economy during the crisis to exercising caution;
  • There is widespread misinformation around vaccines in Georgia, and belief in misinformation is strongly correlated with whether or not someone would want a vaccine for Covid-19 if one were available six months from now;
  • Misinformation was present but not widespread during the crisis. For example, 9% of the public believe that 5G infrastructure spreads the virus;
  • Only 4% of Georgian Orthodox Christians attended Easter Liturgy, as opposed to 44% the year prior;
  • Few people believe that there will be a second wave of the virus, and there would be less support for implementing restrictions on different freedoms if a second wave took place.
The full report is available here. The six datasets are available from CRRC Georgia’s online data analysis tool.

Tuesday, July 21, 2020

Teachers can be encouraged to report domestic violence — but the authorities must respond

[Note: This article first appeared on OC Media, here. It was written by Dustin Gilbreath. Dustin is the Deputy Research Director at CRRC Georgia. The views presented in this article reflect the views of the authors alone and do not reflect the official positions of UN Women, the Danish Government, CRRC Georgia, or any related entity.]

Domestic violence was widely suspected to have increased during the COVID-19 crisis. 

A study CRRC Georgia conducted for UN Women prior to the crisis found a behavioural lever that could encourage teachers to report domestic violence they suspect among their students. Yet, the report suggests that until the government reforms the currently dysfunctional reporting infrastructure, encouraging teachers to report could do more harm than good. 

This suggests that efforts should focus on using behaviourally informed interventions that nudge the authorities to respond to the reports they receive before rolling out large-scale efforts aimed at bystanders. 

The Teacher Reporting of Violence against Children and Women study aimed to understand whether there were behavioural levers that could encourage teachers to report domestic violence. 

Teachers were the target group of interest as they work closely with children. In this regard, it is generally accepted that the signs of domestic violence against both children and their mothers show up in children’s behaviour. 

Moreover, legislation passed in 2016 requires teachers to report suspected domestic violence in Georgia. According to the legislation, there is also a school reporting officer, who is often either a school resource officer (known as a mandaturi — school security guards) or the principal if the school does not have a resource officer.

The problems the authorities need to fix

The study suggests teachers are generally aware of their responsibility to report. Yet, they were hesitant to do so, because they think that the authorities’ and particularly the police’s response will lead to more harm than good.  As one teacher stated: 

‘I think about this frequently — which one is better, to report or not to report. Considering the last case in Kachreti, reporting sometimes results in such a catastrophic consequence… Those who should solve a problem, on the contrary, make it worse… And I was thinking what is better, to speed up such a catastrophe or stay indifferent?’

This fear was also reflected in the responses teachers gave on the study’s survey. When asked what might make their colleagues hesitant to report domestic violence, a fear that institutions would respond ineffectively was among the top three reasons given.

Aside from being afraid of the consequences of a report for the victim’s personal safety (or even the perpetrator’s, as the Kachreti case demonstrates), teachers were often also afraid for their and their family’s safety as the chart above shows. 

As one teacher stated: ‘[If you report] they call you a backstabber, because you collaborate with police […] I want to protect myself since all the violence will be redirected to me, and they will tell my children that your father called the police.’ 

The lack of confidentiality also feeds the fear that teachers will face physical reprisals as a result of making a report. As one teacher stated: ‘guaranteed confidentiality is not in place here.’ 

Even in the absence of such fears or where they are overcome, people who report domestic violence encounter difficulties in getting the police to respond adequately to the situation. As one school resource officer stated: ‘When we call the police, they say, “What happened? Who hasn’t fought?” We even sometimes have to beg them to come to our school.’

One of the sources of the problem is that some police officers do not take domestic violence seriously or even think it should be brushed under the rug. A school resource officer, when discussing an instance where they reported a case of domestic violence, stated, ‘They [the police] advised her [a child who was beaten by her father] not to make the complaint harsh and if she would change her complaint and write it in a “more beautiful” way, her daddy would go home in the evening.’ 

On top of the above, teachers realise that even if the police act effectively, most victims end up worse off as a result of reports. Families lose one or the only breadwinner, and teachers know the economic situation of the abused will deteriorate. 

Economic consequences aside, the abused also face social pressure, shame, and fear of the perpetrator’s return as a result of the report. 

The lack of socio-economic and psychological services, which are quite scattered and fragmented in Georgia, cause or exacerbate these issues. For teachers, the legal responsibility to report violence is outweighed by the moral responsibility to protect the safety of victims and not to add economic problems to the abuse and other issues they already face.

How teachers can be encouraged to report
The above shows that along the full chain of teacher domestic violence reporting, a circle of distrust stemming from ineffective institutional responses to domestic violence discourages domestic violence reporting among teachers. 

If teachers are encouraged to report domestic violence, it seems like they will be met with an ineffective response, discouraging them from reporting suspected domestic violence in the future. 

It is easy to imagine this devolving into a cycle wherein following failed responses, teachers discourage others from reporting through a demonstration effect.

In this context, there should be a reasonable degree of doubt about how much people should be encouraging teachers to report domestic violence. However, if the situation does improve, the study also provides some evidence about messages that could encourage teachers to report.

To test whether different messages might encourage teachers to take concrete steps towards reporting, the study randomly assigned three different messages to teachers and then measured their willingness to take a number of different actions surrounding domestic violence. The three messages included a social norming message, the provision of legal information, and both the previously mentioned messages combined. 

The social norming message highlighted to teachers that most people in Georgia find domestic violence unacceptable. While this may seem like a longshot at making change, previous research has shown the effectiveness of highlighting social norms that people are unaware of in changing behaviour. 

The second message informed teachers of their legal responsibilities to report. It was provided based on the assumption that many might not be aware of this duty. 

The third treatment combined both of these messages, with the goal of seeing whether the sum would be greater than the component parts.

The treatments were measured against attitudes towards reporting domestic violence, whether people were willing to provide their contact information to receive training about domestic violence, and whether they would be willing to sign a pledge against domestic violence. 

Although not direct measures of intention or actual reporting, the logic behind these measures is that a) attitudes relate to action, and b) in Georgia many are unwilling to provide their contact information (as this study re-affirmed). Hence, if the messages could encourage people to change their attitudes or take an action many would be hesitant to, then the message is likely on the right track.

The results of the experiment found few significant results, with one exception. The study asked teachers if they would be interested in participating in training on domestic-violence-related issues, and if they were, to provide contact information. 

Among those that were interested, teachers were 10 percentage points more likely to provide information so that they could be contacted for training if they received the social norming message. 

Aside from this one message having the potential to increase reporting, the study found that the different messages worked for different groups, enabling the targeting of different messages to people likely to be receptive to them. 

For instance, the study found that the legal information messages were effective with men but not women. 

While the study provides extensive detail on what messages are likely to work and for who, the main finding of the study is that before engaging in large-scale encouragement of teachers to report domestic violence, the government needs to adequately respond to the reports they already get. To do so, the study’s findings could be informative. 

Although the social norming message appears to have the most potential, the legal message appears to work with men in particular. Since the police force is largely male, using legally based messages with police officers may be particularly effective at encouraging adequate responses to domestic violence.  However, to confirm this suggestion, further research is needed.

The study has numerous findings and can potentially inform efforts at encouraging the authorities to respond appropriately to reports of domestic violence. It also provides a detailed set of recommendations on how to encourage teachers to report domestic violence. 

However, until the time the problems described in this article are fixed, it is questionable whether encouraging bystanders to report domestic violence will help or hurt.

Monday, July 13, 2020

Social capital in Georgia: how trust becomes solidified when words are backed up with deeds

Social capital is a set of networks between individuals and groups of individuals and the mutual trust related to these networks. It facilitates communication and cooperation between people and makes available resources that would be otherwise out of reach. Thus, social capital is crucial for social and economic development. Caucasus Barometer 2019 data shows that while the level of structural and cognitive social capital in Georgia is somewhat low, with the cognitive component lagging further behind, the bonds between the two are strong and stronger than each’s link to other factors.

While there are different definitions of social capital, most of them point to structural and cognitive elements of the phenomena. The former refers to networks and connections, and the latter to the feeling of trust and reliability among these connections. For example, networks refer to connections people have with people from similar as well as different groups, horizontal as well as vertical. It is relations and repetitive and reciprocal exchanges with others. The cognitive component of social capital points to the feeling of confidence in others and various institutions. Of course these two are interrelated, and they influence and cultivate each other

Structural social capital is measured using questions assessing people’s involvement in various activities and their ties with various groups of people. Caucasus Barometer survey asks a number of such questions, including whether someone:
When it comes to cognitive social capital, it is assessed based on how much people trust others and various institutions. CB contains a number of questions measuring cognitive social capital, including:
Variables measuring trust towards political institutions and specific actors are excluded when calculating cognitive social capital in this writing, as politically loaded attitudes are more likely to be driven by the attitudes towards specific political actors and shift based on ongoing events.

To describe Georgia’s population in terms of structural social capital, the answers to the above questions were recoded so that  respondents receive one point if their answers indicate the presence of structural social capital and 0 if not. Responses were then combined in an index of structural social capital that ranges from 0 to 11, with 0 indicating the lowest level of structural social capital and 11 the highest.

The mean score for structural social capital in Georgia is 4.54 and though it might seem that it is close to the theoretical average (5.5/11), ¾ of the population has structural social capital lower or equal to 6 and only 4% of the population scores higher than 8.

As for cognitive social capital, answers to the questions related to trust were transformed and combined in an index of cognitive social capital that ranges from 0 to 5, with 0 indicating the lowest level of cognitive social capital and 5 the highest.

The mean score for cognitive social capital in Georgia is 1.52 and ¾ of the population has cognitive social capital lower or equal to 2. Only 6% of the population scores higher than 3.

The above shows that more people have higher structural social capital, compared to cognitive social capital. This means that networks and links are more developed in Georgia than the trust component of social capital.

To understand how structural and cognitive social capital varies between different demographic groups, regression models were constructed. They included sex (male, female), age group (18-34, 35-54, 55+), settlement type (capital, urban, rural), education (secondary or lower, secondary technical, tertiary), employment status (employed, not employed), having debt (household has debt, household does not have debt), and an additive index of ownership of different items, a common proxy for wealth. The cognitive social capital index is also included in the model of structural social capital and vice versa. 

In theory all the variables included in the regression model are usually related to social capital. Specifically, people with more connections to various groups and opportunities for interaction have higher levels of both structural and cognitive social capital.

Regression analysis shows that structural social capital is higher in the capital than in other urban or rural areas. Younger people are also more likely to report higher structural social capital than people in their middle years or older people. People with tertiary education are also more likely to have higher levels of structural social capital. This is logical assuming that Tbilisi offers more opportunities to participate in diverse group activities and connect with others. Similarly, younger people might have more time and means of interacting and participating in various activities that form networks. Education may also open up even more opportunities to interact with different groups and participate in various activities. It is noteworthy that as theory would suggest, structural social capital is associated with cognitive social capital. People with higher levels of trust tend to have more connections with different people and groups.

Structural social capital is not associated with other factors. For example, people who are employed are expected to have just slightly higher structural social capital than people who are not employed. There is also no difference in terms of sex, a household having or not having debt, and wealth.

As for cognitive social capital, regression analysis suggest that there is no difference in cognitive social capital in terms of settlement type, sex, age, education, economic situation, employment status or households having or not having debt. The only variable that cognitive social capital is related to is structural social capital, as already shown above. This means that higher levels of trust are not really associated with being part of a certain demographic group or having a specific characteristic, but more closely tied to the amount of networks someone has. The more structural capital one has, the more likely one is to have higher cognitive social capital.

Regression analysis shows that structural social capital is related to various demographic characteristics though cognitive social capital is only predicted well by structural social capital. This relationship is logical and in line with previous research that indicates that these two are bound together. The fact that cognitive social capital is not really related to various demographic characteristics might suggest that structural social capital drives cognitive social capital.

Structural and cognitive social capital in Georgia is average or lower for the majority of the population. While various factors, such as age, education, settlement type, and cognitive social capital predict the level of structural social capital, cognitive social capital is mostly related to structural social capital. Nevertheless cognitive social capital is of great importance, as it simplifies communication and makes connections useful. Therefore, it might be useful to further study the ways in which structural social capital translates into cognitive social capital.

Monday, July 06, 2020

Georgians’ perceptions about equality at court

Attitudes toward the judicial system have been one of the most discussed and researched topics in Georgia. CRRC’s past blogs have shown that Georgians’ perceptions of court system fairness have been at low levels throughout the last decade and that attitudes toward court system (im)partiality are associated with rates of  trust toward the court system and people working in the court system. A recent CRRC study also highlighted division among the public regarding trust in judicial institutions. This blog post contributes to this conversation through describing views on the fairness of courts in Georgia, showing its broader inter-relations with trust in institutions, political views, and general perceptions of the government’s treatment of citizens.

The 2019 Caucasus Barometer survey shows that the majority of Georgians (63%) think that the court system is biased toward some citizens over others. However, the levels of agreeing with that statement vary across different demographic groups. A logistic regression suggests that people living in the capital, those with higher levels of education, and ethnic Georgians are more likely to think the courts favor some citizens, controlling for other factors. People in different age groups, women and men, the employed and those not working, those who use the internet more and less often, and those with more and fewer household assets do not differ in terms of evaluations of court impartiality. 

The belief that the courts are (im)partial is also associated with party support, trust in institutions, and people’s perceptions of whether the government treats people fairly. Controlling for the above social and demographic factors, Georgian Dream party supporters are less likely to agree with the statement that the court system in Georgia favors some citizens over others compared to people who support an opposition party. Those who do not report supporting any particular party fall somewhere in between. With institutional trust, controlling for other factors, lower levels of institutional trust are associated with higher levels of thinking that courts favor some citizens over others. People who think that people like them are not treated fairly by the government are also more likely to think that the Georgian court system treats citizens unequally, when all other factors are held constant.  

Note: The institutional trust index was created from the following variables: Trust in the Healthcare system; Banks; Educational system; Army; Court system; NGOs; Parliament; Executive government; President; Police; Political parties; Media; Local government; Religious institutions respondent belongs to; and the Ombudsman. A 1 represents the lowest level of trust, while a 5 represents the highest level of institutional trust. 

These correlations matter. People who support the opposition, trust institutions less, and think the government does not treat people fairly are all more likely to also think the courts are stacked against citizens. This suggests that people clearly view what should in theory be an impartial umpire as a political one in practice. 

The data presented in this blog post is available here. Replication code for the above analysis is available here.