The aim of the WMS 2011 was to interview the same ‘well-informed respondent’- in each household who had participated in the 2009 survey. The longitudinal dataset enables analysis of changes in household and personal circumstances over a two year period. Fieldwork began on 20 of June 2011 and finished on 30 of July 2011. It was carried out by 84 interviewers, with regional supervisors, all across Georgia.

Sampling

The sampling strategy targeted the 4808 households in which face-to-face interviews had been completed in 2009. Successful interviews were held with respondents from 4147 households, an 86 per cent response rate. 

Of the 659 non-responses, 205 were due to refusals to participate but in the majority of cases respondents had moved away, either temporarily or permanently.

Data weighting

A household weighting variable was provided by the survey research company for use with the complete set of 4147 households in the sample. This was based on the weights used in the 2009 survey, as stratified by region and type of location. Table 2.2 shows that the geographical distribution of the population in Georgia changed little between 2009 and 2011 and the new weight simply adjusts the 2011 households to correct for the change in sample size. Separately, individual weights are applied to questions addressing individual household members. 

Table 2.2: Distribution of population by region at the beginning of 2009 and 2011

Source: http://www.geostat.ge/

Comparability with previous study

The achieved survey sample in 2011 was 4147 households. Examination of the 2009 survey data shows that 4020 households completed a full questionnaire in both survey years (Table 2.3).

Table 2.3: Survey attrition 2009 - 2011

The 127 households included in 2011 but not in 2009 are ones where the 2009 questionnaire was not completed fully. There was no information provided on consumption essential to poverty measurement. Of the 4646 households where we do have consumption data for 2009, the households that went on to participate in 2011 were not significantly different in terms of their 2009 material deprivation, subjective poverty or social exclusion than those who did not. Therefore, no bias occurs from the dropout rate when using only complete surveys in both years.

However, households that went on to participate in 2011 wave had on average lower consumption levels (PAE) and a higher consumption poverty rate than those who dropped out. There is a significant association between consumption poverty and location (rural and urban) so the effect of non-response on consumption levels is likely to be mitigated by the over-representation of urban households lost between the two waves of the survey.

Since the original survey weighting is based on settlement size and altitude within regions, the re-weighting of the 2011 data only adjusts for sample size to reduce any possible bias that may have resulted from attrition between 2009 and 2011. The weighted 2009 sample, for example, includes 50.8 per cent of urban households while the re-weighted 2011 figure is 50.6 per cent.

For most of this report, comparisons are based on the 4646 complete responses for 2009 and the 4147 responses from 2011, weighted appropriately. Where indicated in the report, the analyses tracing the trajectories of individual households through time use only the 4020 households for which a full questionnaire was completed in 2009 and in the 2011 survey. A separate weight was provided by ISSA to account for the number of households in this subset of the sample.

Adjusting for inflation

A sustained increase in the general price level is measured by the consumer price index (CPI). It is based on the cost of a typical basket of consumer goods and services in particular years. The National Statistics Office of Georgia on the website of the National Bank of Georgia (http://www.nbg.ge/index.php?m=306#monetarystatistics) gives a CPI of 171.7 for July 2009 and 201.5 for August 2011 (using a base year of 2000), the months in which fieldwork was completed. For comparing change between the two surveys, monetary data for 2011 are converted to 2009 prices by dividing by 201.5 and multiplying by 171.7.

Income and expenditure per adult equivalent

As in the 2009 analysis, comparison between households of different size and composition use measures of income and expenditure. Some are adjusted, in the relevant parts of this report, to GEL per equivalent adult (PAE). The calculation is based on that of the Georgia Department of Statistics. First an equivalent adult coefficient is calculated for each household (Table 2.4).

Table 2.4: The scale used to calculate number of equivalent adults in a household

To correct for economies of scale in larger households, the number of equivalent adults is then raised to the power , where =1 for a single person household and =0.8 where household size is greater than one.

Statistical significance

In sample surveys we do not have information about the whole population. We infer things about the population from statistical tests carried out on the sample. The probability that the relationships we observe in the sample only occur by chance (rather than from real effects within the whole population) is represented as p. The smaller the value of p, the more confident we can be that the sample relationships do hold true in the population. Probabilities of less than 0.05 mean we can be at least 95% confident and this level is conventionally used to indicate statistical significance.

In the tables that follow the conventions used to represent probability values are:

* means p < 0.05 (95% confidence)

** means p < 0.01 (99% confidence)

*** means p < 0.001 (99.9% confidence)

ns means Not statistically significant (less than 95% confidence)