Survey Of Barriers To Access To Social Services

Dates Collected: 10 Nov 2010 - 01 Dec 2010
Date Released: 01 Nov 2011

The Barriers to Access to Social Services Survey (BASS) was developed to gain a better understanding of why significant shares of poor households in Georgia are not accessing their entitlements. For this purpose, the respondents in the bottom consumption quintile from the 2009 survey were revisited and interviewed. The BASS is particularly important for addressing child poverty.

Highlights

Sample - The sample of the Barriers to Access to Social Services study was drawn from the database of the Welfare Monitoring Survey. For this purpose, the WMS households were divided into five quintiles, according to their welfare level. The first quintile represented the poorest 20% of the households of the sample. Accordingly, the fifth quintile contained the richest 20% of the sample. The ‘welfare level’ was defined as monthly expenses of households counted for one equivalent adult (15-65 years old male) and considering scale-effect. The first quintile included those households whose monthly expenses were less than 117.5 GEL counted for one equivalent adult in this household. Anticipating a possible non-response/refusal rate of 15%, 212 households from the second quintile were added to the sample. The total sample included 1,078 households. They were distributed across all regions of Georgia.

Survey tools – The questionnaire was designed on the basis of focus group discussions. In total, eight specific focus groups were convened: a) Beneficiaries of SSA; b) Applicants of TSA; c) Social workers and social agents; d) Principals and teachers of secondary schools and kindergartens; e) Representatives of civil registry agencies; f) Representatives of insurance companies; g) Representatives of the health care system; and h) Representatives of local authorities. The draft questionnaire was presented by ISSA. The advisory committee provided comments and suggestions. The questionnaire was field-tested with 20 households both in urban as well as rural settlements. After the pilot  questionnaire was finalized and interviewers were trained, the questionnaire was translated and published in three languages (Georgian, Armenian and Azeri).

Fieldwork - Throughout the study 10 interview teams were deployed, each one comprising of 1 supervisor and 1 to 6 interviewers depending on the size of the region. The duration of one interview was on average 1 hour. During the period 10 November – 1 December, 2010, in total 902 interviews were conducted across Georgia. This corresponds to 83.7 per cent response rate. 

Table A1: The distribution of completed interviews and non-responses by region

REGION

NUMBER OF COMPLETED INTERVIEWS

NUMBER OF NON RESPONSES

Tbilisi

111

12

Adjara

21

10

Imereti

191

32

Shida Kartli

57

9

Kvemo Kartli

137

38

Samtskhe-Javakheti

39

10

Samegrelo

81

16

Guria

64

20

Kakheti

127

14

Mtskheta-Mtianeti

57

14

Racha

17

1

Total

902

176

The main three reasons for non-response were: a) Households having changed the address; b) Death of the member of single households; c) Inaccessibility of adult family members (despite several visits by the interviewer).

Data Processing - The completed questionnaires were coded after the field work was completed.

The coded questionnaires were entered into SPSS (13 so ware). Data cleaning consisted of two steps. The first step entailed a frequency count for each variable, in order to check ‘wild codes’ in the data. During the second stage, the statistician checked the frequency listing for inconsistencies and identified problematic cases. The cleaned data was weighted according to weights defined in the framework of the WMS. Weighted data were processed on the basis of different statistical methods: distribution of frequencies, cross-tabulations, mean, mode, median, etc.

Data is presented by different categories- it is divided by Tbilisi, West Georgia and East Georgia and by ethnic groups Georgian, Armenian and Azerbaijanian. For the purposes of this report the sample is also divided by insurance status of the household.