The Cognitive Survey for Mauritius

1
The Cognitive Survey for Mauritius test and
results

• Presented by
• Mr Chettun Kumar ARIANAICK
• Statistician

2
The Cognitive survey

• The survey was conducted in April 2006
• The sample size was 24
• The questionnaire provided by Washington Group
  was used except that the 2000 population census
  question on disability was added
• The questionnaire was translated in creole, a
  local language
• Three interviewers and one supervisor was
  required for the fieldwork
• I had the overall responsibility for the survey
  including the fieldwork

3
The sample frame

• The sample frame used was the 2000 Housing and
  Population Census data
• The population census records with at least one
  person with disabilities within the household
  were merged with the corresponding housing census
  records in order to obtain the geographical
  location of the household

4
Selection criteria

• Given that around 40 of the population lives in
  rural areas, out of the 3 regions selected one
  was urban
• The selection was done in such a way as to cover
  the wide spectrum of disability as well as to be
  representative of age and sex distribution of the
  population

5
Locating the respondents

• Each interviewer was provided with a list of 8
  respondents
• The list had the following information the names
  of the head of household, the address and the hh
  size together with the age, sex and type of
  disability of the person with disabilities
• The interviewers reported having no problem
  identifying the person with disabilities within
  the household
• 25 of respondents could not be contacted and
  were replaced

6
The interviews

• The interviewer filled in the section asking for
  the list of persons in the household first in
  order to locate the respondent
• The interviewer then went through the
  questionnaire, asking the questions as they
  appear, i.e, asking questions about the
  respondent himself/herself and then again about
  another member in the household (proxy)
• Each interview lasted around one hour

7
Data processing

• The completed questionnaires were edited by the
  interviewer, the supervisor as well as by myself
  in the office
• The edited questionnaires were entered on the
  excel worksheet provided by Washington Group

8
Answers to the core question on ability to
communicate (TSCOMM/TPCOMM)

• Out of the 24 respondents, 16 reported as having
  difficulty communicating
• Only one proxy respondent had difficulty
  communicating
• Among the 16 respondents having difficulty
  communicating, the detailed breakdown was some
  diff 6 lots of diff 9 cannot do at all 1

9
Answers to the question Why did you answer that
way?

• Answer 0 No communication problem
  8
• Answer 1 People need to repeat/ nerve problems/
• forget things quickly
  - 6
• Answer 2
• Deaf/dumb/uses sign language/diff. communicating
  - 6
• Weak following dialysis/ Not talk too much
  - 1
• Mental problems ( helped by other persons)
  - 2
• Answer 3 I am handicapped (suffers from acute
  epilepsy)- 1
• 
  Total
  24

10
Comparing responses to TSDIFF and TSEFF for
respondents who reported having difficulty
communicating in core question
11
Comparing responses to TSSAY and TSCONVO for
respondents who reported having difficulty
communicating in core question
12
Correlation matrix
13
Problematic responses Yes to core question and
no to extended questions (pattern N)

• Respondent No 21
• Profile
• Age 28 yrs
• Sex male
• Education 6 years
• Not working health reasons
• Never married
• Low income group
• Has vision problem
• Had hand surgery and two screws inserted in his
  hand
• Refrains from going out because of operation

• Respondent No 23
• Profile
• Age 52 yrs
• Sex male
• Education 6 years
• Own account worker
• Married
• Low income group
• Has vision/ hearing/cognitive/self care problems
• Had a motorcycle accident
• Uses crutches
• I forget quickly. My wife complains on this issue

14
Relation between response patterns and
demographic/socio-economic characteristics of
respondents

• In order to simplify analysis, the response
  patterns have been categorised into three groups
  good, fair and bad
• Given the small sample size, test for association
  between response pattern and demographic/socio-eco
  nomic characteristics could not be undertaken
  given the presence of cells with frequency less
  than 5

15
Thank you for your attention