TANZANIACOGNITIVE TEST RESULTS

1
TANZANIA-COGNITIVE TEST RESULTS

• QUESTIONS ON HEARING
• By. Ms. Albina Chuwa
• October, 2006

2
Background Information

• The primary goal of cognitive testing is to
• Determine whether questions are being interpreted
  as intended. That is, if they are capturing
  aspects of the selected functional domains
• Determine whether questions are interpreted
  consistently across countries

3
How was the Cognitive test was Done in Tanzania

• Pre-test started in August/September 2006
• We received a consultant from WG to assist NBS on
  the preparations of the Cognitive test
• We also received training and interview
  instructions from Miller Washington Group
  expert.

4
How was the Cognitive test was Done in Tanzania

• Recruitment of the Interviewers
• Four experienced Statistician from National
  Bureau of Statistics (NBS) did the interview
• Main stakeholder (Ministry of Health and Social
  Welfare) was first contacted to provide initial
  information about all institutions dealing
  people with disability

5
How was the Cognitive test was Done in Tanzania

• Training
• Took 4 hours CD obtained from Miller was used
  together with inputs from consultant
• Mock interview was done among trainees this was
  to ensure the questions understood well before
  carrying the actual pre-test

6
Selection of the Respondents

• A random purposive sample of 25 respondents were
  drawn from institutions dealing with people with
  disability and private households.
• For each interview, there were two recordings
  (Self-report and Proxy report)

7
How was Data Collection Done

• Each interviewer was provided with enough
  questionnaire
• Consultant was observing the interview.

8
How was Data Collection Done

• Official letter from Director General which
  entailed the purpose of conducting test was sent
  to respondents prior to the data collection
• Interpreter was used for two respondents with
  hearing impairment

9
How was Data Collection Done

• Data collections took about three days
• Each interview took about one hour. Much time was
  taken for those respondents who had hearing
  impairments, illiterate and especially female.

10
Data Processing

• Data processing was done simultaneously with data
  collection
• Data entry was done using WG Format EXCEL.
• Raw data was sent to WG

11
Analysis of the results

• Interpretation of Patterns
• Non-problematic Response Patterns
• A Respondent has no sign of impairment and
  reports no disability
• B Respondent shows clear impairment and reports
  disability

12
Analysis of the results

• Borderline Response Patterns
• D Respondent does not report disability but
  admits some functioning problems may have taken
  a literal interpretation of the crowded question
  OR real impairment but have not come to think of
  themselves as having a real hearing problem

13
RESULTS
14
Figure 1.1 Distribution of Self-Report by
Pattern - Hearing

• Non-problematic Response Patterns
• A Respondent has no sign of impairment and
  reports no disability
• B Respondent shows clear impairment and reports
  disability
• D Borderline Response Patterns
• Respondent does not report disability but admits
  some functioning problems may have taken a
  literal interpretation of the crowded question OR
  real impairment but have not come to think of
  themselves as having a real hearing problem

15
Discussions

• Self Report
• Majority of the respondents reported no sign of
  impairment and reported no disability (72)
• Respondents showed clear impairment and reported
  disability (12)

16
Self- report

• Respondents did not report disability but admits
  some functioning problems and this accounted for
  four percent
• We have noted that pattern D many respondents
  reported conversation problems in the crowded
  room. This could happened to any person.

17
(No Transcript)
18
Figure 1.2 Distribution of Proxy-Report by
Pattern - Hearing
19
Discussion Proxy -Report

• 64 percent of proxy-reporting showed pattern A
  that no sign of impairment and reports no
  disability.
• While
• 16 percent of proxy-reporting showed clear
  impairment and reported disability.

20
Discussion Proxy -Report

• 20 percent of proxy-reporting showed that
  respondents did not report disability but admits
  some functioning problems like conversation in
  the crowded room.

21
Analysis of Social and economic Characteristics
of the Respondents

• This analysis was not done due to time
  constraints
• Analysis will be done be shared with other
  countries through e-mails

22
Agreement and Disagreement with WG Questions

• All WG core questions agreed in Tanzania

23
General Comments on the WG Core Questions

• Questions on hearing aid was not common in
  Tanzania. This type of questions could be
  captured if sample size was large enough.
• To read categories to respondents would tend to
  bias results. This would be applicable for a
  country where literacy rate is very high

24
General Comments on the WG Core Questions

• Questions were not translated into swahili
  although interview as conducted using both
  languages.
• This probably lead to mis-interpretation of some
  responses

25
General Comments on the WG Core Questions

• All respondents were from urban part of Dar es
  Salaam. No rural part was covered.
• Most of the respondents were clever enough to
  evaluate and estimate responses.
• This probably led to biased or estimation
  difficulty.

26
Challenges for conducting the cognitive test in
Tanzania

• Multiple cultures contributed to ambiguous
  responses
• Multiple levels of development All respondents
  were from Dar es Salaam
• In Tanzania pattern A, B and D were the common
  probably was due to small sample size and spread.

27
The Way Forward

• In Tanzania WG Core questions will be included in
  the Disability Pilot Survey that will be
  conducted in Tanzania 2006/07.
• Based on this
• The actual disability survey will be done after
  success of the pilot test in FY 2007/08 with a
  sample of 19,000 households. The results will
  provide national estimates

28
Discussions with CHAWATA
29
(No Transcript)
30
Discussions with CHAVITA
31
ENDS

• Thank you for listening