Aung Kyaw Moe’s Updates
Week 1 community assignment
1.a How many strata does the Harmonia survey will have?
Answer: 10 Strata
1.b List and briefly describe at least two advantages and disadvantages of having these many strata compared to having more, for example 100
Advantages:
- More precise estimate
- More representative
- Can estimate district level coverage
- Subgroup analysis
Disadvantages:
- Need more resource (Time , budget, enumerator and supervisors , equipment)
- Data quality and Interviewer bias if there is no proper training and supervision
2. How much will the sample size decrease or increase (overall)
Total Completed Interviews Needed’.
+/-5% precision per strata: 10481
+/-8% precision per strata: 4645
+/-3% precision per strata: 26282
3.a What kind of data do you need to collect to complete a table like the one below?
Summary Table of Survey Implementation, Vaccination Coverage Survey, Harmonia 2018-2019
Sampling frame | Questionnaires |
Date of ethical review : Date/ Month/ Year |
Immunization Status of each child for each vaccine and dosage (BCG, Hep-B ,OPV at birth , DTP-Hib-Hep B 1, DTP-Hib-Hep B 2, DTP-Hib-Hep B 3,OPV1, OPV2, OPV3, MR1, MR2) -By any source -By HBR -By FBR -By recall -Miss opportunities |
Number of Stratum |
Evidence of Vaccination -Home based record (HBR) -Facility based record (FBR) without HBR -Recall by parents or caretaker without (HBRand FBR) |
Number of Cluster |
Factors associated with incomplete or delayed vaccination
|
Total completed interviews |
Provider of Vaccination -Public health facilities -Public health facilities |
Data collection period : Date/ Month/ Year to Date/ Month/ Year |
Number of fully immunized dhildren |
Number of enumerators | |
Number of supervisors | |
3.b Answers
Potential selection bias : household selection if the interviewers do not visit the centrally preselected households or substitution of households
Exclusion of area with conflict or insecurity , Hard-to-reach area ,
Inappropriate time of conducting the survey
If response rate is not 100% and/or several households could not be interviewed,
Need to make Non-response adjustment by calculating the non-response weight and adjusting the design weight when analysis the findings.
4.a
Answer: ( c ) ‘by documented evidence of vaccination + recall’.
4.b
Answer: Recall bias of the parents and care takers
Proportion of Home based records ,
Any difference of HBR distribution among different stratum and clusters
Need to recheck whether interviewers have not given enough time to parents to search for HBR or have not properly reviewed the health facility records during the data collection.
Need to consider when compare with other studies.
4.c
Answer:
Need to collect the immunization status of each vaccine with age
Need to define zero dose by age and vaccine
4.d
Answer: b) quantify how many records are like this and convert these missing values to not vaccinated with DTP1;
5.a
Answers:
- the probabilities of selection of clusters from each province
- the probability that the household was selected from the list of households
- the probability that the respondent was selected from the list of eligible respondents from the household
- These information are needed and important for calculating the design weight
5.b
Answers:
Population data of National , Province and district and sub-district level
- Under 1 year population (surviving infant)
- Total number of live birth per year
- Population Growth Rate
- 12 months to 23 months population
- Average household population
- Total population
- Male ,female population
Good answers in general. Two important things: make sure you post the assignment in the correct place. It needs to go in Community and with the right title. The other one, please check carefully your answers against the Week 1 Considerations, and if any question is there, please ask the team.