31
MICS Survey Design Workshop Multiple Indicator Cluster Surveys Survey Design Workshop Interpreting Field Check Tables

Multiple Indicator Cluster Surveys Survey Design Workshop

  • Upload
    stella

  • View
    44

  • Download
    0

Embed Size (px)

DESCRIPTION

Multiple Indicator Cluster Surveys Survey Design Workshop. Interpreting Field Check Tables. Introduction. MICS surveys are tools to collect high-quality data on a range of globally agreed-upon indicators and national priorities - PowerPoint PPT Presentation

Citation preview

Page 1: Multiple Indicator Cluster Surveys Survey Design Workshop

MICS Survey Design Workshop

Multiple Indicator Cluster SurveysSurvey Design Workshop

Interpreting Field Check Tables

Page 2: Multiple Indicator Cluster Surveys Survey Design Workshop

Introduction

• MICS surveys are tools to collect high-quality data on a range of globally agreed-upon indicators and national priorities

• MICS surveys typically use various systematic checks before, during and after survey implementation to ensure that data quality is high

Page 3: Multiple Indicator Cluster Surveys Survey Design Workshop

Introduction

• Simultaneous data entry should begin about 1 week after fieldwork begins

• Simultaneous data entry provides a suitable platform for– Data quality checks during field work– Providing recommendations to field staff during

implementing phase to improve fieldwork

Page 4: Multiple Indicator Cluster Surveys Survey Design Workshop

Introduction

• Set of tables covering:– Completion and Response rates – Age reporting– Birth date reporting– Question responses – Quantity of observations – vaccination cards, birth

registration cards, height & weight– Completeness of responses – age of death

Page 5: Multiple Indicator Cluster Surveys Survey Design Workshop

Introduction

• Tables provide totals for each indicator and disaggregate information by the field team

• This information allows field supervisors to identify which specific teams have implementation issues and where these issues are most common

Page 6: Multiple Indicator Cluster Surveys Survey Design Workshop

Introduction• Field Check Tables are generated by standard

MICS programmes • Produced every 1-2 weeks by data processing

staff• Revised by the survey coordinator from the

implementation agency in collaboration with UNICEF

• Suitable recommendations can then be made to field staff for immediate implementation

Page 7: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-1: Household response rate

• This is the first indication of how well teams are able to successfully locate, identify and complete household interviews

• Expected to be from 90 to 95 per cent• Below 90 per cent require immediate action to

increase rates

Page 8: Multiple Indicator Cluster Surveys Survey Design Workshop

Reading Field Check Tables

FC-1: Household response rate

Team

Result of household interview

TOTAL N

Household

response rate (%)*

Completed

No household member or no

competent

respondent at home at time

of visit

Entire household absent for extended

period of time Refused

Dwelling vacant / Address

not a dwelling

Dwelling destroye

d

Dwelling not

found Other

(1) (2) (3) (4) (5) (6) (7) (96) Total 90.9 0.7 2.9 3.6 1.2 0.1 0.1 0.5 100.0 12,433 95.4

Team 1 88.2 0.7 3.3 5.6 1.1 0.2 0.1 0.8 100.0 2,488 93.2Team 2 93.6 0.2 2.7 1.9 0.9 0.0 0.0 0.6 100.0 2,497 97.7Team 3 90.5 0.8 4.0 2.6 1.5 0.1 0.0 0.5 100.0 2,495 96.4Team 4 95.2 0.4 1.2 1.7 0.7 0.2 0.0 0.5 100.0 2,461 97.8Team 5 86.9 1.4 3.0 6.3 1.7 0.2 0.2 0.3 100.0 2,492 91.7

Page 9: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-1: Household response rate

• In this example, the household response rate is high (95.4 per cent)

• Teams 2, 3 and 4 all have response rates higher than 95 per cent (see last column)

• Teams 1 and 5 have rates that are below this and would need to have specific attention from the survey coordinator to improve the quality of their performance

Page 10: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-2C: Eligible children under 5 per household

• Based on information from other sources (national census and other surveys), MICS survey coordinators are able to predict:– how many women, children and men there are in

the households interviewed – determine if field teams are adequately capturing

this population during field work

Page 11: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-2C: Eligible children under 5 per household

• FC-2W, FC-2C and FC-2M examine the mean number of eligible women, children and men that is expected in a household

• As household size varies by place of residence of respondents, figures in these tables are disaggregated by urban/rural

Page 12: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-2C: Eligible children under 5 per household

• The target is the minimum mean number of eligible children under 5 years of age per HH that we hope to find

• It should be > 80% of what was expected at the time of sample design

• Thus, if we expected to find 1.2 children under 5 years of age per HH at the time of sample design, teams should be finding a minimum of 0.96 children under 5 years of age per HH

Page 13: Multiple Indicator Cluster Surveys Survey Design Workshop

Reading Field Check Tables

FC-2C: Eligible children under 5 per household

Team

Urban Rural

Number of completed households

Number of eligible

children in those HHs

Mean number of

eligible children per HH

Target not met

Number of completed households

Number of eligible

children in those HHs

Mean number of

eligible children per HH

Target not met

Total 7,322 2,780 0.38 0.38 3,977 1,614 0.41 0.41

Team 1 1,568 532 0.34 0.34 627 229 0.37 -

Team 2 1,159 482 0.42 - 1,179 514 0.44 0.44

Team 3 1,197 450 0.38 0.38 1,060 418 0.39 -

Team 4 1,975 753 0.38 0.38 369 148 0.40 0.40

Team 5 1,423 563 0.40 - 742 305 0.41 0.41

Page 14: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-2C: Eligible children under 5 per household

• Only two teams were able to meet the target (for both urban and rural areas)

• The supervisors and editors should revise the household questionnaires for teams that didn’t meet the target to ensure that the ages of children are correctly recorded and calculated by the interviewers

Page 15: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-4W: Age displacement: women

• Age displacement is an especially critical issue as the eligibility for interview in MICS is age-dependent

• Ages may be displaced to make an eligible respondent older or younger which may affect the eligibility of the respondent

Page 16: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-4W: Age displacement: women

• FC-4W, FC-4C, and FC-4M show the ages of women, children and men by team

• Survey coordinators should examine the overall smoothness of the ages of women, children and men in these tables and the ratio of ages specified in the tables

Page 17: Multiple Indicator Cluster Surveys Survey Design Workshop

Reading Field Check Tables

FC-4W: Age displacement: women

Team

Women's age (12 - 18 years)

TOTALAge ratio (women

15/ women 14)Target

not met 12 13 14 15 16 17 18

Total 150 155 155 120 140 112 119 951 0.77 0.77

Team 1 23 33 37 24 16 24 21 178 0.65 0.65

Team 2 42 33 30 39 51 36 36 267 1.30 -

Team 3 37 30 32 21 32 22 16 190 0.66 0.66

Team 4 19 24 25 15 30 19 20 152 0.60 0.60

Team 5 29 35 31 21 11 11 26 164 0.68 0.68

Page 18: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-4W: Age displacement: women

• The totals do not appear to be smooth• From age 14 to age 15 there is a steep decline

in the number of women interviewed, producing a ratio of 0.77

• This indicates that many women were displaced from age 15 to age 14

Page 19: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-4W: Age displacement: women

• In cases where the age displacement ratios are lower than 0.8, field supervisors and editors should review questionnaires first to ensure that age calculations are correctly made

• In cases where specific teams have excess age displacement, supervisors may decide to observe interviews while they are in progress

Page 20: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-7: Birth date reporting

• Completeness of birth reporting refers to the amount of data that an interviewer is able to get from a respondent regarding age

• Age is composed of the year of birth, month of birth and age of the respondent, according to the respondent

• Role of the interviewer is to gather a complete set of information on age i.e. year of birth and month of birth

Page 21: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-7: Birth date reporting

Team

Completeness of reporting

TOTAL NTarget not

met

Year and month of

birth

Year of birth and

ageYear of

birth only Age onlyOther/No

data

(1) (2) (3) (4) (5)

Total 95.1 2.5 0.0 2.4 0.0 100.0 4,371 95.1

Team 1 83.2 8.1 0.0 8.7 0.0 100.0 749 83.2

Team 2 100.0 0.0 0.0 0.0 0.0 100.0 991 -

Team 3 100.0 0.0 0.0 0.0 0.0 100.0 866 -

Team 4 90.0 5.2 0.0 4.8 0.0 100.0 898 90.0

Team 5 100.0 0.0 0.0 0.0 0.0 100.0 867 -

Page 22: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-7: Birth date reporting

• In the case of teams 1 and 4, there were many respondents who only provided year of birth only (column 3) or year of birth and age (column 2)

• For these teams, it is recommended that interviewers probe respondents more to determine the missing information before further proceeding with the interview

Page 23: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-8: Question responses by team

• For many questions in MICS, there is already an expected level of certain indicators

• From past recent surveys, we may already know, for example, the percentage of women who are currently pregnant or who have had a live birth in the previous two years

• Tables FC8 verify that levels of selected indicators in the MICS are similar to expected levels

Page 24: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-8: Question responses by team

• These tables also show the variation in the levels of these indicators by team

• Overall, the variation in levels should be small• When different teams show widely differing

levels on the survey coordinators and field supervisors should also review these tables to ensure that teams show a deviation in the overall pattern of responses

Page 25: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-8: Question responses by team

Team

Completeness of reporting

Mean number of "Yes" to child

disciplineNumber of children

age 1-14 with a completed child

discipline module

Percentage of women with a live birth in the

last 2 years

Percentage of women

currently pregnant Number of

women age 15-49(1) (2) (3)

Total 1.3 415 3.3 2.4 7992

Team 1 1.3 87 3.7 2.8 1453Team 2 1.4 75 3.8 2.8 1915Team 3 1.5 48 3.4 2.3 1539Team 4 1.1 102 3.0 2.1 1691Team 5 1.4 103 2.6 1.8 1394

Page 26: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-8: Question responses by team• Table shows that 3.3 per cent of women had live birth in the

last two years• We may know from a previous MICS survey that this level was

5.6 per cent• One possible interpretation of this finding is that interviewers

may record some women as “no live birth in the last two years which effectively reduces the workload of interviewers

• Field supervisors should look out for unusual levels of these indicators and verify that results are accurate with revisits to households if necessary

Page 27: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-9: Birth registration

• This table shows the proportion of birth certificates seen by interviewers for children under five

• As a measure of data quality, interviewers ask to see the birth certificates (or equivalents in the country) of children under five to ensure that these documents exist

• In MICS, at least 90 per cent of birth certificates should be seen

Page 28: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-9: Birth registration

Team

Percent of children

reported to have a birth certificate

Percent of children

whose birth certificate

was seen by interviewer

Percent of children

whose birth was

registered but does not have a birth certificate

Percent of all births with a

birth certificate or

registeredNumber of

children

Proportion of birth

certificates seen (%)

Target not met

(1) (2) (3) (4) (5) (6)=(2)/(1)

Total 99.6 47.3 0.3 99.9 4,371 47.5 47.5

Team 1 99.7 45.8 0.3 100.0 749 45.9 45.9

Team 2 99.6 60.0 0.4 100.0 991 60.3 60.3

Team 3 99.3 36.1 0.5 99.8 866 36.4 36.4

Team 4 99.4 54.7 0.4 99.9 898 55.0 55.0

Team 5 99.9 37.4 0.0 99.9 867 37.4 37.4

Page 29: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-9: Birth registration• Overall, no team surpassed the 90 per cent

requirement • Supervisors should observe if interviewers

request to see the birth certificates from mothers/caretakers at the beginning of the interview

• Supervisors should ensure that interviewers probe respondents for the birth certificates and explain why it is important to see them

Page 30: Multiple Indicator Cluster Surveys Survey Design Workshop

FC-9: Birth registration

• Interviewers should note that birth registration may carry a legal burden in some countries and may be a sensitive issue

• Therefore, interviewers should re-affirm to respondents that the information that the respondent gives is confidential

Page 31: Multiple Indicator Cluster Surveys Survey Design Workshop

Thank you