the weai forward
Post on 25-Jul-2015
92 Views
Preview:
TRANSCRIPT
The “WEAI” forward:What we’ve learned and where we’re
going
Photo credits: Chiara Kovarik and Katie Sproule
Hazel Malapit - Research Coordinator/IFPRIWEAI Stakeholder Convening, Washington DC, 27 May 2015
Overview• Revising the WEAI overview• 10 indicator vs 6 indicator WEAI comparison • Indicator by indicator highlights from the pilot– Modifications to questionnaire– Recommendations of what to keep– Rationale
• Weighting issues • Name challenge • Next steps
p. 2
Revising the WEAI
• USAID Goals:– Streamline survey– Reduce time to administer by ~30% – Improve problematic modules (time use, autonomy in
production, credit, and speaking in public) • Process – Conducted cognitive testing– Conducted pilot fieldwork in Bangladesh and Uganda
• Outcome: Version of WEAI with 6 indicators and streamlined questions
Survey times
Pilot 1.1 Pilot 2.0Ban Uga Ban Uga
DM in production & income 6 4
DM in production, income & hh 10 5
Assets 11 6 Assets 10 5Credit 5 3 Credit 6 3Leadership 10 5DM in hh & autonomy 12 9 Autonomy vignettes 16 8Time use 15 9 Time use-7dr 10 7
Time use-24hr 12 7Leisure - - Leisure 3 1
Leadership - -TOTAL* 62 37 TOTAL 7dr** 65 34Target (30% reduction) 43 26 TOTAL 24hr** 67 34
Median minutes spent collecting each sub-module, both countries
*Includes Leadership time from 2.0 **Includes Leisure time from 1.1
p. 4
DOMAIN INDICATORS1 Production Input in productive decisions
2 Resources Ownership of assetsAccess to and decisions on credit
3 Income Control over use of income4 Leadership Group membership
5 Time Workload
DOMAIN INDICATORS1 Production Input in productive decisions
Autonomy in production
2 Resources Ownership of assetsPurchase, sale, or transfer of assetsAccess to and decisions on credit
3 Income Control over use of income
4 Leadership Group membershipSpeaking in public
5 Time WorkloadLeisure
Old vs new comparison
Original: 5 domains, 10 indicators New: 5 domains, 6 indicators
ProductionIndicator 1: Input in productive decisions
Indicator 2: Autonomy in production
• Pilot modifications– Input in productive decisions: no significant
modifications piloted – Autonomy in production: tested original autonomy
questions against a series of vignettes that ask about autonomy in various agricultural activities
• Recommendations– Input in productive decisions: keep– Autonomy in production: drop
Autonomy
missing
never true
not very true
somewhat true
always true
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00
Figure 7: Distribution of responses to "no option" vignettes, by activity & sex of respondent, Bangladesh
Livestock raising female
Livestock raising male
Taking crops to market female
Taking crops to market male
Types of crops to grow female
Types of crops to grow male
p. 7
ResourcesIndicator 1: Ownership of assets
Indicator 2: Purchase, sale or transfer of assetsIndicator 3: Access to and decisions on credit
• Pilot modifications– Ownership of assets
• Ownership definition changed from “who owns most” to “do you own any”? • Eliminated question asking for number of assets.
– Purchase, sale or transfer of assets: Tested combined rights question (sell, give away, mortgage, rent) versus separate rights questions as analysis of original pilot data suggested rights are bundled together.
– Access to and decisions on credit • Added question on hypothetical borrowing. • Changed definition of adequacy for credit indicator – respondents from non-borrower
households are counted as “adequate” if their household can borrow from at least one credit source (excluding friends and family).
• Recommendations– Ownership of assets: keep, but modify – Purchase, sale or transfer of assets: drop – Access to and decisions on credit: keep, but modify
ResourcesIndicator 1: Ownership of assets
Indicator 2: Purchase, sale or transfer of assetsIndicator 3: Access to and decisions on credit
UGANDA
Empowerment rates conditional on asset ownership empowerment
1.0 VERSION 1.1 VERSION 2.0 VERSION Men (%) Women (%) Men (%) Women (%) Men (%) Women (%)Asset rights indicator 94.36 89.61 97.79 91.41 97.92 86.7Credit indicator 33.21 25.41 64.39 65.36 78.68 75.28
Empowerment rates conditional on asset rights empowerment
1.0 VERSION 1.1 VERSION 2.0 VERSION Men (%) Women (%) Men (%) Women (%) Men (%) Women (%)Ownership indicator 99.60 93.88 100 100 100 99.39Credit indicator 32.80 25.86 64.34 65.71 79.1 79.35
OVERALL CORRELATIONS (PWCORR)UGANDA Men
joint own asset joint any right
asset joint any right asset -- 1.0000 -- joint any credit -- 0.0855 -- (0.3221) Women
joint own asset joint any right
asset joint any right asset 0.1905*** 1.0000 (0.0083) joint any credit 0.0254 0.2382*** (0.7338) (0.0012)
IncomeIndicator 1: Control over use of income
• Pilot modifications – Combined into section with input in production decisions– Not clear if it saved time
• Recommendations – Control over use of income indicator: keep – Change definition: Adequate if there is at least one
domain where individual has some input in income decisions [or feels she/he can make decisions regarding wage, employment and minor hh expenditures; as long the only domain in which the individual feels that he/she makes decisions is not minor household expenditures]
LeadershipIndicator 1: Group member
Indicator 2: Speaking in public
• Pilot modifications– Group member: Minor changes (changes in sequence in qx, group categories,
added DK as response)– Speaking in public: New questions designed to be more broad and less
controversial; new adequacy cutoff if reported comfort in speaking in public or has spoken up in last 3 months
• Recommendations– Group member indicator: keep – Speaking in public indicator: drop
• Rationale– Group member indicator emerged as a top constraint in 6 out of 13 countries from
the FtF WEAI baseline data – While the new speaking in public questions seem to do better than the old ones,
this indicator has not been useful for analyses and in many places has been sensitive to collect (was not collected in Cambodia or Tajikistan)
TimeIndicator 1: Workload
Indicator 2: Leisure
• Pilot modifications– Workload: collected a 7-day recall (both last 7-day and typical 7-days), 24-hr recall w/ primary and
secondary activities and a 24-hr recall w/ primary activities only– Leisure: no change except sequence in qx
• Findings– 7 day recall faster to implement than 24 hr recall w/ primary and secondary; however, concern
about recall accuracy and burden on enumerators to calculate – 24 hr recall w/ primary only saves 2-3 minutes compared to 24 hr recall w/ primary and secondary – Both Uganda and Bangladesh teams prefer 24 hr recall to 7 day recall because of recall accuracy
concerns • Recommendations
– Workload indicator: keep -- collect 24 hr recall w/ primary activities only – Leisure indicator: drop
• Rationale– Workload indicator was identified as a top constraint in 6 out of 13 countries in the FtF WEAI
baseline data – Leisure indicator has not been useful for analyses because concern on women’s adaptive
expectations
Workload 7dr vs 24hr
Average work hours by method, country and sex of respondentBangladesh Uganda
male female total male female totalPilot 1.1N 179 221 400 136 140 27624hr primary & secondary 9.4 8.8 9 8.1 10 9.124hr primary only 9.3 8.8 9 8.1 10 9.1
Pilot 2.0N 175 225 400 144 189 3337dr actual 8 8.7 8.4 8.7 13 117dr typical 9.8 11 10 11 14 13
N 175 225 400 142 149 29124hr primary only 8.5 8.9 8.7 6.3 9.1 7.7
p. 13
DOMAIN INDICATORS WEIGHT
1 Production Input in productive decisions
1/5 1/5
2 Resources Ownership of assetsAccess to and decisions on credit
1/10 2/15
1/10 1/15
3 Income Control over use of income
1/5 1/5
4 Leadership Group membership 1/5 1/5
5 Time Workload 1/5 1/5
TOTAL 100%
DOMAIN INDICATORS WEIGHT1 Production Input in productive
decisionsAutonomy in production
1/10
1/10
2 Resources Ownership of assetsPurchase, sale, or transfer of assetsAccess to and decisions on credit
1/15
1/15
1/15
3 Income Control over use of income
1/5
4 Leadership Group membershipSpeaking in public
1/101/10
5 Time WorkloadLeisure
1/101/10
TOTAL 100%
Old vs new comparison - weights
Original: 5 domains, 10 indicators New: 5 domains, 6 indicators
Old vs new comparison - 5DE decomposition
10 6 - 1/10 w 6 - 1/15 w 10 6 - 1/10 w 6 - 1/15 wFemale Male
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
17.111.3 11.7 8.3
1.7 1.7
23.5
16.2 13.214.8
2.6 2.3
14.4
16.917.5
13.0
12.0 12.0
31.5
37.3 38.7
39.8
46.2 46.3
13.318.3 19.0
24.1
37.6 37.7
Bangladesh: 10 vs 6 indicators
Production Resources Income Leadership Time
10 6 - 1/10 w 6 - 1/15 w 10 6 - 1/10 w 6 - 1/15 wFemale Male
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
15.319.8 20.2 21.7 23.3 22.7
19.0 13.5 10.817.4 14.0 11.4
26.3 27.0 27.6 8.72.3
2.3
14.4 18.0 18.4
26.037.2
38.6
25.0 21.6 23.0 26.0 23.3 25.0
Uganda: 10 vs 6 indicators
Production Resources Income Leadership Time
Top contributors to disempowerment
Bangladesh Uganda Indicators Women Men Women Men 10 1. group
member2. credit3. speaking in
public
1. group member
2. workload3. control over
income
1. control over income
2. workload3. credit
1. group member
2. credit3. workload
6 credit weight=1/10
1. group member
2. workload3. control over
income
1. group member
2. workload3. control over
income
1. control over income
2. workload3. input in
decisions
1. group member
2. workload2. input in decisions
6credit weight =1/15
1. group member
2. workload3. control over
income
1. group member
2. workload3. control over
income
1. control over income
2. workload3. input in
decisions
1. group member
2. workload3. input in
decisions
Old vs new comparison - rankings
The new WEAI
• Reflects 5 domains
• Collects on fewer (6) indicators
• Smaller and quicker to implement
DOMAIN INDICATORS WEIGHT
1 Production Input in productive decisions
1/5 1/5
2 Resources Ownership of assetsAccess to and decisions on credit
1/10 2/15
1/10 1/15
3 Income Control over use of income
1/5 1/5
4 Leadership Group membership 1/5 1/5
5 Time Workload 1/5 1/5
TOTAL 100%
Name that WEAI contest!
• Your mission: Come up with a new name for the 6 indicator WEAI!!
• Instructions: Write your name, email, and name ideas on the index card. You can propose as many names as you like. Be CREATIVE !!
• Logistics: The WEAI team will select the winning entry. The winner gets bragging rights and a token.
Next Steps
• Finalizing the new WEAI• Name the new WEAI• Creating a “Community of Practice” for WEAI
users
Thank you!
Any questions?Contact Hazel Malapit at h.malapit@cgiar.org
Visit the WEAI Resource Center:http://www.ifpri.org/book-9075/ourwork/program/weai-resource-center
Proportion of respondents who have inadequate achievements in control over use of income (raw and censored), by country and sex of respondent
IncomeIndicator 1: Control over
use of income
p. 21
Bangladesh Ugandamale female all male female all
Pilot 1.1N 179 221 400 136 178 314% missing 0% 0% 0% 0% 0% 0%N with 10 indicators 175 219 394 73 133 206raw headcount 0.006 0.128 0.074 0.082 0.263 0.199censored headcount 0.006 0.128 0.074 0.055 0.226 0.165Pilot 2.0, 24 hr time moduleN 175 225 400 145 193 338% missing 0% 0% 0% 1% 2% 1%N with 10 indicators 172 221 393 128 143 271raw headcount 0.041 0.131 0.092 0.023 0.224 0.129censored headcount 0.041 0.127 0.089 0.023 0.210 0.122Pilot 2.0, 24 hr time module + new control over income definition% missing 61% 48% 54% 41% 35% 38%N with 10 indicators 69 116 185 74 107 181raw headcount 0.391 0.655 0.557 0.487 0.682 0.602censored headcount 0.319 0.603 0.497 0.378 0.626 0.525
top related