Mindanao Youth for 
Development (MYDEV)
Program
FY15 Impact EvaluationReport
Measuring Youth Employment, Perceptions, and Skills 
November 23, 2016
Prepared by: 
Education Development Center, Inc. 
(EDC)
1025 Thomas Jefferson St.
Suite 700W
Washington, DC 20007
This report was prepared by the Education Development Center, Inc. for review by the United States Agency 
for International Development
Cooperative Agreement Number AID-492-A-13-00012
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report i
Table of Contents
Table of Tables.............................................................................................................................................. ii
Table of Figures............................................................................................................................................. ii
Acronyms and Abbreviations....................................................................................................................... iii
EXECUTIVE SUMMARY ................................................................................................................................. iv
Methodology............................................................................................................................................. v
Impact Findings and Outcomes for MYDev Youth.................................................................................... v
Key Takeaways and Recommendations................................................................................................... vi
I. THEORY OF CHANGE .............................................................................................................................1
II. METHODOLOGY ....................................................................................................................................2
A. Study Design......................................................................................................................................2
B. Sample...............................................................................................................................................3
C. Data Collection..................................................................................................................................4
D. Tools..................................................................................................................................................4
E. Limitations.........................................................................................................................................6
III. DEMOGRAPHICS................................................................................................................................6
IV. IMPACT FINDINGS and OUTCOMES for MYDev YOUTH ...................................................................7
A. Employment......................................................................................................................................7
1. New Employment..........................................................................................................................8
2. Impact Findings...........................................................................................................................10
3. Previously Employed MYDev Youth............................................................................................11
4. Unemployed Youth .....................................................................................................................14
5. Conclusions from Employment Study.........................................................................................15
B. Life Skills, Work Readiness, and Leadership Skills...........................................................................15
1. Skill Gains for MYDev Youth........................................................................................................15
2. Impact Findings...........................................................................................................................16
C. Government and Community Perceptions.....................................................................................17
1. MYDev Youth’s Government and Community Perceptions........................................................17
2. Impact Findings...........................................................................................................................18
D. Youth’s Assets and Contexts...........................................................................................................19
1. MYDev Youth’s Assets.................................................................................................................19
2. MYDev Youth Contexts...............................................................................................................20
3. Impact Findings...........................................................................................................................21
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report ii
E. Connections Between Employability, Skill Gains, and Perceptions................................................23
1. Relationships with Employment Outcomes................................................................................23
2. Relationships with Improved Perceptions..................................................................................23
V. CONCLUSION.......................................................................................................................................24
VI. RECOMMENDATIONS......................................................................................................................25
VII. ANNEX-IMPACT EVALUATION SLIDE PRESENTATION .....................................................................27
Table of Tables 
Table 1: Reliability of the YES and YPS..........................................................................................................5
Table 2: Reliability of DAP in MYDev Context...............................................................................................5
Table of Figures 
Figure 1: MYDev Theory of Change ..............................................................................................................2
Figure 2: Study Design...................................................................................................................................3
Figure 3: Area by Group................................................................................................................................7
Figure 4: New and Better Employment ........................................................................................................8
Figure 5: Profile of a Newly Employed Youth ...............................................................................................9
Figure 6: Difference in Difference Estimate of Employment Gains............................................................10
Figure 7: Employment, by Group and Sex ..................................................................................................11
Figure 8: Profile of A Continuously Employed MYDev Youth .....................................................................12
Figure 9: Index for Better Employment ......................................................................................................13
Figure 10: Profile of an Unemployed Youth................................................................................................14
Figure 11: Skills Gains for MYDev Youth.....................................................................................................16
Figure 12: MYDev Youth's Community and Government Perceptions Gains.............................................18
Figure 13: The DAP's Asset Categories and Contexts .................................................................................19
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report iii
Acronyms and Abbreviations
ALS Alternative Learning System
CDCS Country Development Cooperation Strategy
EDC Education Development Center, Inc.
ENGAGE Enhancing Governance Accountability and Engagement
FY Fiscal Year
LGU Local Government Unit
LoP Life of Project
OSY Out-of-School Youth
OSYDA Out-of-School Youth Development Alliance
TVET Technical and Vocational Education and Training
USAID United States Agency for International Development
USG United States Government
YES Youth Employment Survey
YPS Youth Perceptions Survey
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report iv
EXECUTIVE SUMMARY
The MYDev Project supports USAID-Philippines’ overall CDCS goal through Development Objective 2: 
Peace and Stability in Conflict-Affected Areas in Mindanao Improved. MYDev provides experiential 
training and post training support to improve life skills, increase civic engagement, and increase
employability for 19,000 vulnerable out-of-school youth (OSY) in eight (8) conflict-affected areas: 
Cotabato City, Marawi City, Isabela City, Lamitan City and Zamboanga City, and the municipalities of 
Parang in Maguindanao, and Jolo and Indanan in Sulu. In addition, MYDev organizes and mobilizes local 
government officials, local businesses, and other stakeholders into OSY Development Alliances (OSYDA) 
aimed at coordinating multi stakeholder support to job skills training and livelihood pathways.
MYDev program implementation is focused on three main components (Intermediate Results):
 Strengthening life skills, leadership, and civic engagement for vulnerable out-of school-youth 
through capacity building of local government units, communities, and stakeholders to improve 
education and training services for out-of-school youth;
 Increasing out-of-school youth access to education through the Department of Education’s 
Alternative Learning System (ALS);
 Increasing out-of-school youth employability by providing equitable access to relevant education 
and skills training.
EDC has designed this quasi-experimental impact study to better understand MYDev’s contribution to 
improving OSY community and government perceptions, skills, and employment outcomes. To evaluate 
MYDev’s progress in these areas, the impact study seeks to answer the following three evaluation 
questions, using both a pre-test and post-test for MYDev intervention groups and comparison groups.
Q1. Do youth participants demonstrate improved livelihoods as a result of project activities? 
Q2. Do youth participants demonstrate improved perceptions of government and community as a 
result of project activities?
Q3. Do youth participants demonstrate improved life, work readiness, and leadership skills as a result 
of project activities?
The choice of a quasi-experimental approach has allowed the MYDev team to avoid excluding certain 
youth from participation in MYDev. This has been important to avoid some youth possibly feeling further 
marginalized, as many of these youth are already routinely excluded in a variety of ways. Instead, 
admission into MYDev’s training programs has remained open to any OSY in participating geographic 
areas that are among MYDev’s target population, with comparison group youth consisting of similar youth 
from neighboring barangays.
Initial results from this impact study of MYDev’s work in FY15 were presented to USAID by the MYDev 
leadership and M&E team, during a meeting hosted by the USAID Mission Director, Dr. Brems, in Manila,
on November 2, 2016. The presentation was also attended by the USAID Deputy Mission Director, Office 
of Education Director, MYDev AOR and other USAID representatives. The impact evaluation presentation 
is attached as an annex to this report. The valuable feedback provided by the USAID leadership team has 
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report v
been incorporated into this final evaluation report. EDC looks forward to sharing additional M&E data 
associated with this impact evaluation that will be collected from subsequent MYDev youth cohorts during 
the coming two years of project implementation. 
Methodology 
To understand MYDev’s contribution to improving OSY’s community and government perceptions, skills, 
and employment outcomes, MYDev designed and implemented a quasi-experimental study with both a 
pre-test and post-test for the intervention groups and comparison groups. Intervention and comparison 
group youth were assessed at baseline and four to six months after the intervention group completed 
training.
The use of a quasi-experimental approach means that the equivalence of the comparison and intervention 
groups cannot be assumed. EDC took care in selecting comparison group youth who were very similar to 
intervention group youth in five key characteristics: sex, age, education status, employment status, and 
place of residence. 
The findings of this evaluation are based on sample data that were generalized to the population of MYDev
youth who participated in trainings beginning in February of FY15. The sample data was weighted to this 
population and was collected from youth using three tools: the Youth Employment Survey (YES), the Youth 
Perceptions Survey (YPS), and the Development Assets Profile (DAP). 
Impact Findings and Outcomes for MYDev Youth 
This quasi-experimental impact evaluation of MYDev’s work in FY15 has revealed that many MYDev youth 
seem to be improving their economic and levels of social engagement:
 Almost 24% of the intervention group was newly employed at endline.
 42.2% of MYDev youth were working at baseline and improved their work situations between 
baseline and endline.
 Almost 70% of MYDev youth improved in either life, leadership, or work-readiness skills over 
the course of the intervention. 
 51.5% of intervention group youth improved their perceptions of community and 52.5% 
improved their perceptions of the government.
This study can conclude that MYDev youth are better off than non-MYDev youth in terms of having 
gained more new employment, though among working MYDev and non-MYDev youth, the quality of 
employment is statistically similar. Further, new and continuously employed youth tend to be urban males 
working in one temporary/substitution job at a time, while continuously unemployed youth tend to be 
rural females. 
New Employment: MYDev youth obtained new employment at a higher rate than did comparison 
group youth, with almost 24% of intervention group youth newly employed at endline compared to about 
14% of youth in the comparison group. This approximate 10 percentage-point difference in the rates of 
newly employed youth between the intervention and comparison groups is significant.
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report vi
Better Employment. Approximately 42.2% of MYDev youth and 44.4% of comparison group youth who 
were working at both baseline and endline had better employment at endline. The difference between 
the rates of better employment of MYDev and comparison group youth is not statistically significant. 
This study also examined a cumulative measure capturing both newly employed youth and youth with 
better employment. In the intervention group, about 32.2% of youth reported new or better 
employment at endline compared to only 17.9% of comparison group youth.
Life Skills, Work Readiness, and Leadership Skills. A statistically similar proportion of intervention 
group and comparison group youth improved their skills across the study period, with about 69.8% of the 
intervention group and about 70.0% of comparison group youth improved in either life skills, work 
readiness or leadership skills over the course of the MYDev intervention. The difference in the rates of 
improvement between MYDev and comparison group youth is not statistically significant. While gains in 
the intervention group were smaller than gains in the comparison group, intervention group youth’s 
endline scores were higher than the comparison group’s endline scores.
Government and Community Perceptions. Statistically, a similar proportion of intervention and 
comparison group youth improved in their community or government perceptions. An examination of 
youth’s perceptions of community and government respectively shows that comparison group youth’s 
perceptions of the government improved, by more than half a level on average, more than intervention 
group youth’s perceptions, which improved by less than a quarter of a level. Similarly, comparison group 
youth’s perceptions of their communities improved by more than half a level on average more than 
intervention group youth’s perceptions (which improved by roughly a quarter of a level). These
differences are statistically significant.
Key Takeaways and Recommendations 
This impact evaluation has afforded us the opportunity to learn more about the mechanisms behind 
improved economic and social outcomes, and how those two elements interact with each other and with 
youth’s demographics. Recommendations are grouped into five categories.
Improved livelihoods. A central intended outcome of MYDev and OSYDA programming is improved 
employment outcomes for out-of-school youth. All analyses point to success for MYDev regarding 
improving the employment situations of MYDev youth. 
Recommendation: Future programing should strengthen and broaden partnership with private 
sector, identifying market driven opportunities with a greater emphasis emerging skills 
opportunities.
Recommendation: Future programing should support activities to encourage entrepreneurship 
and self-employment by connecting OSY to local government and private sector programs, 
identifying support mechanisms for strengthening the new entrepreneurship skills for long term 
sustainability and profitability.
Improved Perceptions of Government and Community. MYDev programing aims to improve youth’s 
engagement within their communities, while also improving youth’s perceptions of their communities 
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report vii
and governments. The majority of MYDev youth had improved their government or community 
perceptions by the time of endline administration. 
Recommendation: Future programming should continue to promote deepening of OSY￾community connections through increased partnerships with government institution, community 
leaders and other community groups in civic engagement, peace building and youth development 
activities.
Experiential Learning to Improve Work Readiness. The Work Readiness component encompasses 
eight questions measuring skills like solving conflicts at work, looking for a job, and working towards a pay 
upgrade. MYDev youth that were employed at endline had higher work-readiness skills than unemployed 
youth.
Recommendation: Further investigate approaches to developing work-readiness skills to 
determine the ideal combination of technical training and experiential learning. 
Recommendation: Future programming should promote the recognition by government agencies 
and private sector of work readiness skills as a facilitating factor for youth’s success.
The Role of Gender. Gender and skill development seem to be related, with female youth improving 
significantly less than their male counterparts in life skills and work-readiness components, though the 
mechanisms at play behind this relationship are unclear. Further, females have lower rates of employment 
than males and tend to enter different sectors when they do have employment. These females also tend 
to feel safer at work and save more, all while working almost ten hours more than male youth on average. 
Recommendation: Future programming should investigate unique challenges and needs of 
female out-of-school youth and determine if supplemental program may be helpful to lessen the 
gender gaps in employment outcomes and skill gains.
The Role of Geography. One’s geographic area matters in a way that is complex and likely related to 
more than just personal skills and opportunities, but also to contexts, including local markets and 
communities. Rural youth had significantly worse perceptions of the government and community than 
youth from other geographic areas and reported having less internal and external assets than urban 
youth.
Recommendation: Future programming should aim to target the needs of rural youth through
tailoring MYDev programming to the rural labor market context and engagement with 
government services.
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 1
I. THEORY OF CHANGE
Within the context of the eight conflict-affected areas targeted by MYDev external factors, combined with 
local socio-economic circumstances, bear acutely on vulnerable out-of-school youth populations. MYDev 
therefore aims to strengthen positive factors for pro-social engagement of OSY by supporting 
reaffirmation of the social contract between and among different local government and other stakeholder 
groups, including OSY. Thus, MYDev’s implementation is informed by USAID’s youth policy, USG’s 3D 
approach and MYDev’s development hypothesis: If the social contract between local government and 
communities—vulnerable out-of school-youth (OSY) in particular—are strengthened through incremental 
gains in mutual trust, conflict affected communities will begin to benefit from the accretion of social capital 
that underpins local area stability, security and prosperity. 
Taken together, OSY and OSYDA programming is designed to produce an environment in which OSY cease 
to feel alienated from their communities and instead are able to contribute through work and community 
service. Such integration into the community—as produced by both OSY training programs and OSYDA 
activity—would be evidenced by first by improved leadership, life, and work readiness skills accompanied 
by improved perceptions of the community and government. The improvement of skills and perceptions 
in the context of an environment that is conducive to OSY’s development and growth would lead to 
improved work outcomes for OSY.
Yet can such a theory—in which a holistic approach to creating an enabling environment for OSY is 
combined with skills-focused training—be proven in the context of MYDev? This impact evaluation seeks 
to determine whether or not links can be found between youth participation in MYDev’s programming 
and their outcomes in three key areas: 
 Employment, including self-employment
 Perceptions of the community and government
 Life, Leadership, and Work-readiness skills
The MYDev theory of change model is depicted in the following graphic.
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 2
Figure 1: MYDev's Theory of Change
These three outcome areas that serve as the basis for this evaluation study, align with three of MYDev’s 
key indicators and are investigated through three evaluation questions that this impact evaluation seeks 
to answer:
Q1. Do youth participants demonstrate improved livelihoods as a result of project activities? 
 MYDev M&E Indicator 12: Number of OSY receiving new or better employment / livelihoods (FAFI 
4.6.3-2: Number of persons receiving new or better employment (including better self-employment) as 
a result of participation in USG-funded workforce development project. 
Q2. Do youth participants demonstrate improved perceptions of government and community as a 
result of project activities?
 MYDev M&E Indicator 10: Percentage of targeted OSY with improved perceptions of community and 
government.
Q3. Do youth participants demonstrate improved life, work readiness, and leadership skills as a result 
of project activities?
 MYDev M&E Indicator 14: Percentage of OSY with improved life skills, work readiness, and leadership 
skills for civic engagement.
II. METHODOLOGY
A. Study Design
In order to understand MYDev’s contribution to improving OSY community and government perceptions, 
skills, and employment outcomes, MYDev designed and implemented a quasi-experimental study with 
both a pre-test and post-test for the intervention groups and comparison groups. The choice of a quasi￾Train learners in ALS 
instruction and life skills.
Train OSY in skills and life 
skills.
OSY complete skills training.
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 3
experimental approach has allowed the MYDev team to avoid actively excluding certain youth from 
participation in MYDev. This was has been important to avoid some youth feeling further marginalized, 
many of these youth are already routinely excluded in a variety of ways. Instead, admission into MYDev’s 
training programs has remained open to any OSY in participating geographic areas who was a member of 
MYDev’s target population and who wanted to participate, with comparison group youth consisting of 
similar youth from neighboring barangays.
Intervention and comparison group youth were assessed at baseline and four to six months after the 
intervention group completed training.
B. Sample
The use of a quasi-experimental approach means that the equivalence of the comparison and intervention 
groups cannot be assumed (there may be some observable and unobservable differences between youth 
in our comparison and intervention groups). However, care was taken to choose comparison group youth 
who were very similar to intervention group youth in key characteristics:
 Sex
 Age
 Education Status
 Employment Status
 Place of Residence
Intervention group youth who participated in the impact evaluation were randomly selected from the 
population of OSY who began MYDev training after January of fiscal year 2015 (FY15) as data collection 
tools were undergoing revision until that point. To account for a high expected rate of attrition from the 
sample, a sample of 1,264 youth, stratified by sex and LGU, was randomly selected from the FY2015 
participants. A much smaller sample of 416 comparison group youth was drawn as well, also stratified by 
sex and LGU. These sample sizes were chosen to allow for the detection of even small changes over time 
or small differences between groups, sometimes referred to as a small effect size. While this sample size 
Figure 2: Study Design
Comparison
Group
Intervention 
Group
Baseline data 
collection (pre)
Endline data 
collection (post)
Study Design for Each Cohort
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 4
allows for comparisons between males and females, it does not allow for comparisons between males 
and females by LGU.
In order to allow for findings of this evaluation—which is based on sample data—to be generalized to the 
population of MYDev youth who participated in trainings beginning in February of FY15, the sample data 
was weighted to this population. Weighting allows for the sample—which did not reflect key 
characteristics of the population such as the percent of males and females or the percent from rural/peri￾urban/urban settings—to be re-adjusted so that the composition of the sample reflects the composition 
of the population in key ways. In the MYDev context, where a stratified sampling strategy meant that an 
equal number of males and females were sampled from each area, the composition of the sample in terms 
of sex and being from a rural/peri-urban/urban setting did not reflect the population. Thus, weights were 
applied to the sample based on the sex and rural/peri-urban/urban setting composition of the population. 
This allows the MYDev team to argue that findings represent the entire population. Both the intervention 
and comparison groups were weighted to the population of FY15 MYDev youth who participated in 
MYDev trainings during the time period in which we implemented our data collection surveys.
C. Data Collection 
Data was collected from intervention group youth both within the first week of their participation in 
MYDev training and 4-6 months after they graduated from their training program. Since MYDev youth 
begin and graduate from their trainings at different points throughout the fiscal year, which runs from 
October 1 through September 30, baseline data was collected monthly throughout the entire year. The 4-
6 month lag between graduation and endline data collection was intentional: it would allow youth the 
time necessary to realize the expected outcomes that we intended to measure. 
Comparison group youth were assessed at the beginning of the fiscal year—in October and November—
with an endline survey implemented 4-6 months later in March and April. 
D. Tools
Data was collected from youth using three tools: the Youth Employment Survey (YES), the Youth 
Perceptions Survey (YPS), and the Development Assets Profile (DAP). While these three tools were 
implemented with MYDev youth in FY14, this initial use in MYDev revealed the necessity of considerable 
revision and adaptation of the tools. As a result, a period of testing and revision was undertaken from 
October of 2015 through January of 2015. 
The YES was developed and validated in previous EDC projects, and then adapted for use in MYDev. This 
process of adaptation included translation, back translation, piloting, and revision to the tool to ensure its 
usefulness for the MYDev context. In addition, a new section was added to measure youth’s life, work￾readiness, and leadership skills. To establish the reliability of this new section, principal component 
analysis was implemented, revealing the tool to be relatively reliable with a Cronbach’s alpha of .85 for 
work readiness, .81 for leadership, and .74 for life skills1
(see Table 1).
The YPS was developed specifically for the MYDev context with the goal of measuring youth’s perceptions 
of their governments and their communities. Principal component analysis revealed that both dimensions 
measured—perceptions of community and government—were reliable, with α=0.90 and α=0.91, 
respectively (see Table 1). 
 
1 A level of .70 or higher is viewed as acceptable for Cronbach’s alphas, though researchers prefer to see alphas of .85 or higher.
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 5
Tool Dimension Cronbach’s 
Alpha
How 
Reliable is 
it?
Youth Employment 
Survey
Work Readiness 0.85 Good
Leadership 0.81 Good
Life Skill 0.74 Acceptable
Youth Perceptions 
Survey
Perceptions of 
Community
0.90 Excellent
Perceptions of 
Government
0.91 Excellent
Table 1: Reliability of the YES and YPS
The DAP2 was developed by the Search Institute to allow organizations to understand the youth they serve 
from two perspectives: youth’s internal/external assets and youth’s contexts. The DAP has been tested 
throughout the world and proved reliable. Reliability analysis of the DAP in the MYDev context revealed 
the tool to be reliable in this context as well (see Table 2).
The reliability of these three tools—as revealed in principal component analysis using baseline data—
suggests that these tools are measuring consistently each time they are implemented in this context.
 
2 The DAP was contextualized and used successfully under EDC’sUSAID EQ2 project and further contextualized for MYDev.
Asset/Context Category
Cronbach’s 
Alpha
How Reliable 
is it?
External 
Assets
Whole External Asset Category 0.918 Excellent
Support 0.778 Acceptable 
Empowerment 0.726 Acceptable
Boundaries and Expectations 0.840 Good
Constructive Use of Time 0.600 Questionable
Internal 
Assets
Whole Internal Asset Category 0.925 Excellent
Commitment to Learning 0.802 Good
Positive Values 0.794 Acceptable
Social Competencies 0.766 Acceptable
Positive Identity 0.712 Acceptable
Contexts Personal 0.814 Good
Social 0.849 Good
Family 0.855 Good
School 0.851 Good
Community 0.830 Good
Table 2: Reliability of DAP in MYDev Context
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 6
E. Limitations
While this impact evaluation was implemented with attention to rigor and detail, some limitations persist:
 Claims to Causality. While this quasi-experimental impact evaluation is relatively rigorous, it still 
cannot claim the rigor of an RCT. Thus, the findings in this evaluation cannot be claimed to be caused 
by MYDev’s programming. 
 Comparison Group Weighting. As data on the sex and geographic area composition of all out-of-school 
youth in Mindanao was not available, comparison group youth were weighted to the population of 
MYDev Cohort 2 youth. This means that, regardless of the reality, comparison and intervention groups 
will seem statistically comparable in terms of sex and location.
 Contamination of the Comparison Group. Comparison group youth—selected from within the same 
LGUs as intervention group youth—may have experienced positive spillover effects from any 
encounters with MYDev youth within their LGUs or from the work of OSYDAs in improving the 
environment for OSYs. Further, other projects—such as USAID’s ENGAGE project, which specifically 
targets civil society organizations with civic engagement strengthening activities that include OSY—
are operating in the same areas as MYDev. While MYDev youth are likely not participating in ENGAGE 
activities since the two projects require participation at many of the same times, comparison group 
youth may be participating in these projects. Since ENGAGE is focused on improving civic engagement, 
measures of civic engagement with comparison group youth may be biased upward if comparison 
group youth are ENGAGE participants. This would bias downward the estimate of MYDev youth’s gains 
relative to the comparison group.
 Different Timing of Data Collection for Intervention and Comparison Groups. MYDev’s trainings have 
rolling admission, which means that baseline surveys were implemented on a rolling basis as well, 
whereas comparison group baselines were implemented at one point in time. If time of year has a 
differential effect on skills, perceptions, or employment outcomes, observed differences between 
intervention and comparison group youth may be artificial. For example, perhaps comparison group 
youth were surveyed during an election cycle that made them feel unheard, while intervention group 
youth were surveyed at all points in the year, including this one. In this case, intervention group youth 
as a whole would appear to have a better perception of the government at baseline, even if this 
weren’t true.
 Timing of Baseline Data Collection for Intervention Group. Baseline data collection is implemented 
within the first two weeks of training, during which youth are taking a life skills course that encourages 
them to begin to develop skills such as interpersonal and leadership skills. Depending on scheduling, 
youth may take baseline surveys on day two or day eight of training, after they may already have 
started to benefit from the effects of training. If this is the case, then baseline data would be biased 
upward, making the intervention group appear better off at baseline than the comparison group and 
minimizing the apparent gains of the intervention group between baseline and endline.
III. DEMOGRAPHICS
The weighted sample of intervention group youth that this study examines was primarily male, with just 
40.3% of the sample female. The sample was relatively evenly divide between urban (34.8), peri-urban 
(27.3%), and rural youth (37.9%). The largest group of youth was from Cotabato (21%), the next largest 
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 7
from Indanan (20%), with the fewest youth coming from Marawi (8%). The average age of intervention 
group youth was 19.5 years, with a minimum age of 14 and a maximum age of 25. The majority (68.8%) 
of intervention group youth had participated in MYDev’s skills training programs, with the rest (31.2%) 
participating in ALS training.
Comparison group youth were statistically equivalent to the intervention group youth in age, sex 
composition, geographic area, and age. This suggests that—at least in terms of these characteristics that 
one can observe—the intervention and comparison groups can be assumed to be equivalent at baseline 
and thus provide appropriate comparisons against each other.
While the intervention and comparison groups were equivalent in terms of basic demographics, a baseline 
study revealed that intervention group youth at baseline were more likely to be employed, were more 
able to achieve their goals with their savings, and ate more meals per day on average. This suggests that 
comparison group were more economically vulnerable at baseline than were intervention group youth.
 
IV. IMPACT FINDINGS and OUTCOMES for MYDev YOUTH 
A. Employment
A central intended outcome of MYDev and OSYDA programming is improved employment outcomes for 
out-of-school youth. All analyses point to success for MYDev regarding improving the employment 
situations of MYDev youth. Across the course of the intervention, 24% of MYDev youth obtained new 
employment, statistically more than the percent of the comparison group (14%) who gained new 
employment at endline. At endline, more MYDev youth were working relative to comparison group youth
regardless of their employment status at baseline. Further, almost half (42%) of the roughly 19% of MYDev 
youth who were working at both baseline and endline had improved their employment by endline, 
constituting 8.2% of the sample with improved employment. 
This rest of this section builds on the above analysis, examining both the rates of new employment among 
17%
9% 8% 10% 8%
18% 19%
11%
21%
8%
0% 7%
12% 14%
20% 18%
Cotabato Isabela Jolo Lamitan Marawi Zamboanga Indanan Parang
Area, by Group Intervention
Comparison
Figure 3: Area, by Group
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 8
MYDev youth as well as the quality of employment for both newly employed youth and youth who already 
had work at baseline. In addition, newly and better employed MYDev youth are compared to comparison 
group youth in an effort to examine the MYDev project’s possible impact on its participants.
1. New Employment
Profile of A Newly Employed MYDev Youth
About 24% of MYDev youth obtained employment over the course of the intervention. What do these 
youth look like on average? Who are they, and what are their work lives like?
Demographics. First, newly employed MYDev youth were more likely to be male than female, with 27% of 
all males in the sample obtaining new employment compared to 19% of all female youth in the sample 
(significantly different at p<.05). These findings may suggest that one’s sex may influence their 
employment opportunities or their ability to take advantage of them. Youth’s geographic area may also 
influence their employment opportunities. Newly employed MYDev youth were more likely to be from 
highly urban areas (44.6%) as compared to partially urban areas (20.4%) or rural areas (35%). Interestingly, 
more MYDev youth from rural areas were employed than from partially urban areas, which perhaps 
indicates that the relationship between employment outcomes and geographic area is complex, and 
merits further investigation to better inform future MYDev programming.
Work Lives. Most MYDev youth obtained new employment in either the business/enterprise (26%) or 
transportation services (21%), though male and female youth did not enter these sectors at the same 
rates. Male MYDev youths entered the transportation services sector at higher rates (31%), while female 
MYDev youths entered the business/enterprise sector at higher rates (36%). These differences may be 
worth exploring further to determine whether there are opportunities for programming to promote 
gender balance within different industrial sectors.
24.0%
8.2%
NEW AND BETTER EMPLOYMENT 
ACROSS THE SAMPLE
Newly Employed Improved Employment
19.5% of all youth had 
employment at both 
baseline and endline, and 
42% of these youth
improved their 
employment, which 
represents 8.2% of the 
70% of youth did not have 
employment at baseline, 
and 34.3% of these youth 
gained employment by 
endline, which represents 
24% of the sample
Figure 4: New and Better Employment
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 9
The average newly employed MYDev youth is more likely to have only one job, with only about 16% of 
MYDev youth indicating that they had 
more than one job. Newly employed 
MYDev youth are also much more 
likely to be paid monetarily (97%) 
rather than through in-kind payment. 
Newly employed MYDev youth are 
more likely to be employed in 
temporary or substitution positions 
(53.5%) rather than permanent or 
contractual positions (19.6%) or than 
own their own business (21.8%). 
On average, newly employed MYDev
youth work about 42 hours per week. 
Female MYDev youth work almost ten 
hours more than male youth on 
average, with female youth working 
about 48 hours per week and male 
youth working about 39 hours per 
week, on average, and the difference 
in the number of hours worked per 
week is statistically significant at 
p=.0406.
Quality of Employment. About half (51.1%) of newly employed MYDev youth were only sometimes able 
to save, with about 40.3% indicating that they were never able to save, and only 8.6% always able to save 
money. This perhaps indicates that even though MYDev youth have obtained employment in the period 
leading up to endline, they are underemployed and their employment is insufficient to satisfy their 
needs. MYDev youth from rural areas are more likely to be able to meet their needs from their earnings 
than youth from other highly urban areas, which perhaps demonstrates that future MYDev programming 
should further investigate the characteristics of rural economies and job markets. Even though the 
average MYDev newly employed youth is unable to save money frequently, most MYDev youth (59.3%) 
save enough to meet their needs or goals, with female youth responding more positively about their 
savings than male youth. 
About half of newly employed MYDev youth had their earnings increase (57.9%) or savings increase at 
least a little (53.6%) in the six-month period leading up to endline. Additionally, almost all (81.4%) newly 
employed MYDev youth ate three or more meals per day indicating that their earnings or savings are 
sufficient to meet their nutritional needs, though this is not statistically different from baseline. While 
MYDev does provide youth with financial fitness training, these findings in conjunction with those that 
indicate that most MYDev youth are never able to save money suggest that additional focus on developing
money management skills or savings strategies may be beneficial to MYDev youth.
MYDev youth have more varied responses when it comes to job safety and job satisfaction. A small 
majority (58.6%) of newly employed MYDev youth feel safe at their jobs and only 7% feel very safe, while 
Figure 5: Profile of a Newly Employed Youth
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 10
about 32% feel very unsafe. Female youth reported feeling safer performing their jobs than male youth 
(p=.007), which perhaps indicates that gender influences the types of jobs youth perform as well as the 
industries they enter. On average, newly employed MYDev youth are satisfied with their jobs with 57.3% 
responding that they were either happy or very happy at work. 
2. Impact Findings
While the above profile of a MYDev youth reveals much about the work lives of MYDev youth after 
participating in MYDev, it does not speak to the impact of MYDev’s programming. While the study design 
does not allow us to categorically claim impact, in order to better estimate the effects of MYDev and 
OSYDA programming on participating youth, the following analysis compares participating (intervention 
group) youth to comparison group (non-participating) youth. The differences between MYDev
intervention and comparison group youth are highlighted in this section; any differences may suggest that 
it is MYDev programming that affected outcomes.
MYDev youth obtained new employment at a higher rate than did comparison group youth, with almost 
24% of intervention group youth newly employed at endline compared to about 14% of youth in the 
comparison group. This approximate 10 percentage-point difference in the rates of newly employed 
youth between the intervention and comparison groups is significant at p<.001, and it gives a moderate 
effect size of .39. These findings suggest that MYDev may have been successful in improving the 
employment outcomes of MYDev youth. Similarly, as shown in Figure 5 below, where all employed youth 
at baseline and endline, respectively, are represented, MYDev youth made significant gains (11.3 
percentage points) in employment relative to the comparison group. Further, though more of the 
employed MYDev population is male and from urban areas, when these factors are held equal across both 
the intervention and comparison group, more MYDev youth than non-MYDev youth still gained 
employment by endline.
15.0%
20.0%
25.0%
30.0%
35.0%
40.0%
45.0%
50.0%
Baseline Endline
Employment Gains 
(% Working at Baseline and Endline)
Intervention Group
Comparison Group
Positive (11.3 percentage points) 
and Significant (p<.001)
Gain for the 
Intervention Group 
Relative to the
Comparison Groupe
Figure 6: Difference in Difference Estimate of Employment Gains
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 11
Furthermore, when controlling for other characteristics such as sex and geographic area, the treatment 
status, meaning youth’s participation in MYDev programming, remained significant (p<.001) with MYDev
youth more likely than comparison group youth to have obtained new employment. Interestingly, sex also 
remained significant indicating that female youth, regardless of their treatment status or geographic area, 
obtained new employment less often than male youth (p<.001). 
This further emphasizes the 
influence of gender norms on 
employment opportunities and 
youth’s employment choices. The 
findings also show geographic area 
was correlated with obtaining new 
employment: youth from partially 
urban and rural areas obtained 
new employment at lower rates 
than youth from highly urban areas 
(p=.001). Specifically, youth from rural areas obtained new employment at lower rates (about 15% less) 
than youth from highly urban environments (regardless of youth’s sex or treatment status). 
While MYDev youth were more likely to have new employment than non-MYDev youth, the quality of 
employment of these newly employed MYDev youth was not statistically different from that of non￾MYDev youth. An index that sums all quality of employment variables into a value ranging from a 
minimum value of 0 (worst quality employment) to a maximum value of 8 (best quality employment) was 
built to measure quality of employment. Newly employed MYDev youth scored higher than comparison 
group youth, though not significantly differently, with MYDev youth scoring 3.00 and comparison group 
youth scoring 2.94. 
3. Previously Employed MYDev Youth
In addition to improving employment outcomes for out-of-school youth through new employment, 
MYDev and OSYDA programming aims to improve employment outcomes for previously or currently 
employed participating youth. This section presents findings that describe MYDev youth employed at the 
time of baseline and endline survey administration and highlights the changes in their employment 
situations. Specifically, this analysis takes a comprehensive view of the changes in different employment 
characteristics, including change in job type, change in any of eight quality of employment characteristics, 
and the interaction between the amount worked (number of jobs and number of hours worked) and 
changes in the ability of a youth’s earnings to meet his/her needs. The combination of these different 
characteristics is summed into an index so that each youth received a single numeric score to represent 
the quality of their employment.
Profile of a Previously Employed MYDev Youth
Youth who were employed at baseline still could benefit from MYDev training related to technical and 
non-technical skills. So, what did these youth—who began the program with some form of employment—
look like at endline? Who are they, and what are their work lives like?
Demographics. Approximately 19.5% of MYDev youth were employed both at the time of baseline and 
endline survey administration. Male MYDev youth were more likely to be previously employed and 
44.01%
21.97% 23.39%
11.71%
Male Female
Intervention
Comparison
Figure 7: Employment, by Group and Sex
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 12
maintain employment over the course of the intervention than female youth: 25.8% of males were 
previously employed compared to 10.0% of female youth, and this difference is significant at p<.001. This 
finding indicates that gender may influence not only the ability of a youth to obtain employment, but also 
to maintain it. A youth’s geographic area is also influential with most of the previously employed MYDev
youth from highly urban areas (about 51%) with only about 21.5% from rural areas. The difference in the 
rates of continuous employment between youth from highly urban and rural areas is significant at p<.001, 
further indicating the need for MYDev and workforce development programming more broadly to be 
tailored to the rural labor market context. 
Work Lives. The average MYDev youth is likely to have only one job, which is probably an unstable, 
temporary position, and this youth did not change his/her job type between baseline and endline. At 
endline, about half (49.9%) of previously employed MYDev youth worked in temporary positions with 
32.7% working in permanent or contractual positions. The average previously employed MYDev youth 
likely did not change to either a better or worse job type over the course of the intervention. Indeed, 
findings show about 87% of youth stayed in the same type of job3
. At endline, about three-quarters of 
previously employed MYDev youth worked only one job, meaning that on average, previously employed 
MYDev youth maintained the same number of jobs between baseline and endline. 
Quality of Employment between Baseline and Endline. Almost half (42%) of working youth (youth with 
jobs at both baseline and endline) experienced an overall improvement in the quality of their 
employment between baseline and endline. This overall improvement is based on an index that takes 
into consideration improvements in job type, job safety, job satisfaction, earnings, and savings as 
described above. While almost half of MYDev youth employed at both baseline and endline improved 
their employment overall, none statistically improved their quality of employment when these quality 
indicators were examined individually. Female MYDev youth and youth from rural areas were more likely
 
3 During the administration of both baseline and endline surveys, youth were asked to report which job category best 
matches their primary job. They were able to choose between 1) permanent/contractual, 2) temporary/substitution, 3) 
internship/apprenticeship, and 4) owning their own business. Among these four options, permanent/contractual and 
business ownership are considered the better job types (but are equal to each other), followed by 
internship/apprenticeship, and then temporary/substitution. Consequently, this analysis is able to consider the change 
between different job types from baseline to endline survey administration as an aspect of better or worsening 
employment.
Figure 8: Profile of A Continuously Employed MYDev Youth
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 13
to see an increase in the number of meals they ate per day than male youth and youth from highly urban 
areas, respectively (p=.015 and p=.044). 
MYDev youth who experienced changes over the course of the intervention in the number of jobs they 
held and in their employment type were also more likely to report positive changes in several quality of 
employment aspects. For example, when holding constant4 demographics such as sex and age, a MYDev
youth with an increase in the number of jobs was more likely to report an increase in earnings (p=.0240). 
Yet, a MYDev youth who worked more hours at the end of 
the intervention period was less likely to have seen an 
increase in his/her savings over the course of the 
intervention. This indicates that improved employment 
outcomes are important to be framed within the context 
of better work, and not simply more work. More work is 
not necessarily better work. This is further emphasized by 
the findings that youth who did change their job type to a 
better job type, such as from a temporary to a permanent 
position, were more likely to have increased savings than 
those with no change or with a change to a worse job type 
(p=.022). 
Impact Findings
In order to better estimate the impact of MYDev and 
OSYDA programming on participating youth, this analysis 
also compares participating to comparison group (non￾participating) youth. The differences between MYDev and 
comparison group youth are highlighted in this section. 
About 42.2% of MYDev youth and 44.4% of comparison 
group youth had better employment at endline. The 
difference between the rates of better employment of 
MYDev and comparison group youth are not statistically 
significant. Interestingly, when controlling for other characteristics, there were no correlations found for 
gender or geographic area, meaning that, for example, female youth had the same rates of better 
employment, on average, as male youth. This may suggest that—while gender matters for getting a first 
job--it may matter less for improving one’s work situation.
An Aggregate Look into New or Better Employment 
While it is clear that MYDev youth experience more positive new employment outcomes and equivalent 
quality of work outcomes when compared to non-MYDev youth, this study also examined a cumulative 
measure capturing both newly employed youth and youth with better employment. In the intervention 
group, about 32.2% of youth reported new or better employment at endline compared to only 17.9% 
 
4 We know already that sex and geographic area change an OSY’s employment story. However, in many cases, we want 
to investigate how employment operates in OSY’s lives separately from how sex or geographic area operates. In these 
cases, we statistically control for these demographic variables. When we control for these other characteristics, we can 
isolate how an employment or other variable of interest operates regardless of the different effects of these 
characteristics.
Better Employment Index
Variable Change from 
Baseline to 
Endline
Job Type -1, 0, 1
Job Satisfaction -1, 0, 1
Job Safety -1, 0, 1
Earnings 
Change per 
hour worked
-1, 0, 1
Savings 
Frequency
-1, 0, 1
Savings 
Increase
-1, 0, 1
Savings Meet 
Needs
-1, 0, 1
Meals 
Eaten/Day
-1, 0, 1
Total Index 
Scale
-8 through 8
Figure 9: Index for Better Employment
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 14
of comparison group youth. The difference in rates of new or better employment is statistically significant 
at p<.001. When controlling for other characteristics, in addition to treatment status, several interesting 
relationships appear. On average, female youth were about 30% less likely to have new or better 
employment than male youth (p=.001). Geographic area was also a significant influence on new or better 
employment with youth from partially urban areas 29.3% less likely and youth from rural areas 34.4% less 
likely than youth from highly urban areas to have new or better employment (p=.011, p=.001). Thus, males 
and highly urban youth are better able to improve their employment situations than other groups.
4. Unemployed Youth
In order to better achieve the positive impact of MYDev and OSYDA programming, it is important to 
identify the group of MYDev youth who are unemployed. This section analyzes the rates of unemployment 
among MYDev youth as well as the factors that may influence unemployment in order to inform future 
MYDev programming. 
Profile of an Unemployed MYDev Youth
Without taking into consideration baseline 
employment status, about half (56.6%) of MYDev
youth were unemployed at the end of the MYDev
intervention period. Of these youth who were 
unemployed at endline, 81.1% were unemployed 
at baseline as well. Thus, about 45.9% of MYDev
youth were unemployed both at the beginning and 
end of the intervention. What do these youth who 
did not have work at endline look like? And how are 
they different from employed youth?
Demographics. While female MYDev youth were 
more likely than male youth to be unemployed 
throughout the course of the intervention, (with 
62.4% of intervention group females being unemployed at both baseline and endline, significant at 
p<.001), the group of youth who was not working at endline was evenly split between males and females. 
A continuously unemployed MYDev youth is more likely to be from a rural area as well as female, with 
about 48.2% of continuously unemployed youth being from rural areas and 44.7% of those without 
employment at endline being from rural areas. The higher rates of continuous unemployment among 
these groups perhaps demonstrate that MYDev and OSYDA programming should investigate ways to 
further impact to improve employment outcomes for these groups.
Work Life. The average unemployed MYDev youth is currently looking for a job and is unhappy with his/her 
unemployment situation. 74.3% of MYDev youth reported that they are currently looking for a job. Male 
MYDev youth are more likely to be looking for a job than female MYDev youth, with over 78.7% of males 
looking compared to about 70.0% of females, and this difference is significant at p=.0469. Of the 
remainder of youth that replied that they are not currently looking for a job, the most common reasons 
given for not looking were that youth felt “There are no jobs” (28.2%), that youth were in school (22%), 
and that youth feel they “Do not have the skills necessary” (18.5%). 
Figure 10: Profile of an Unemployed Youth
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 15
When unemployed youth in the 
intervention group were asked whether 
they were happy with their present 
situation, the majority of youth reported 
being either unhappy (53.2%) or very 
unhappy (32.1%) with the situation. 
5. Conclusions from 
Employment Study
This study of MYDev’s employment 
outcomes can conclude that MYDev
youth are better off than non-MYDev
youth in terms of having gained more 
new employment, though among working MYDev and non-MYDev youth, the quality of employment is 
statistically similar. Further, new and continuously employed youth tend to be urban males working in 
one temporary/substitution job at a time, while continuously unemployed youth tend to be rural females. 
Yet beyond demographic differences between employed and unemployed youth, other factors such as 
life skills, work readiness skills, and leadership skills have been linked to employment outcomes. The next 
section examines these skills.
B. Life Skills, Work Readiness, and Leadership Skills 
MYDev’s approach to OSY programming is characterized by a holistic perspective of productive youth 
development, focused on developing life, leadership, and work readiness skills to enable OSY to more 
productively participate in their communities. This section examines the changes in MYDev youth’s life, 
leadership, and work-readiness skills and whether and how much improvement occurred over the course 
of the MYDev intervention. 
A skills index was created in order to reliably measure each of the skills components. To construct this 
index, each question of the three skills components was assigned a response scale of one to four, with 
four reflecting the strongest or best response. The index was constructed by taking the average of points 
within each component at baseline and at endline. The net gains are analyzed below and findings are 
presented to highlight skills improvement among MYDev youth.
1. Skill Gains for MYDev Youth
Almost 70% of MYDev youth improved in in either life, leadership, or work-readiness skills over the 
course of the MYDev intervention. There were no significant differences in improvement between male 
and female youth, between youth from different geographic areas, or between those that did and did not 
have new or better employment outcomes. 
Gains in Life Skills for MYDev Youth
The Life Skills component encompasses four questions measuring skills like goal-setting, communicating 
with a variety of people, and speaking clearly and is measured on an average-based index ranging from 0 
to 4. The average MYDev youth improved about .10 points along this index, with 46.7% of MYDev youth 
improving in Life Skills between baseline and endline (significant at p=.002). At endline, the average Life 
Skills score for MYDev youth was 2.84. Female MYDev youth improved less than their male counterparts 
with an average gain of only .05 points on the Life Skills Index (p=.006). 
The Top Three Reasons Why Some Unemployed 
Youth are NOT Looking for Work
1. There are no jobs (28.2% of the sample)
2. I am in school (22% of the sample)
3. I do not have the skills necessary (18.5% of 
the sample)
3=I have a lot of 
knowledge about 
this and can do 
this correctly and 
consistently
2=I have enough 
knowledge about 
this and can do 
this correctly 
sometimes.
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 16
Gains in Work Readiness Skills for MYDev Youth
The Work Readiness component encompasses eight questions measuring skills like solving conflicts at 
work, looking for a job, and working towards a pay upgrade, and is measured along an average-based 
index from 0 to 4. About 55.6% of intervention group youth improved in work readiness skills, with the 
average MYDev youth improving about 0.16 points along the index. At endline, the average work 
readiness score for MYDev youth was 2.78. Female MYDev youth improved less than their male 
counterparts with an average gain of about .06 points on the Work Readiness index (p=.009). In contrast, 
MYDev youth that obtained either new or previous employment over the course of the intervention 
averaged about .14 points higher than unemployed. 
These findings reaffirm the relationship between work readiness skills and employment outcomes.
While it is not possible to determine causality, these findings, combined with other findings, suggest that 
the challenges facing female youth in obtaining employment may similarly impede female youth from 
improving in work readiness skills. Many of the skills covered in the Work Readiness skills index can be 
practiced through employment experience, which female youth have not had access to at the same rates 
as male youth. 
Gains in Leadership Skills for MYDev Youth
The Leadership Skills component encompasses two questions measuring leadership skills at work and in 
the community, and is measured along an average-based index from 0 to 4. About 43% of MYDev youth 
improved in leadership skills over the course of the intervention, with the average MYDev youth 
improving about .20 points along the index. At endline, the average leadership skills index score for 
MYDev youth was 2.58. While one might assume that improving one’s leadership skills would spill over 
into improved life or work readiness skills, there were no correlations in the gains in leadership skills with 
other characteristics such as work readiness and life skills. This suggests that programming targeted at 
improving each skill area—leadership, work readiness, and life skills—is important as youth do not 
automatically transfer skill improvement in one area to skill improvement in another.
2. Impact Findings
Findings that examine the intervention group alone cannot speak to the effects of the MYDev project. 
Only comparing intervention group outcomes to comparison group outcomes over time will allow us to 
examine the differential outcomes that may have occurred with the intervention. This examination, 
however, does not produce positive findings for MYDev youth compared to comparison group youth. For 
Figure 11: Skills Gains for MYDev Youth
2.62
2.72
2.39 2.78
2.84
2.58
Work Readiness Life Skills Leadership Skills
3=I have a lot of Baseline Endline
knowledge about this 
and can do this correctly 
and consistently
2=I have enough 
knowledge about this 
and can do this correctly 
sometimes.
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 17
example, a statistically similar proportion of intervention group and comparison group youth improved 
their skills across the study period, with about 69.8% of the intervention group and about 70.0% of 
comparison group youth improved in either life skills, work readiness or leadership skills over the course 
of the MYDev intervention. The difference in the rates of improvement between MYDev and comparison 
group youth is not statistically significant. When controlling for other characteristics, there were no 
correlations found. 
Further, when life and work readiness skills are examined separately, the comparison group’s gains are 
statistically greater than the intervention group’s gains from baseline to endline, though these gains are 
small. Specifically, comparison group youth improved their life skills scores by .22 points and their work 
readiness scores by .25 points. Intervention group youth, on average, increase their life skills scores from 
baseline to endline by .1 point and their work readiness scores by .16 points, and these differences are 
statistically significant by at least p=.05. Yet, while gains in the intervention group were smaller than gains 
in the comparison group, intervention group youth’s endline scores were higher than the comparison 
group’s endline scores.
C. Government and Community Perceptions 
MYDev programming aims to improve youth’s engagement within their communities, while also 
improving youth’s perceptions of their communities and governments. This section analyzes the changes 
in youth’s perceptions of their community and government over the course of the MYDev intervention. 
Separate indices were created for each component, in which points were allocated based on a scale with 
the number of points matching the answer chosen, meaning someone who chose one (“I Don’t Know”) 
received one point.5
1. MYDev Youth’s Government and Community Perceptions
The majority of MYDev youth had improved their government or community perceptions by the time 
of endline administration. In fact, 66.7% of MYDev youth had improved government or community 
perceptions. There were no significant differences in improvement between male and female youth, 
between youth from different geographic areas, or between those that did and did not have new or better 
employment outcomes.6
Changes in Government Perceptions
The Government Perceptions Component encompasses twelve questions measuring youth’s attitudes 
towards government and civic participation. Change in government perceptions scores are measured 
along an average-based index ranging from 0 to 5. The average MYDev youth had a 0.22 score increase 
along the index between baseline and endline. At endline, the average MYDev youth scored a 3.17 on the 
index. There were no significant correlations found between increase in government perception scores 
and other characteristics. However, there were differential effects based on sex and geographic area.
Female MYDev youth scored about .16 points lower at endline than did male youth (p=.02). Similarly, rural 
youth scored .26 points lower than youth from highly urban areas at endline (p=.004). However, these 
 
5 Within the Youth Perceptions Scale “I don’t know” is interpreted as the out-of-school youth is not aware 
of the situation or is detached or indifferent about his/her government or community. The survey design 
consequently rates “I don’t know” as the lowest level of perception. 
6 The multivariate analysis (regression) uses variables from different surveys, and consequently, a sample 
weight that accounts for all three survey samples. 
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 18
differences between groups at 
endline are statistically similar to the 
differences at baseline, suggesting 
that geographic area and sex did not 
negatively affect perception 
improvements across the course of 
the MYDev intervention. Thus, one 
could argue that the MYDev 
curriculum was similarly effective 
across different geographic areas 
and with both male and female 
youth.
Changes in Community Perceptions
The Community Perceptions 
Component encompasses twelve 
questions measuring youth’s 
attitudes towards their communities 
and community organizations. Change in community perceptions scores are measured along an average￾based index ranging from 0 to 5. The average MYDev youth had a .24 point increase in community 
perceptions along the index, with an average score of 3.42 at endline. There were no significant 
correlations found between increase in community perception scores and other characteristics. At 
endline, the average MYDev female youth scored .15 points lower than the average MYDev male youth 
(p=.017). Additionally, the average MYDev youth from rural areas scored about .27 points lower than the 
average MYDev youth from highly urban areas (p=.001). 
2. Impact Findings
Statistically, a similar proportion of intervention and comparison group youth improved in their 
community or government perceptions: about 69.9% of the comparison group improved while 66.7% of 
MYDev youth improved their perceptions on at least one question overall though the difference is not 
statistically significant at conventional levels. When controlling for other characteristics including sex, 
geographic area, and whether or not youth had new or better employment as well as treatment status, 
no significant correlations were found with perception improvements. 
An examination of youth’s perceptions of community and government respectively shows that 
comparison group youth’s perceptions of the government improved (by more than half a level on average 
(.6244)) more than intervention group youth’s perceptions (which improved by less than a quarter of a 
level (.2168)). Similarly, comparison group youth’s perceptions of their communities improved (by more 
than half a level on average (.5797)) more than intervention group youth’s perceptions (which improved 
by roughly a quarter of a level (.2429)). All of these differences are statistically significant at p<.001.
2.95
3.17
3.17
3.42
BASELINE ENDLINE
Government 
Perceptions
Community 
Perceptions
Figure 12: MYDev Youth's Community and Government Perceptions 
Gains
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 19
D. Youth’s Assets and Contexts
MYDev and OSYDA programming aims to engage OSY holistically to enable OSY to become active members 
of their communities. A central aspect of this approach is a comprehensive and nuanced understanding 
of OSY’s contexts and self-perceptions. This section analyzes the changes in these aspects reported by 
MYDev youth over the course of the intervention. The tool used to capture these change is the 
Development Assets Profile (DAP), which measures youth’s assets and youth’s contexts through the 
summation of different questions into categorical indices (see Figure 9). These indices reflect a score out 
of 30, and the cutoff points for each level of assets and contexts score are 0-14 (challenged); 15-20 
(vulnerable); 21-25 (adequate); and 26-30 (thriving). 
1. MYDev Youth’s Assets
Internal Assets. The internal assets index encompasses questions relating to an individual’s commitment 
to learning, positive values, social competencies, and positive identity. The average MYDev youth had an 
internal asset score of about 21.0 at endline, which is at the lower end of the adequate bracket. While 
males and females’ endline scores were statistically the same, youth from highly urban areas assessed 
themselves as having more internal assets (21.6) than youth from rural areas (20.4), and this difference is 
significant at p=.013 and remains when sex is held constant across youth from different areas.
On average, MYDev youth improved about 1.3 points in their internal asset scores between baseline and 
endline. Male youth obtained higher, but statistically similar, gains compared to female youth, improving 
1.6 points compared to female youth’s .95 points. Highly urban youth improved more than the average 
youth with gains of 1.8 points; however, the difference in gains between youth of different geographic 
areas is not statistically significant. 
External Assets. The external assets index encompasses questions relating to an individual’s support, 
empowerment, boundaries and expectations, and the constructive use of time. At endline, the average 
MYDev youth had a score of 20.6 on the index, placing him/her at the edge of the vulnerable bracket and 
almost in adequate. On average, male MYDev youth’s external asset scores (20.9) were higher than female 
youth’s (20.2), further emphasizing the gender-based discrepancies seen in employment and skills gains 
findings (p=.0472). Youth from highly urban areas scored above average on the index with a score of 21.3, 
Figure 13: The DAP's Asset Categories and Contexts
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 20
placing them within the adequate bracket, and significantly higher than youth from rural areas, who scored 
19.9 on average (p=.002). 
On average, MYDev youth improved about 1.07 points on the external assets index between baseline and 
endline. Male youth and youth from highly urban areas had the greatest score gains over the course of the 
intervention. On average, male youth gained about 1.57 points, which is significantly different to the .36 
points gained by female youth (p=.0162). Youth from different geographic areas had significantly different 
score gains as well. Youth from highly urban areas gained approximately 2.02 points while youth from rural 
areas lost about .05 points over the course of the intervention (p=.002). When controlling for both sex and 
geographic area, both of these characteristics remain significant indicating the impact that a MYDev youth’s 
sex and geographic area have on their external asset scores.
2. MYDev Youth Contexts
Personal Context
The average MYDev youth scored 20.7 points on the personal context index at endline, placing him/her 
at the higher edge of the vulnerable bracket. There were no statistically significant differences in personal 
context scores between youth from different geographic areas or between male and female youth. 
On average, MYDev youth improved about .95 points on the personal context index between baseline and 
endline. Male youth improved more, but not significantly more, than female youth in their scores. Highly 
urban youth improved about 1.6 points, on average, which is significantly more than the improvements 
by rural youth, who improved by just .1 points (p=.0351). When controlling for both sex and geographic 
area characteristics, these significant differences remained. On average, female youth gained about one 
point less in their scores than did male youth (p=.047). On average, rural youth gained about 1.4 points 
fewer than did highly urban youth (p=.035). 
Social Context
The average MYDev youth scored about 20.9 points on the social context index at endline, placing 
him/her at the higher edge of the vulnerable bracket. Male and female youth scored similarly at endline. 
Youth from rural areas scored significantly lower with scores of about 20.4 points compared to youth from 
highly urban areas that scored about 21.5 points (p-.0138). The difference in scores between rural and 
highly urban youth remained significant when controlling for both sex and geographic area, with rural 
youth scoring about 1.6 points lower than highly urban youth (p=.002).
On average, MYDev youth improved about 1.8 points on the social context index between baseline and 
endline. Male youth improved significantly more than female youth with average gains of 1.07 points 
compared 2.38 points gained by female youth (p=.0425). When controlling for both sex and geographic 
area, the findings show that rural youth score about 1.9 points lower, on average, than youth from highly 
urban areas (p=.005). 
Family Context
The average MYDev youth scored about 23.0 points on the family context index at endline, which is within 
the adequate category. Female youth scored higher, though statistically similarly, than male youth. There 
were no differences in endline scores based on geographic areas. 
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 21
On average, MYDev youth gained about .5 points on the family context index over the course of the 
MYDev intervention. Male youth had significantly higher score gains than female youth, who actually lost 
about .09 points between their baseline and endline scores while male youth gained about .9 points 
(p=.0402). There were no differences in score gains based on geographic area. 
School Context
The average MYDev youth scored about 22.2 points on the school context index at endline, which is within 
the adequate category. There were no significant differences in endline scores based on geographic area 
or sex. 
On average, MYDev youth gained about .4 points on the school context index over the course of the 
intervention. There were no differences in score gains based on either sex or geographic area. 
Community Context
The average MYDev youth scored about 18.0 points on the community context index at endline, which 
is within the vulnerable bracket. Male youth scored significantly higher than female youth, with scores 
of 18.6 at endline compared to 17.0 (p=.0004). Youth from highly urban areas also scored significantly 
higher, though still within the vulnerable bracket, than youth from rural areas (19.2 compared to 16.7), 
and this difference is significant at p<.001. When controlling for both sex and geographic areas, these 
findings remain significant. Female youth scored, on average, 1.6 points lower than male youth (p<.001) 
and youth from rural areas scored, on average, 2.5 points lower than youth from highly urban areas 
(p<.001).
On average, MYDev youth gained about 1.6 points in their community context index scores. Male youth 
had significantly higher score gains, improving about 2.1 points on average compared to about .83 for 
female youth (p=.0321). There were no differences in score gains based on geographic area. 
3. Impact Findings
Assets
MYDev youth scored significantly higher than comparison group youth in both internal and external 
asset scores at endline. MYDev youth scored, on average, 21.0 points in internal asset scores and 
comparison group scored 18.4 points (p<.001). In terms of external asset scores, MYDev youth scored 20.6 
at endline and comparison group youth scored 17.8 points (p<.001). This places intervention group youth 
above in the adequate bracket and comparison group in the vulnerable bracket for both asset indices. 
When controlling for other characteristics, such as sex and geographic area, the differences in scores 
between MYDev and comparison group youth remain significant at the most conservative confidence 
threshold for both external and internal assets (p=.000). Geographic area similarly remains highly 
significant with youth from rural area scoring about 2.06 points lower on external assets and about 2.52 
points lower on internal assets than comparable youth from highly urban area, regardless of sex or 
treatment status (p=.000). These findings highlight the importance of the rural context in conjunction with 
the impact of MYDev programming. 
Comparison group youth improved more, but not significantly more, than intervention group youth 
gaining about 1.7 points on the internal asset score compared to the intervention group’s 1.3 points. 
Comparison group youth also improved more than MYDev youth in their external asset scores gaining 
about 1.7 points compared to the intervention group with gains of 1.1 points; however, the difference is 
not significant at conventional levels. When controlling for other characteristics, only geographic area is 
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 22
significant with youth from rural areas gaining improving less on average than youth from highly urban 
areas. For internal asset scores, rural youth gained about 2.2 points less than youth from highly urban 
areas with average gains of only 0.6 points (p<.001) For external asset scores, rural youth went down in 
scores between baseline and endline, and on average, scored 2.24 points lower than highly urban youth 
(p<.001). 
Contexts
Personal Context
MYDev youth scored significantly higher than comparison group youth at endline with scores of 20.7 
points compared to 17.9 points (p<.001).
However, MYDev youth and comparison group youth had similar score gains between baseline and 
endline though the comparison group improved only slightly more at 1.0 points compared to .95 points. 
When controlling for other characteristics, only geographic area remained significant with rural youth 
improving about 2.0 points less between baseline and endline than youth from highly urban areas 
(p=.001). 
Social Context
MYDev youth scored significantly higher than comparison group youth at endline with scores of 20.9 
points compared to 18.4 points (p<.001). When controlling for other characteristics, the treatment status 
remained significant. The findings also showed that youth from rural areas were more likely to score 2.6 
points lower than youth from highly urban areas at endline (p<.001). 
MYDev and comparison group youth had statistically similar score gains on the social context index. When 
controlling for other characteristics, only geographic area was significant with rural youth improving about 
2.1 points less between baseline and endline than youth from highly urban areas (p<.001). 
Family Context
MYDev youth scored significantly higher than comparison group youth at endline with scores of 23.0 
points compared to 19.7 points on the family context index (p=.000). When controlling for other 
characteristics, the treatment status remained significant with intervention group youth scoring, on 
average, 3.3 points higher than intervention group youth (p=.000). Interestingly, female youth were more 
likely to score about 1.0 point higher than male youth (p=.002). The findings also showed that youth from 
rural areas were more likely to score 1.8 points lower than youth from highly urban areas at endline 
(p<.001) and youth from partially urban areas were more likely to score about 1.0 points lower than youth 
from highly urban areas (p=.018). 
MYDev and comparison group youth had statistically similar score gains on the family context index. When 
controlling for other characteristics, only rural area was found to be significant with rural youth more 
likely to score about 1.7 points lower than youth from highly urban areas (p=.004). 
School Context
MYDev youth scored significantly higher than comparison group youth at endline with scores of 22.2 
points compared to 18.6 points on the school context index (p<.001). When controlling for other 
characteristics, treatment status remained significant. The findings also showed that youth from rural 
areas were more likely to score about 2.1 points lower than youth from highly urban areas at endline 
(p<.001). There were no gender-based correlations found. 
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 23
Comparison group youth had significantly higher score gains than did MYDev youth on the school context 
index. On average, comparison group youth improved about 1.6 points and MYDev youth improved .4 
points, and this difference is significant at p=.011. When controlling for other characteristics, the 
treatment status remains significant. Rural area is also significant with rural youth gaining on average 1.4 
fewer points than youth from highly urban areas between baseline and endline (p=.016). 
Community Context
MYDev youth scored significantly higher than comparison group youth at endline with scores of 18.0 
points compared to 16.4 points on the community context index (p<.001). When controlling for other 
characteristics both sex and rural area were significant. Female youth were more likely to score about 1.0 
points lower than comparable male youth at endline (p=.006). Youth from rural areas were more likely to 
score about 2.4 points lower than youth from highly urban areas (p<.001). 
MYDev and comparison group youth had statistically similar score gains on the community context index. 
When controlling for other characteristics, only rural area was found to be significant with youth from 
rural areas improving, on average, about 3.0 fewer points than youth from highly urban areas (p<.001). 
E. Connections Between Employability, Skill Gains, and Perceptions
The complexity of MYDev and OSYDA programming in its multiple and interrelated intended outcomes 
necessitates a robust examination of the connections of the different measurements central to MYDev’s 
aims. These include 1) new or better employment; 2) gains in either life, work readiness, or leadership 
skills; 3) improved perceptions of either government or community; and 4) changes in assets or contexts. 
This section aims to explore the relationships between these components to provide additional and 
contextualized insight on key findings. 
1. Relationships with Employment Outcomes 
Intervention group youth who were employed at endline (regardless of their employment status at 
baseline) improved their government perceptions and work readiness skills more than youth who did 
not have employment at endline. Specifically, youth with employment improved their government 
perceptions by almost half a level (.38) more than youth without employment (statistically different at 
p<.001). Employed youth also improved their work readiness skills more than non-employed youth,
and these differences are significant at least p=.015.
When newly employed youth were compared to youth who had maintained employment since baseline, 
these newly employed were found to have gained statistically more leadership skills (p=.037), though the 
other relationships between employment and government perceptions and work readiness skills fade.
2. Relationships with Improved Perceptions
When examining whether key variables have a correlation with the improved perceptions of community 
or government, change in youth’s assets and contexts were found to be significant. In fact, youth with an 
increase in their asset scores at endline were 18.8% more likely to have improved perceptions of their 
community or government (p<.001). This correlation remains significant even when accounting for other 
demographic characteristics, such as treatment status, sex, and geographic area. Similarly, youth with an 
increase in their context scores were about 18.0% more likely to have improved perceptions of their 
community or government (p<.001). This correlation also remains significant when controlling for other 
characteristics. 
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 24
V. CONCLUSION
This quasi-experimental impact evaluation of MYDev’s work in FY15 has revealed that many MYDev youth 
seem to be improving their economic status and level of social engagement:
 Almost 24% of the intervention group were newly employed at endline. On average, these 
employed youth tend to be male, urban workers who hold one temporary job in which they are 
remunerated in money. These workers tend to struggle to save, though they eat at least three 
meals a day, report that their savings or earnings have increased and that they earn enough to 
meet their needs.
 42.2% of MYDev youth were working at baseline and improved their work situations between 
baseline and endline.
 About half of MYDev youth were unemployed at endline, and over three quarters of these youth 
had been unemployed at baseline as well. The average unemployed youth is rural, looking for 
work, and dissatisfied with his or her work situation.
 Almost 70% of MYDev youth improved in either life, leadership, or work-readiness skills over 
the course of the intervention. 
 51.5% of intervention group youth improved their perceptions of community and 52.5% 
improved their perceptions of the government.
These findings offer the following progress towards MYDev’s key indicators:
Indicator Findings IE Extrapolated to Population # 
OSY Supported
Employment (new or better 
employment)
32.2%
(69.4% male; 30.6% 
female)
1,601
(1,111 males; 490 females)
Perceptions 66.7%
(61.6% male and 38.4% 
female)
3,322
(2,047 males; 1,275 females)
Skills 69.8%
(60% male and 40% 
female)
3473
(2,084 males; 1,389 females)
While findings on MYDev youth alone are quite positive, relative to their non-MYDev peers, MYDev youth 
made gains in employment, but not in other areas:
 MYDev youth obtained new employment at a higher rate than did comparison group youth, with 
almost 24% of intervention group youth previously unemployed youth employed at endline 
compared to about 14% of youth in the comparison group (significant at p=.000).
 MYDev youth and non-MYDev youth have statistically similar quality of employment.
 MYDev youth’s perceptions of community and government, respectively, improved less than did 
non-MYDev youth’s.
 MYDev youth’s skills improved less than did non-MYDev youth’s.
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 25
Beyond these findings that satisfy our research questions, this impact evaluation has afforded us the 
opportunity to learn more about the mechanisms behind improved economic and social outcomes, and 
how those two elements interact with each other and with youth’s demographics:
First, gender matters. Gender and skill development seem to be related, with female youth improving 
significantly less than their male counterparts in life skills (p=.006) and work-readiness (p=.009) 
components, though the mechanisms at play behind this relationship are unclear. Further, females have 
lower rates of employment than males and tend to enter different sectors when they do have 
employment. These females also tend to feel safer at work and save more, all while working almost ten 
hours more than male youth on average. Beyond economic outcomes, female youth have worse 
perceptions of their communities and government.
Next, one’s geographic area matters, but in a way that is complex and likely related to more than just their 
personal skills and opportunities, but also to contexts, including local markets and communities. For 
example, while rural youth had lower rates of employment than highly urban youth, they were more likely 
to meet needs with earnings. And these rural youth were more often employed than youth from partially 
urban areas. Rural youth had significantly worse perceptions of the government (p=.004) and community 
(p=.001) than youth from other geographic areas and also reported having less internal and external 
assets than urban youth.
Finally, there is a relationship between being employed and exhibiting more work readiness skills. In 
addition, youth who are employed also had more positive perceptions of the government.
VI. RECOMMENDATIONS
This impact evaluation has afforded the opportunity to learn more about the mechanisms behind 
improved economic and social outcomes, and how those two elements interact with each other and with 
youth’s demographics:
Improved livelihoods. A central intended outcome of MYDev and OSYDA programming is improved 
employment outcomes for out-of-school youth. All analyses point to success for MYDev regarding 
improving the employment situations of MYDev youth. Across the course of the intervention, 32.2% of 
intervention group reported new or better employment.
Recommendation: Future programing should strengthen and broaden partnership with private 
sector, identifying market driven opportunities with a greater emphasis emerging skills 
opportunities.
Recommendation: Future programing should support activities to encourage entrepreneurship and 
self-employment by connecting OYS to local government and private sector programs, identifying 
support mechanisms for strengthening the new entrepreneurship skills for long term sustainability 
and profitability.
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 26
Improved Perceptions of Government and Community. MYDev programing aims to improve youth’s 
engagement within their communities, while also improving youth’s perceptions of their communities 
and governments. The majority of MYDev youth had improved their government or community 
perceptions by the time of endline administration. In fact, 66.7% of MYDev youth had improved 
government or community perceptions.
Experiential Learning to Improve Work Readiness. The Work Readiness component encompasses 
questions measuring skills like solving conflicts at work, looking for a job, and working towards a pay 
upgrade. MYDev youth that were employed at endline had higher work-readiness skills than unemployed.
The Role of Gender. Gender and skill development seem to be related, with female youth improving 
significantly less than their male counterparts in life skills and work-readiness components, though the 
mechanisms at play behind this relationship are unclear. Further, females have lower rates of employment 
than males and tend to enter different sectors when they do have employment. These females also tend 
to feel safer at work and save more, all while working almost ten hours more than male youth on average. 
The Role of Geography. One’s geographic area matters in a way that is complex and likely related to 
more than just their personal skills and opportunities, but also to contexts, including local markets and 
communities. Rural youth had significantly worse perceptions of the government and community than
youth from other geographic areas and also reported having less internal and external assets than urban 
youth.
Recommendation: Future programming should investigate unique challenges and needs of female out￾of-school youth and determine if supplemental program may be helpful to lessen the gender gaps in 
employment outcomes and skill gains.
Recommendation: Future programming should aim to target the needs of rural youth through tailoring 
of MYDev programming to the rural labor market context and engagement with government services.
Recommendation: Further investigate approaches to developing work-readiness skills to determine 
the ideal combination of technical training and experiential learning. 
Recommendation: Future programming should promote the recognition by government agencies 
and the private sector of work readiness skills as a facilitating factor for the success of young people.
Recommendation: Future programming should continue to promote deepening of OSY-community 
connections through increased partnerships with government institution, community leaders and 
other community groups in civic engagement, peace building and youth development activities. 
USAID Mindanao Youth for Development Program FY 15 Impact Evaluation Report 27
VII. ANNEX
USAID’s Mindanao Youth for Development 
(MYDev) Project
Improving Peace and Stability in Conflict Affected Areas
FY15 Impact Evaluation Results Presentation
November 2016