Reforming labor policies and community programs for lower-income workers

OVERVIEW:

Mixed-methods research to inform labor policy reforms, community initiatives, and communications strategies that benefitted 234,000 lower-income workers.

METHODS:

Interviews, Co-design, Focus Groups, Surveys, Data Analytics (Latent Class Analysis, Regressions, t-tests)

DURATION:

18 MONTHS

2020.09 - 2022.03

ROLE:

Sole researcher overseeing 6 surveys, 23 focus group discussions and 30 interviews; design facilitator

DELIVERABLES:

Quantitative and qualitative research reports, research insights

Jump to: Context • Methods and Timeline Overview • Policy • Community • Communications • Reflection

🌰 In a nutshell…

Pillar 1: Policy

Inform new policy terms through a study of stakeholders’ lived experiences

💡

• Used surveys, statistical modelling, interviews and focus groups to understand and size the pain points workers, employers and consumers faced with the current policies • Findings directly translated into 8 of the 18 policy amendments in the reform

Pillar 2: Co-design

Pilot community-led initiatives

💡

• Identified 4 challenge areas for the community co-design program • Facilitated 1 of 8 project groups in a 6-month co-design program, comprising 54 people, public and private sector participants

Pillar 3: Communications

Formulate a data-driven communications strategy

💡

• Used surveys and focus groups to pinpoint what the public found confusing about the policies and inform the design of creative assets • Collaborated with communications colleagues on a data-driven communications strategy that significantly improved public understanding and public perception of the policies

Context

Lower-income workers were amongst those most impacted by COVID-19 in Singapore.

The Ministry of Manpower (Singapore) came together with employers and union leaders to explore policy and community solutions to the challenges they faced in the wake of the pandemic.

Lower-income workers are defined as workers with wages in the bottom twentieth percentile. In Singapore, these workers tended to be from cleaning, security, landscaping, retail and F&B industries.

I was the sole user researcher responsible for end-to-end research design, vendor management, synthesis and reporting.

This was the first policy review in the Ministry that heavily made use of user research data in conjunction with administrative data, such as labor market statistics. To advocate for the value of user research in the policymaking cycle, I collaborated closely with counterparts in policy and communications departments, and directly reported findings to cross-functional leaders and political leadership.

I worked with research vendors and colleagues in other departments to deliver findings and recommendations directly to senior leadership.

The impact that user research had on this project led to a researcher being tagged to each major policy review that the Ministry took on after.

Prior to this project, our research team had only contributed to ad-hoc research requests and managing a monthly perception tracking survey. After I demonstrated the value that user research could bring to the policy formulation, our team gained the trust of our inter-departmental colleagues and leadership, and have been able to direct research efforts more proactively in subsequent policy reviews.

I was eventually able to expand my team when I became manager, doubling in size from 2 to 4 researchers, to support the expanded responsibilities.

I was invited to present this project at a Community of Practice gathering of government-wide user and public opinion researchers

As it was a lighthearted peer-led brownbag, we chose a pixel game theme and framed our research journey as a quest to rally community support

⬆️ Back to top

Methods & Timeline Overview

Historically, my internal stakeholders have favored quantitative data such as employment rate and income growth in data-driven policymaking. They focused strongly on the generalizability of research findings, as any resulting policy changes would impact the population at large.

💡

Hence, I chose to run large sample surveys where possible, and triangulated the data with findings from interviews, co-design and focus groups. This way, I could provide them with the confidence that various problem spaces have been appropriately sized, while socializing them to the value of qualitative methods in humanizing these pain points.

Across 18 months, I oversaw 6 surveys, 23 focus group discussions and 30 interviews across 3 pillars of the policy reform.

⬆️ Back to top

Pillar 1: Policy

Methods

Inform new policy terms through a study of stakeholders’ lived experiences

Sep 2020 - Mar 2022

Qualitative
30 contextual interviews with lower-income workers to understand their lived experience with current policy terms
18 focus groups with workers, employers, and consumers to identify opportunities for policy and community interventions
Quantitative
2 online surveys (n=1,000 each) to compare the general public’s understanding and perception of the old and new policy terms
1 telephone survey (n=1,000) to understand workers’ reactions to reform
Latent class and regression analysis in R using the monthly perception survey (n=1,500 monthly) that our team also manages
Participants were recruited by vendors, using screening criteria I provided.

💡

I chose to start with qualitative methods to better understand workers’ lived experiences, and how workers, employers and consumers viewed each other in this landscape. Using the emerging themes from the interviews and focus groups, I designed the surveys and analytics studies to get a representative sense of how the population at large interacted with these policies. I chose to run the workers’ survey over the phone, as they were not well represented on online survey panels, and preferred verbal conversations.

Discoveries

My research uncovered frustrations that workers and employers had with the current policies that would have been absent from an analysis of labor market statistics alone.

Finding #1

Lower-income workers, who tended to perform frontline and essential service duties, have been overworked and under-appreciated since the pandemic started. The public were harsher towards them, as they were unhappy about the workers upholding stricter pandemic protocols and mandates.

Finding #2

While 70% of the public expected the government and employers to be responsible for workers’ wages, employers shared that it was difficult to pay their workers more than existing baseline wages as they needed to be competitive with their prices.

Hence, beyond adjusting the policy parameters to improve wage growth, we wanted to support the workers holistically by also tackling public appreciation and consumer attitudes.

A focus group discussion with workers.

💡

I invited policy colleagues to observe the focus groups, as they rarely had an opportunity to directly hear from the citizenry.

Bonus: Are there consumer archetypes?

💡

To better understand different consumer archetypes and their attitudes towards supporting lower-income workers, I mentored the intern on our team to run latent class and regression analyses on attitudinal and perception data from our team’s recurring monthly survey. We ran the latent class analysis purely on psychographic data to uncover attitudinal archetypes without being biased by our assumptions about demographics. Demographics were mapped onto the output of the latent class analysis after to flesh out the archetypes.

We uncovered 4 archetypes that were used by policy colleagues to scope an accreditation program for progressive employers (external link) after my departure from the Ministry.

Reports by my amazing intern, Tzu Zheng.

Data redacted due to sensitivity.

Outcomes

The policy reform uplifted the employment outcomes of 234,000 workers.

💡

I regularly updated a tripartite workgroup of political, union, and business leaders on my research recommendations. These added a human perspective to the labor market analyses and policy impact projections tabled by colleagues from policy teams.

The workgroup subsequently presented 18 policy recommendations, of which 8 were derived from insights from my research, that impact 234,000 workers in total - these included :

Establishing an accreditation system to encourage consumers to shop from progressive businesses Establishing standards for protecting workers’ well-being Providing transitional support for employers to ensure that their businesses stay viable

The recommendations have been gradually implemented since 2021.

The full report tabled by the Tripartite Workgroup on Lower-Wage Workers.

⬆️ Back to top

Pillar 2: Community

Methods

Pilot community-led initiatives

Dec 2020 - Sep 2021

Details

Co-design
Design facilitation for a 6-month program to co-design community initiatives
Participants were recruited by colleagues on the citizen engagement team.

Discoveries

We brought together 54 community leaders, employers, and union representatives to co-design community initiatives that would complement top-down policy changes.

Over 6-months, they developed 8 pilot initiatives that addressed 4 challenge areas derived from the focus groups:

Employment opportunities

Employer-employee relationships

Public appreciation

Workers’ well-being

The team I facilitated launched Makan&Shine (makan means “to eat” in Malay) - an initiative that brings together lower-income workers for microlearning brown bags over casual meals to help them access new job opportunities and connections.

🤔

There were operational challenges in co-designing directly with the workers due union regulations. Hence, we worked with their union representatives as a proxy. Nonetheless, all 8 project teams worked with lower-income workers in the development of their respective projects.

The first of 6 co-design workshops. Participants ideated on the 4 challenge areas during this session.

At the second workshop, we formed 8 groups after a round of dot voting on ideas across the 4 challenge areas.

❤️

These 6-months cemented my decision to go to grad school for design. Everything felt right when I was facilitating these workshops on weekends, outside of my work week. It reignited my love for facilitation, it made me realize how much I valued exploration and creativity, and I loved seeing the teams drive their own design projects after providing them with the tools to do so. This approach to empower ground-up action was a breath of fresh air from the top-down approach in past projects I was assigned to.

Guiding participants through a Theory of Change map at the final workshop.

Outcomes

This project transformed the way we work with citizens.

The first run of this co-design program was so successful that it was renewed for a second run in 2022 - the only governmental co-design program to be extended. Several project teams from 2021 are still carrying on with their projects!

It was also awarded the Public Sector Transformation Award for its innovative approach to citizen engagement.

The cohort from the first ever Alliance for Action for Lower-Wage Workers.

⬆️ Back to top

Pillar 3: Communications

Methods

Inform communications planning for the reform and quantify impact on policy perception

Dec 2020 - Nov 2021

Quantitative
3 waves of surveys (n=1,000 each; pre-, during, and post-campaign) to evaluate the impact of the public communications campaign
Qualitative
5 focus groups with workers and consumers to concept test the framing of the policy reform announcement
Participants were recruited by vendors, using screening criteria I provided.

💡

I chose large sample surveys (oversampling lower-income workers ) so that I could track awareness and attitudinal metrics over the course of the campaign and compare attitudinal shifts between people who saw campaign assets and those who did not. This helped colleagues to quantify the impact of effective public communications beyond traditional publicity metrics such as impressions and reach.

Discoveries

Low awareness of the wage protection policies impacted the public’s perception of government efforts to support lower-income workers.

At a general population level, awareness of the wage protection policy was particularly low, as tracked in the recurring monthly survey our team managed.

Data redacted due to sensitivity.

The pre-campaign survey helped to us identify specific parts of the policies that were causing confusion and misconceptions, and the insights informed the design of publicity assets.

Our research showed that lower-income workers wished to see how peers have benefited from the policies, hence we featured real-life workers in the main campaign visuals.

The general public, on the other hand, wanted to see tangible numbers of how much support was given to workers. We created infographics that spelled out these figures to accompany the main campaign assets.

Outcomes

Awareness and perception of these policies improved significantly post-campaign.

There was a statistically significant improvement in awareness and perception of the policies among people (both general public and lower-income workers) who have seen the campaign assets.

Data redacted due to sensitivity.

This research also helped to showcase the value of a data-driven communications strategy to colleagues and management.

Our team had previously assisted the communications teams on ad-hoc concept tests and small A/B tests of assets. This project forged greater trust between our teams and built communications colleagues’ confidence in using more data in their day-to-day work. It also increased budgetary support on future communications projects to incorporate more research into the strategy planning and performance tracking.

The surveys showed that the emotive TVC was the most effective asset in this campaign.

⬆️ Back to top

Reflections

Process

Code-switching for impactful research

I quickly became adept at “code switching” and data storytelling as I was working in a cross-functional workgroup with colleagues from policy and communications departments. I had to translate their deeply technical business questions into research questions and then back into actionable insights. I also had to present the insights differently to different leaders to best align with their respective thinking styles.

Missed opportunity for contextual inquiry

That said, as a junior researcher running the study during the pandemic, I was not able to introduce more methods into the studies. A contextual inquiry would have been most suitable to understand the day-to-day experiences of the workers we were trying to support.

Personal

Where passion, purpose and persistence intersected

This remains my largest and most challenging project to date. There were many moments where juggling the multiple workstreams on this project, managing junior researchers, and fielding ad-hoc requests got overwhelming, but I pushed through with a strong belief in the purpose of the work. Being able to use a bunch of different methods in this project brought me a lot of pure, nerd joy too!

⬆️ Back to top