How workers in San Francisco’s Chinatown harnessed data for community change

Published 1/20/2015

Last month Yank Sing, one of the biggest and most successful dim sum restaurants in San Francisco, agreed to pay $4 million in back pay to 280 employees for wage and overtime violations. The story made national headlines and was the largest minimum wage settlement that California’s labor commissioner had ever been involved in.

How a group of low-income immigrant workers who spoke limited English brought exploitive working conditions to light and got the world to listen is a complicated story. It was years and lots of hard work in the making. But this story is unusual for another reason, as it turns out. Data collection and analysis play a big part.

An Immigrant Neighborhood Adapts to a Changing Economy

In 2008, for Chinese immigrant workers who wash dishes, push dim sum carts, and cook in the many restaurants that make up San Francisco's Chinatown, conditions had gone from bad to worse. After the decline of manufacturing and the closing of the city's many garment factories in the early 2000s, many of the area's workers were forced to compete for jobs in the neighborhood's restaurants — jobs where they often had to work long hours, below the minimum wage, often without breaks, and at the mercy of frequent verbal abuse from the bosses in the back of the house.

The Chinese Progressive Association (CPA) had been organizing around these issues in the community for some time, fighting for enforcement of the city’s minimum wage law, and had won successful wage theft cases. But workers who joined CPA were looking for ways to make more systemic change. Many community members were skeptical that anything big could be done to improve working conditions. “This is just the way things are in Chinatown,” many said, according to CPA organizer Shaw San Liu.

Community members at large, Liu said, were often fatalistic, and the challenges they'd faced in getting the legal and regulatory systems to enforce the law had solidified these attitudes. Policymakers, many neighborhood residents felt, just didn't care about low-income, non-English-speaking, immigrant laborers.

Leaders at the CPA wanted to do more to help workers get their story out to the public. But how?

One of the CPA's founders, Pamela Tau Lee, had formerly worked with the University of California at Berkeley's Office of Labor and Occupational Health. Maybe researchers there, she hoped, would be able to help Chinatown's restaurant workers document the unfair labor practices and abusive conditions.

An Unlikely Partnership

Meredith Minkler is a professor at UC Berkeley's School of Public Health with a special interest in community-based participatory research. When she heard the Chinatown group's request, Minkler said she had only one concern.

They called us because they wanted hard data that they believed would support their point,” she said, that poor working conditions and wage theft were widespread in Chinatown restaurants. “We said yes, we were happy to help.”  Although Minkler was fairly confident that the data would support the workers concerns, she wanted to make sure the group understood that the data collected could go either way. She told the CPA organizers: “You may not get back what you want to hear.”

For workers and organizers, that was a risk they were willing to take. “As we were doing this, the thing that was clear to us,” Liu said, “was that we were never going to solve this one restaurant at a time, considering how endemic the issues were.”

"They called us because they wanted hard data."

For policymakers and the public to take note of working conditions in Chinatown, the organizers knew they needed more than personal experiences and anecdotes. “We wanted to be able to use our resources to move forward a longer institutional plan to combat conditions and organize workers across the industry,” Liu said. “And we knew we needed data to do that.”

Minkler had agreed to partner with the Chinatown group on the study, but she knew that the traditional way of doing a research study wasn't going to work. Minkler had tired of what she and others call “parachute research,” where outside researchers drop into a community they are unfamiliar with, collect data, and then disappear, taking the data and any hope of using it to support near-term community change, with them.

“I think it is very important to work with, rather than on, communities,” Minkler says on the UC Berkeley website. She notes that collecting data as an outsider can result in lower participation rates and biases. In this Chinatown project, she knew the researchers would need the help of community residents to collect this information.

“Communities often have sophisticated insider knowledge and understanding,” Minkler writes in an essay in “What Counts: Harnessing Data for America's Communities.” This knowledge, Minkler says, can be crucial in ensuring that researchers “ask the right questions and gather data in ways that will increase the ‘relevance, rigor and reach' of the findings to effect change.”

This was certainly the case in Chinatown. Minkler and the organizers at CPA set about building a broad coalition of researchers, public health officials, workers, and community partners to collect the information. In addition to the University of California and the Chinese Progressive Association, their coalition eventually included the San Francisco Department of Public Health (SFDPH); San Francisco Medical School, University of California; the UC Berkeley School of Public Health and its Labor and Occupational Health program. The study included both a worker-administered community survey and an SFDPH-administered observational checklist.

"IN the end the team conducted interviews with 433 restaurant workers. The SFDH team observed 106 of the 108 restaurants in Chinatown."

Minkler and her team, including then doctoral students Charlotte Chang and Alicia Salvatore, worked closely with a corps of restaurant workers recruited through the CPA. The full team trained the workers on how to use and administer the survey to collect the data that was needed. At the same time, workers gave the researchers feedback on what questions to include on the survey and how to ensure the survey was sensitive to cultural and other issues.

According to both Liu and Minkler, there were a lot of challenges. For starters, the survey instrument, a standard occupational health and safety tool, was long and intimidating. With 118 questions, many of which had a long list of subquestions, the instrument took more than an hour to fill out. Workers were concerned it would be difficult to find Chinatown restaurant workers who had the time, and the courage, to be interviewed. Liu said there is a lot of fear among low-wage restaurant workers, not unfounded, that if they speak out they will be fired.

Organizers worked closely with volunteers to overcome these fears as well as the cultural norm that problems at the workplace are not acceptable to talk about. “There are not a lot of shortcuts when you are trying to collect data from folks who are pretty economically and socially vulnerable and aren't used to sharing their data and information and experiences,” Liu said.

Workers also had to learn to understand the value of including things like standardized scales which would allow responses to be compared across populations. Understanding human subjects review, privacy, and why questions need to be asked in a neutral way were all new ideas for the community interviewers.

“We'd role-play,” Minkler said, “about how you tell someone about the study and ask if they want to participate and explain about confidentiality. When a worker-interviewer would say to prospective interviewees, ‘We're doing this survey because we want to show how bad things are in Chinatown,' we'd say, ‘No, you can't say that.' We'd role-play until they understood how to ask a question that wasn't biasing the reponse.”

But both Minkler and Liu said working through these challenges was worth it. First and foremost, the connection to CPA and the workers made it possible for researchers to enter the community in a trusted way, with connections to workers, language capabilities, and working knowledge of the community that they would not have had otherwise. The partnership also made it possible for researchers to ask the right questions.

There were several instances, Minkler said, where workers pointed out flaws in the survey instruments in ways that were invaluable. For example, in a standardized scale for measuring depression and anxiety, one of the questions included the idiom "butterflies in your stomach." The workers were puzzled. There was no such idiom in Chinese. They rewrote the question.

The workers also helped the researchers ask the proper questions about how wage theft happens, which many of the researchers didn't fully understand coming into the study. For example, many of the researchers had no idea that employers can (and do) steal tips in several different ways, including through credit cards.

“Many of these details we didn't know — we wouldn't have asked them if we hadn't had the workers,” said Niklas Krause, now a professor of Epidemiology and of Environmental Health Sciences at UCLA who participated in the study when he was Professor of Medicine in the Division of Occupational Medicine at UCSF.

“Armed with the data, we were able to actually get policymakers to pay so much more attention to the issues and deepen their commitment to addressing it,”

After perfecting the survey, the workers with an additional group of 17 who were trained as surveyors, hit the streets in the summer of 2008. Traditional researchers would start a survey of restaurant employees inside the restaurants. But in Chinatown, that would have left a sample size of zero.

“We learned from the community workers that people are too afraid in this close-knit community to be blacklisted if they speak out against their boss,” Krause said. Workers were afraid that if they did anything the employer perceived as disloyal, they'd never get another job in Chinatown. And because most workers only speak Cantonese or Mandarin, they lacked job options.

“That was quite stunning to me,” Krause said. “We had to take more care about how to approach the workers so they wouldn't be afraid of losing their standing in the community. So we couldn't survey them near their restaurant. You wouldn't know that from the outside.” To ensure privacy, the research team interviewed workers at pastry shops, parks, or in the single room occupany hotels where many lived.

In the end, the team conducted surveys with 433 restaurant workers — an astonishing number given the sensitive issues involved — and SFDPH observed 106 of the 108 restaurants in Chinatown.

Workers also helped with the analysis, Krause said. The research team had been training the workers in data analysis by reading news or other reports and asking them to analyze basic results. It didn't take a lot of work, said Krause. “These folks are very smart. Waitresses have to hold a lot of data in their heads. They're very good at it already.” Later, when Krause, Salvatore, and other academic researchers came back with initial results, they would ask the worker advisory group, “Does this sound true? Can you explain it? What does it say?”

The Power of Data

Krause said worker insights were again invaluable. The rate of workers receiving paid sick leave, for example, seemed higher than most labor studies typically show. When asked about it, the workers said, yes, they're allowed to take a sick day. But when pressed, they also said that they work an additional day without pay for each day off. To them, that was sick pay.

In the end, the results showed just how widespread labor violations and poor working conditions were. Among the major findings:

  • Wage theft in Chinatown restaurants was rampant. One in two workers report being paid less than minimum wage.  Only 5% are paid a living wage.
  • Of those whose wages are withheld for various reasons, one-third never receive the back pay. 
  • 30% of those relying on tips say their boss takes a portion of the tips. 
  • Workers report long work days: 42% worked more than 40 hours a week, with half of those workers working over 60 hours a week. Three-fourths are not compensated for overtime.
  • Breaks are short, though smokers tend to get longer breaks.
  • Workers experience injuries, work in hazardous workplaces, and most do not receive training.
  • Workers report high levels of stress and do not have necessary healthcare and time off to address their medical conditions or injuries.

These basic data — simple frequencies or rates — which are often not sophisticated enough to publish in academic journals, proved to be very powerful.

Armed with the data, we were able to get policymakers to pay so much more attention to the issues and deepen their commitment to addressing it,” CPA's Liu said.

In 2011, the San Francisco Board of Supervisors passed two wage theft ordinances that tightened regulations and to created a wage theft task force. Then, in 2013 Dick Lee Pastry Inc. settled with the city for $525,000 in back wages and penalties. The city sued the restaurant for forcing employees to work 11-hour days, six days a week for less than $4 an hour. The survey data also helped organizers win the $4 million settlement with Yank Sing. The restaurant also agreed to other changes such as base wage increases, holiday and vacation pay, fully paid health care for full-time employees, a workers compliance committee, and workers' rights education.

As a result of the project the SFDPH began to consider worker health and safety more regularly in their observations across the city, and former Director of Environmental Health Ragiv Bhatia reached out to other national regulatory bodies about doing so as well.

"A community doesn't need a statistician to get powerful data that has an impact."

More researchers should consider the power that basic data had in Chinatown, Krause thinks. Sometimes, he says, the bar for what is being considered “scientific" is set overly high. Often people assume that results are less valid unless they're “statistically significant.” The results may show that something is three times more dangerous, but because the sample is small and the result is not statistically significant, the increased risk may get ignored.  “But it's still three times more dangerous,” says Krause. (And likewise, very small differences in risk may appear as statistically significant as long as the size of the study population is very large.)

“I'm not saying this to dumb down community research, but rather to assert that even with small samples and even without applying statistical tests you can produce useful information,” he said. “A community doesn't need a statistician to get powerful data that has an impact.” However, he says, “researchers can help communities to know when and how to apply research methods in order to increase the validity of their data.”

Beyond the data, Liu says the major victory of the survey is how it helped build the capacity of workers who participated to become leaders, to talk about issues that matter, and believe in their own agency to fight for and make change.

Workers who were involved in the study formed a leadership body at the CPA, and continue to meet monthly to discuss workers rights and conduct outreach and education activities.

“No one is going to do it for us,” Liu said. “If communities want to make change, we have to know that it's up to us to tell our stories. In a data-driven world and in a context where, like it or not, the lived experience of poor and working-class immigrant communities is not going to garner the same attention and care as facts that are vetted by the academy, we have to play ball if we are going to have the voice of our communities heard.”

Photos: San Francisco History Center, San Francisco Public Library; Meredith Minkler; Kathleen Costanza

Other examples in the What Counts volume of how data can empower community residents:

Join the conversation on Twitter using #WhatCountsForUSA or share this on Facebook.