United States Department of Labor
I tell anyone who will listen that BLS staff love to talk about our data. We have LIVE people at the end of the phone line (or email request) who are happy to answer questions about BLS data and the methods behind those data. The COVID-19 pandemic has not stopped our ability to respond to public questions. Even in our telework posture, we pride ourselves on outstanding customer service. All BLS statistical programs have staff who answer public information requests. We also have a central information staff out of our national office and eight information offices scattered around the country. Yes, we get questions, and we are more than happy to provide answers.
Recently, we’ve received some general questions about our methods, which cover multiple BLS programs. Here are a few of those questions and our answers.
Why does BLS revise published estimates?
One of the hallmarks of BLS economic indicators is their prompt release. We provide a “first look” at a variety of economic conditions, including employment and unemployment, price change, wages, productivity, and more. To release these data in a timely manner, we follow very strict data collection and processing schedules. Data obtained after the collection deadline are not included in the initial release but can be incorporated later. We identify data subject to these revisions as preliminary. Revisions are a necessary part of the statistical estimation process to ensure accuracy.
The Producer Price Index (PPI) recently expanded the amount of revised data available to the public. PPI data are revised for 4 months following initial release, again to account for information received following the initial deadline, thus providing a clearer picture of price change. Until recently, revised data were only available in the fourth month. For example, July data originally published in August would be revised with the November release in December. The expanded data now available show monthly revisions for each of the 4 months following initial release. So, following the initial release of July data in August, revised data for July are available in September, October, and November, before we release final data in December. This change is in response to requests from data users for these interim values.
Other BLS programs release periodic revisions as updated data become available, providing a clearer view of the economy. For example, the Current Employment Statistics program has more information about the monthly revisions to payroll employment data. Details about the methods behind all BLS programs are available in the BLS Handbook of Methods.
Why is it important to respond to BLS surveys?
We carefully design our survey samples to represent the people and businesses in the United States. Without input from these sample members, BLS indicators would not accurately reflect the economic and social conditions in our country. We strive to make completing our surveys as easy as possible, and we often offer multiple ways to provide information. We design survey questions that are easy to understand and answer in a short period of time.
Nearly all of our surveys are voluntary, which means the people, households, and organizations selected can choose whether to participate. We are grateful that the great majority of them agree to participate. The information benefits all of us.
BLS maintains response rate information on our website and updates this information on a regular basis. This information can be very technical, which is why BLS staff stand ready to answer any questions you might have about response rates.
Check out this video to learn more about the importance of responding to BLS surveys.
What effect did the pandemic have on BLS survey participation?
With some careful planning, a lot of hard work, and a little bit of luck, BLS has been able to release all planned data products on schedule, despite the pandemic. We weathered both internal and external challenges. While many of our tasks had been successfully tested in remote environments, we had to change a few processes. Fortunately, those changed processes were successful, and some even spawned innovations we will continue. Externally, we were mindful that many businesses had limited operations or were closed, and many households were preoccupied with illness, childcare, and other responsibilities. Response rates did decline. Since the start of the pandemic, each BLS program has provided more information about survey response and methods. In some cases, response rates have recovered from their pandemic lows, but many are still below levels before the pandemic.
What steps has BLS introduced to combat weak survey response during the pandemic?
BLS takes many different approaches to data collection and works closely with our partners in the states and other statistical agencies to obtain high quality information from businesses and households. Traditionally, some data collection is done in person, where BLS builds a relationship with survey respondents and shows them the importance of response. BLS also offers many options designed to make ongoing response easy, including use of the internet, email, file transfer, and others. At the start of the pandemic, BLS suspended all in-person data collection. We were fortunate that many businesses, even many of those with limited operations during the pandemic, maintained electronic records they provided to BLS, allowing us to continue producing key economic data.
For our part, the pandemic provided an opportunity to accelerate our ongoing move away from paper and mail. We used phone and email to contact respondents and obtain their data. We also began to experiment with video data collection, a process that proved very successful and is now a vital part of our data collection toolkit. While we started slowly with video collection, and took particular care to ensure confidentiality, we quickly discovered huge benefits. BLS staff can use video communications systems to share their screen, demonstrate BLS confidentiality procedures, show data products, and more. In person, shuffling all these papers can be a little unwieldy. With a little practice and planning, video data collection has proved invaluable.
BLS also has explored ways of capturing information without burdening respondents at all. In some cases, we are able to use web scraping to obtain needed data. We are also exploring supplemental data sources, such as data aggregators and crowd sourcing websites. We have accelerated these explorations during the pandemic. We are learning a lot and obtaining more and more data through these alternative approaches, which can mitigate the effects of declining response rates on data quality. These efforts will ensure that BLS data products remain of high quality with enough detail for stakeholders, while lessening respondent burden.
We will return to some in-person data collection over time and will use those interactions to build ongoing relationships. But we also will continue to advance these innovations, such as video collection and web scraping, as options to make data collection more efficient in the future.