When it comes to solid data on jobs in K-12 education, government policymakers are flying blind. There is no valid dataset that can tell us definitively what is happening to K-12 staffing levels across the country. In that absence, media reports of considerably reduced staffing in some locales have created a false narrative of drastic overall cuts since the start of the Great Recession in 2008.

To address this lack of data, we built a better -- though still imperfect -- dataset by combining information from multiple sources and correcting for time lags and outright errors to track the numbers of local school-system employees, including not only teachers and instructional aides but also school and district administrators and support staff. Our analysis finds that 2008-09 was actually a peak for K-12 staffing after several decades of growth. Student-staff ratios after the Great Recession are now equivalent to those that were the norm a decade ago -- a far cry from historic lows.

Setting 2008-09 as the benchmark staffing level creates an artificially high watermark compared to historical patterns. That makes it hard for journalists to see the real deal when it comes to education-sector jobs. And it's not just the media who are affected by the lack of data. Policymakers are making huge financial decisions in the dark. When the federal government allocated $100 billion in education stimulus funds in 2009, and then another $10 billion for the Education Jobs Fund in 2010, lawmakers were doing so to stem the recession's effects on public schooling. Both allocations, however, were made in the absence of solid data on historical staffing trends.

That information gap persists: We don't know if the Jobs Fund worked, or if it was needed in the first place. We don't know whether new state investments being proposed -- such as hiring more teachers to reduce class size -- make sense, because we don't know the actual size of staff-student ratios in schools. When policymakers want to know how public-pension or retiree health-care costs are affecting education budgets, they'll continue to operate in the dark without solid, up-to-date numbers on key staffing indicators.

Here's what our dataset shows:

Staff-student ratios nationally are the same as they were a decade ago. The number of public-school K-12 staff per 1,000 students peaked at 129 in 2009-10 and has now dipped to 123, which is the same ratio as in 2003-04.

Graph: K-12 staff-student ratios

Even though staffing ratios have dipped, they climbed so much before the recession that staffing ratios today are still higher than those of previous decades. In 1995-96, for instance, there were only 111 staff members for every 1,000 students. Despite the recessionary dip in total staffing, the system still employs 12 more adults today for every 1,000 students than the average in 1996.

Staffing ratios vary tremendously by state. For every thousand students, some states have up to three times as many staff as other states. Today, for example, Vermont employs 211 K-12 workers per 1,000 students (about one adult for every five students), while Nevada employs 72 (about one adult for every 14 students). In general, northeastern and midwestern states have higher staffing ratios than southern and western states. Since 2000-01, staff-student ratios have decreased in southwestern states and increased in most other states.

Big changes are underway in our country's public-education system, and better data will be critical in monitoring the effect of these changes. The trends can tell us how state finance policies affect staffing ratios, highlight changes in delivery across states, and clarify the changing role of technology on staffing policies. And as the evidence here shows, any averages used at the national level are unlikely to capture the dramatic variations in staffing ratios from state to state. In fact, as long as federal efforts such as the Education Jobs Fund rely on national estimates of staffing, these blunt levers may be effectively subsidizing one state with twice the available staffing of another.

Today's data tools simply are not suitable for a sector of this size and significance to public priorities. To address this information challenge, the National Center for Education Statistics and the Bureau of Labor Statistics should team up to determine the best way to capture more-relevant school staffing information rather than each continuing to produce data with such obvious shortcomings. Otherwise, policymakers and education leaders will continue to make decisions based on misperceptions.