{
  "id": "2003-julyaugust-data-for-covering-workers-compensation",
  "slug": "2003-julyaugust-data-for-covering-workers-compensation",
  "url": "https://dwillis.github.io/uplink/article/2003-julyaugust-data-for-covering-workers-compensation.html",
  "year": 2003,
  "month": "July/August",
  "monthNum": 7,
  "headline": "Data for covering workers, compensation",
  "kicker": "SPOTLIGHT: JOBS, PAY & PERKS",
  "author_name": "David Herzog",
  "author_title": "",
  "affiliation": "NICAR",
  "summary": "Journalists can access three key databases—payrolls, pensions, and check registers—to investigate public employee compensation, benefits, and spending patterns.",
  "keywords": [
    "payrolls",
    "pensions",
    "databases"
  ],
  "topics": [
    "Investigative Methods",
    "Data & Databases",
    "Government & Politics"
  ],
  "technologies": [],
  "full_text": "Journalists can use several key databases to cover public employees, their salaries and benefits. Here are a few of the most valuable and what you can expect to find inside them:\n\nPayrolls\n\nPayrolls list the names, departments, jobs and pay of government workers. The workers' hiring dates, pay rates, work status (part- or full-time) also should be included. Check the pay field carefully because agencies sometimes mix the hourly pay rates of wage earners in the same field with the monthly or weekly salaries of professional workers. Also, make sure you request the most detailed pay records you can get so you can look at overtime pay.\n\nDepending on your state's open records law, you might be able to get even more detailed information about workers, including home addresses and dates of birth. The more detailed information you can get, the better.\n\nRequest this data from agency human resources or personnel offices. Sometimes finance offices maintain the data. Update this data every year so you can build an archive that shows salaries over the years.\n\nPensions\n\nThese databases list the names of pensioners, last date of employment and amounts of annual pensions. Many of these databases include the addresses of the pensioners and the types of pensions (disability or retirement).\n\nDepending on the complexity of the retirement system, there may be several types of pensions with distinct rules and cost-of-living adjustments.\n\nIf there is more than one pension system, make sure your data includes a field that indicates the type of pension.\n\nJournalists have used pension data to show how some pensioners get more money off the job than when they worked. They've also used it to project how much pensioners will collect in the future.\n\nYou can request this data from agency pension boards. If they do not keep the data, try the human resources, personnel or finance offices.\n\nCheck register\n\nThe check register can be a great place to find information about public agency worker travel. A check register is a database that contains a record of every check issued by a government agency.\n\nRegisters often include the check number, name of person or entity that the check was issued to, date, amount, department and purpose of payment (often a code). Sometimes the data includes an employee identification number.\n\nJournalists can use this data to scour for employee travel, which may or may not be legitimate, and reimbursed business expenses. In addition, journalists can use the data to look for payments to consultants and contract workers who are not paid through the usual payroll channels.\n\nTry requesting this data from the treasurer, comptroller or another government fiscal officer.\n\nContact David Herzog by e-mail at dherzog@nicar.org.\n\nuniversity's medical schools, a finding that was true back in 1990.\n\nThe breakdown was almost too good and too easy to be true. So I did more reporting, calling officials at the medical schools and experts on medical school salaries. I found that the database did not tell the complete story. At face value the highest-paid state employee, a renowned surgeon, made about $600,000 a year from the state. But, as my reporting later revealed, most of the medical school deans and professors also practice medicine, thereby earning most of their so-called state salaries from private sources.\n\nAfter culling through names of individuals who fit this category, I found that the highest-paid employee, whose $359,570 salary was completely state-funded, was the chancellor of the University System of Maryland. That revelation changed the focus of the story, especially since the salary and deferred compensation package of the incoming chancellor had recently raised questions at a time when the state was trying to close a $1 billion deficit.\n\nInstead of analyzing the database in the aggregate, I decided to focus my story on the chancellor's salary and look at why medical school personnel received such high six-figure salaries, outranking the governor, judges and heads of state agencies.\n\nI also decided to write a separate story on the two coaches whose state salaries did not accurately reflect their overall compensation package. The database showed that the two coaches, Friedgen and Williams, made $183,820 and $207,584, respectively, when in fact, they were making nearly triple the pay of the chancellor. Because I was still waiting for my request to be answered at the time of deadline, I relied on published reports of the coaches' salaries to write my story. The story also touched upon the controversial issue of paying college coaches million-dollar salaries.\n\nThe stories I found went further than the original intent of my reporting: to update the 1990 analysis of the 100 highest-paid state employees. If I had taken the database at face value, I would have incorrectly written that medical school personnel received the top salaries, a trend unchanged from 1990. Instead, I found that the university chancellor was the highest-paid employee whose salary was paid completely from state funds.\n\nThe salary database did not require complicated queries or have many quirks. But the project was worthwhile because I learned a valuable lesson: No database is a substitute for good, thorough reporting. I also learned it's crucial to know the state's public information law so you can be armed when going into battle with lawyers or agency administrators who don't want to give up information easily.\n\nAs a final note, a few weeks after the story ran, the University of Maryland athletic department responded to my request and provided more details of the coaches' salaries, such as the amount of bonuses tied to performance. It pays to be persistent.\n\nContact Hanah Cho by e-mail at hanah.cho@latimes.com.\n\nSince Charlotte County stores delinquent taxpayer information on a mainframe system, I asked for the data in ASCII. They provided the entire database for free. I imported the text into Microsoft Access 97.\n\nNot long after beginning a search through all the records, I discovered my first roadblock. The tax collector kept detailed information on each delinquent, but surprisingly, no one had tallied up how much people owe in back taxes over the years. In fact, I was told no one had ever done a countywide calculation showing how much each delinquent owed in penalties over multiple years. Also, because interest kept adding up, the dollar figure changed every day. But the tax collector's database showed only how much delinquents originally failed to pay. To get an accurate picture of the top tax delinquents, I needed to calculate the back tax information for more than 40,000 records.\n\nPrep work\n\nAfter talking with the information technology office for the tax collector, began writing a program in Access to calculate back taxes. To accomplish our task, we had to first calculate how many months had passed since each delinquency began. The more months without a payment, the more in back taxes owed.\n\nI created two new fields in Access and ran the calculations to show the number of delinquent months and the monthly interest rate for each delinquent. We used this information to calculate how much each delinquent owed in back taxes, but I still had to add a $4.25 penalty to each record to cover the cost of advertising the delinquency in a local newspaper. After completing the work, I finally thought we had all the information we needed.\n\nIt didn't take long for us to discover I needed to merge the file with another county database to get an accurate picture of local delinquents. The database of tax certificates showed only the name of the person who owned the property the year its taxes became delinquent. We needed a second database to see the name of the current owner. Florida law states the current owner of property is responsible for all back taxes.\n\nI called the property appraiser's office and arranged to receive the electronic list of properties in Charlotte County. I merged the tax collector's information with the property appraiser's list by joining the files using the \"strap number\" field that serves as a property identifier. Finally, with information about current property owners in hand, we were able to generate the true list of top tax delinquents.\n\nBeyond the data\n\nOur queries found that one company had piled up more than $950,000 in back taxes on more than 500 properties. With a quick visit to the county clerk's office, we discovered the county's top tax delinquent was going to have a very difficult time paying its bill. Court records showed the company had gone bankrupt and hadn't paid its taxes for seven years in a row. Despite this fact, the county continued to spend tax dollars sending out delinquent notices and advertising its delinquency in the newspaper year after year.\n\nThis struck us as odd, since Florida law allows county governments to auction property if the owner remains delinquent for more than two years. For some reason the county was bypassing the option to auction the properties off, a move that could recoup money and return the real estate to the tax rolls.\n\nBy law, the tax collector must spend tax dollars advertising delinquencies three times a year for each piece of property that qualifies. We began more detailed queries, searching for all the delinquents that hadn't paid for years but were still draining the county's coffers. We calculated the tax collector had spent $251,712 since the mid-1990s trying to collect from deadbeats it could have cracked down on long ago.\n\nWe eventually uncovered a total of $8.3 million in uncollected taxes across the county since the mid-1990s. It's money that wasn't getting to essential services such as schools, fire departments and road construction crews. More than $1 million of that would have gone to the Charlotte County Fire Department. Fire Chief Dennis DiDio told us it could have added new fire trucks, firefighters and equipment to fight bio-terrorism.\n\nPutting it together\n\nFurther research revealed Florida law requires county governments to auction off all land worth more than $5,000, if the land has been delinquent for more than two years. If it's worth less than $5,000, counties can decide for themselves if they want to go after the property. The $8.3 million in uncollected revenue we discovered was largely made up of properties worth less than $5,000.\n\nWe found the county failed to put together all the pieces on its own. It looked only at how much each delinquent owed on individual pieces of land. No one ever thought to keep track of just how much delinquent land existed under one person's name. We eventually discovered systematic proof that the county wouldn't crack down on a majority of delinquents who owned land assessed at $5,000 or less.\n\nWe took our findings to the tax collector and the real property manager for Charlotte County. The tax collector told us she wanted to immediately respond by auctioning off all the delinquent land we discovered. However, that wasn't her decision to make. The real property manager decides which land to sell at tax auction. When we showed him our findings, he pledged his support for the tax collector's idea. The real property manager told us he would begin the process of auctioning off all the delinquent properties that qualified for the program.\n\nThe county clerk has agreed to postpone some of its normal fees to speed the sale of the large number of properties that will be auctioned off on the courthouse steps. The Charlotte County tax collector and the real property manager credit WBBH for discovering a potential source of millions of dollars in uncollected revenue. We're told Charlotte County will complete its auctions over the course of the next year.\n\nContact Mark Greenblatt by e-mail at mark.greenblatt@nbc-2.com\n\nWhy is steak sauce important to a political campaign? That was the key element of my first computer-assisted reporting project: A question.\n\nFinding New York's campaign finance database online was easy. After that, I had to fall back on my Boot Camp training. Two pieces of advice became immediately useful - even critical - in doing the story:\n\nFirst, find a buddy to review your work. In my case, I tapped Brad Heath. He's a reporter in Detroit now, but he was my predecessor as Binghamton's CAR person and understands my experience as a reporter and my virtually nonexistent database skills. In fact, he knew the data I was working with, too, because he had briefly played with it.\n\nSecond, give the database to the source. The campaign treasurer had done all the paperwork in filing the information, and it was in his interest to make sure I got my facts straight. He found a couple of errors in my queries. And with more than 3,000 potential sources, I still kept total control of what the story said.\n\nI learned a few other lessons.\n\nDo everything two or three times. It wasn't until I exported my query results to Excel and checked the math that I realized the numbers didn't add up, and it was because of a problem in my query. At Boot Camp, they called these data integrity checks, and they are important.\n\nReview the data. The quick and easy task is to find the extremes in your data, and that frequently can be useful. But go through the rest of it, too. You'll come up with more questions. Why, for example, couldn't I find the local Republican leaders contributing to the area's most prominent Republican, but I could find their spouses? You're not going to notice that unless you know both the data and your subject.\n\nI'm primarily a beat reporter so that kind of analysis was easy, even fun, once my co-workers and I made a game of finding how many names we recognized. I don't know how a reporter could do that coming into the data cold.\n\nI also had to learn what the data wasn't telling me, which meant I had to learn what it was supposed to say. Candidates for New York's state legislature generally contribute to a party \"housekeeping\" fund to pay for legal advice, polls and whatnot. Other politicians had contributed to it, according to my analysis, but the state senator hadn't. Either he was the state's biggest political cheapskate (unlikely) or there was another fund out there the data didn't have a record of.\n\nI had to learn to keep focus on my story - how the senator's campaign spent its money. There were plenty of other tempting angles to pursue. Why was the ambassador to Grenada my boy's biggest individual contributor? And was the fact that he was indicted about 15 years ago (and later exonerated) on charges alleging he blew up trailers of a union protesting one of his facilities relevant? Fascinating stuff, but it wasn't part of the story I had in mind. Maybe next time.\n\nThat's the lure of a good database. It gives you a garbage load of ideas. You can't do them all. Even if you could, you probably shouldn't. Overkill can be a nasty habit, and with all the time I sank into this project, it was a temptation.\n\nThe story described how campaigning works and what it means to a community.\n\nAnd that was the toughest lesson of all. Making the database work took a solid 20 hours - as much as I normally devote to a basic, no-frills enterprise package. And because a database generates about three numbers and a nut graf - and maybe some charts or lists or facts for a box - that meant that once I found the good stuff, I still had 15 or 20 hours of work remaining.\n\nCAR stories are inexpensive compared with other investigative stories of similar depth, but they're inefficient compared with traditional methods. That can be an issue at a community newspaper, where cutting somebody free from the daily need to feed the beast can be a burden on everybody else. Fortunately, Heath taught my editors that so I didn't have to.\n\nThrough all of this, Heath double-checked my work and cheered me up when I felt dumb. With his help, I learned to rewrite queries to make them as simple as possible. The easy",
  "provenance": {
    "extracted_by": "anthropic/claude-sonnet-4-6",
    "extracted_at": "2026-07-05",
    "verified": true
  },
  "similar": [
    {
      "slug": "2002-march-april-mining-workplace-data",
      "score": 0.238
    },
    {
      "slug": "2006-july-august-extra-extra-your-guide-to-the-latest-investigative-work",
      "score": 0.217
    },
    {
      "slug": "1996-january-bits-bytes-and-barks",
      "score": 0.175
    },
    {
      "slug": "1995-december-who-profits-at-nonprofits",
      "score": 0.167
    },
    {
      "slug": "1996-january-putting-the-heat-on-pensions",
      "score": 0.157
    }
  ],
  "html_url": "https://dwillis.github.io/uplink/article/2003-julyaugust-data-for-covering-workers-compensation.html",
  "json_url": "https://dwillis.github.io/uplink/article/2003-julyaugust-data-for-covering-workers-compensation.json"
}