Even within the newsroom, most people don't know the details of the work that goes into a particular story, photograph or investigation.

To get at those details I'm going to depart from the usual column format today for a Q&A with Charles Piller, an investigative reporter for The Bee who has spent much of the past year on a series of stories about problems within the California National Guard.

Piller exposed fraud involving up to $100 million in improper benefits paid to Guard members. Among his many stories he also reported that pilots and other Guard members double-dipped improperly, meaning they were paid two days' salary for a single day of work.

Piller's stories culminated recently in the abrupt removal of Maj. Gen. William H. Wade II from active service, an unprecedented action, because of controversy over improper double-dip earnings.

Let's hear how Piller reported his stories.

How did you get started on the Guard investigation?

I received a tip from an inside source who claimed that hundreds, if not thousands, of Guard members had improperly received millions of dollars in incentive payments.

You get many, many tips. How did you know this one was worth pursuing?

In this case, the initial tip came from an authoritative source – an auditor within the Guard who was able to provide thousands of internal documents to justify his concerns. He met a key test – he was in a position to know and could back up his claims.

Let's talk about the kind of document research you did for these stories. You weren't handed complete audits with findings, but analyzed stacks of pay records and other documents. What were the key records you reviewed?

This series comprised a wide range of stories and records. For the initial story, I worked with thousands of pages of raw pay records, plus partly completed analyses (in the form of Excel spreadsheets) created by several sources within the Guard.

I confirmed and refined the data in interviews with inside sources and experts on the topic. I also obtained confidential personnel documents, emails and policy memos. These allowed me to determine "who knew what when" – that is, to identify many leaders within the Guard who either overlooked, ignored or were involved in the fraudulent or improper payments.

In the pilot story, I received a completed audit, but it only told part of the story – how much money was improperly obtained by a subset of the pilots.

I used pilot schedules within that audit to prove that the pilots abused rules associated with "pilot rest" – a fundamental regulation set to ensure safety when operating $20 million jets, and just as important, the security of our airspace.

By consulting Air Force policies and vetting my assumptions with pilots in other regions and with outside experts, I found that the Guard pilots had gamed the system to maximize their incomes.

For the generals' double-dip story, after months of delay, I received raw data via a Public Records Act request and then cross-referenced it to data leaked to me from an inside Guard source. I combined the data into a spreadsheet with thousands of records – part of the grunt work to see who had likely received funds improperly.

Think back over the months of reporting. How many hours were devoted to number crunching?

Wow – hard to estimate. Including poring over raw data to make sure I understood it correctly, inputting data, creating my own spreadsheets, analyzing provided spreadsheets and fact-checking data, I'm going to say 100 to 120 hours for all the stories combined.

How do you ensure your findings are fair and accurate?

In general, I work "conservatively" – that is, I test my assumptions against all the reasons my thinking might be wrong. I look at all the innocent explanations I can imagine for what I see in the data. Only after I've ruled those out do I draw conclusions that improper or fraudulent actions might have occurred. I've never seen a story that was not strengthened by getting the views of those implicated in the wrongdoing. For that reason, and for accuracy and basic fairness, I go to the people implicated in the data – if needed, multiple times – to ask, encourage and urge them to give their perspectives. But unfortunately, they often decline to comment.

Fairness is more than offering people a chance to comment. I often give people a lot of information about what I see in their situations to help them understand the context of what I plan to write about them.

Finally, I go through an elaborate fact-checking process on every story, tracing every fact to a source (often, several sources) I feel sure are accurate and authoritative. I check every word and every number at least three times once the story has been completed, to ensure accuracy.

© Copyright The Sacramento Bee. All rights reserved.

Reach Executive Editor Joyce Terhaar at (916) 321-1004.

• Read more articles by Joyce Terhaar