Originally printed in the Milwaukee Journal-Sentinel, December 19th, 2011

I was surprised to read PolitiFact's assessment of the accuracy of a recent statement of mine. I pointed out that "President Obama has increased the federal workforce 192,000 individuals - about 10% since he's taken office - while we've lost 2 million jobs."

PolitiFact acknowledges that "Johnson is on target on the 2 million jobs lost overall, and accurate or very close on the federal workforce trend - at least using FedScope," but rates my assertion "half true." But Politifact's justifications for not acknowledging that this is true do not hold up to scrutiny.

First, PolitiFact says that I am "comparing apples and oranges"' in my utilization of FedScope data for federal payrolls and Bureau of Labor Statistics data for private sector payrolls. But if your goal is to produce the most accurate counts of both, you use the most accurate data for each. For federal payrolls, the most accurate data comes from FedScope.

As PolitiFact itself notes, "FedScope is considered reliable and is used by researchers who prefer an actual head count to the sample-based approach" of the BLS. This is precisely the point: to get an actual head count. Is PolitiFact really arguing that relying on a statistical sample would be more "true" than an actual count?

FedScope does not provide data on private sector payrolls, so it is impossible to meet Politifact's "apples to apples" standard by using FedScope's data for both. Instead, I used BLS data for private sector payrolls because it is the gold standard for employment levels, unemployment rates and economic trends. With an economy worth $15 trillion, BLS must use best-available and correlative data collected through survey methods.

PolitiFact also implies that I should have included Postal Service payrolls in my estimates of the federal workforce. The Postal Service is a quasi-independent federal agency, whose employee levels are outside the direct control of both the administration and Congress. Had I included those payroll levels in my counts, PolitiFact likely would have criticized me for suggesting that the president had control over them. This is an inappropriate argument.

I stand by my statements, which are justified. PolitiFact seems inclined to rate my assertion as half true out of concern that it's not fair to hold the president accountable for his policies. It's legitimate to argue with my conclusion that those policies have made our economic situation far worse. But it is incorrect to conclude that my facts are anything but completely true.

Mr. Johnson is a Republican senator from Wisconsin.