In March of 2019 we noticed some discrepancies in a gender wage gap article by Warren Fiske of PolitiFact Virginia. We sent messages asking about the discrepancies and received replies from Fiske.
Notably, PolitiFact did make at least three changes to its story in apparent response to our messages.
The email exchange occurs below, excluding some extra messages stemming from PolitiFact Virginia publishing an inactive email address for Fiske.
We provide explanatory notes between the email images.
Note: Craig Carper was listed as Fiske’s editor for the fact check.
The change Fiske mentions was made without any correction notice.
Note that Fiske does not address the discrepancy in PolitiFact Virginia’s reporting. The fact check said the Bureau of Labor Statistics did not provide data in job category subdivisions for men and women broken down by race. And the fact check also mentioned the advantage white men possess when looking at the data for different categories of jobs.
Most likely we did too little to make the discrepancy clear.
In my next reply to Fiske we focused on getting enough information to reproduce his math.
Fiske’s reply was gracious and specific.
Fiske’s instructions apply to a specific page at the Bureau of Labor Statistics website listed, but not hotlinked, in the reference section of PolitiFact Virginia’s fact check.
In a subsequent messages to Fiske we questioned how we obtained his original total number of job categories (29). Not long after that inquiry, PolitiFact Virginia updated its article again, changing the number of job categories to 22.
After that we questioned whether PolitiFact Virginia had performed the correction in keeping with PolitiFact’s statement of principles. The article was subsequently updated with a correction notice. However, the PolitiFact Virginia article to this day does not appear on PolitiFact’s list of corrected or updated stories.
PolitiFact populates that list by adding a subject tag to the corrected item, “Corrections and Updates.” The PolitiFact Virginia story continues to lack that tag.
As for the original discrepancy we saw in PolitiFact Virginia’s reporting?
We think Fiske meant to assert that the BLS breaks down data for men and women in the 22 broad categories of work by race (he wrote “sex” instead).
Final note: This resource page does not address the central problem with the PolitiFact Virginia fact check. We intend the notes to simply offer the context needed to understand the email exchange.