This was a great catch and a computation error on my part. I️ added rather than averaged work increases among women and the top 5% earners into the mix. I’ve taken out that section since you’re right, it’s not supported. I️ appreciate your close reading.