If you're fitting a model to observations, they'll have something of an inverse (but of course non-linear) relationship. So if your virulence estimates are wors

If you're fitting a model to observations, they'll have something of an inverse (but of course non-linear) relationship. So if your virulence estimates are worse than your transmissivity estimates, or vice versa, the bad measurement is going to push the model out of whack.