Bias-free model fitting of correlated data in interferometry

In optical and infrared long-baseline interferometry, data often display significant correlated errors because of uncertain multiplicative factors such as the instrumental transfer function or pixel-to-visibility matrix. context model fitting, this situation leads to a bias in parameters. most severe cases can result fit lying outside range measurement values. This is known nuclear physics Peelle's Pertinent Puzzle. I show how arises interferometry determine that relative order square root component uncertainty times number measurements. It impacts preferentially large sets, those obtained medium high spectral resolution. then give conceptually simple computationally cheap way avoid issue: without covariances, estimate covariance matrix by error propagation using modelled instead actual data, perform fitting also more imprecise but unbiased be from ignoring correlations fitting.