Simulated analysis of observational selection bias. The panels in the top row represent N_test (with a larger number of sample sets set to 10 000), while those in the bottom row represent N_obs (the same number of sample sets as the observational data). The left panels show the results of uniformly random sampling within the range, and the middle panels show the random sampling in a way that uses observational data but applies the biases of transit detection. Points represent the distribution of simulated data, with the color bar indicating the distribution of P_transit. Black and orange dashed lines indicate the fit results (γ₁ of the correlation between the inner edge and stellar mass). Each of these four panels represents the median result from 1000 simulation runs. The right panels display the distribution of correlations between M_★ and a_in obtained from 1000 simulations, with meanings consistent with Fig. 4.