We investigate the hidden gender gap, the unexplained wage inequality that could be related to direct discrimination. A General Linear Model combining the variables of the 2018 Spanish Structure of Earnings Survey is estimated to construct typologies with maximum pay gap. We build a matrix with all combinations of explanatory variables including sex (>3 million rows). We predict salary with the model and calculate the wage gap between rows whose only difference is sex. We order rows selecting those with the maximum pay gap (≥ 99th percentile) to identify typologies. Therefore, we are able to adjust the calculation of the hidden gap (13.1%) and identify the type of occupation where it mostly occurs (permanent jobs, in companies of 1 to 49...