A response to Sherley: FISHERIES/2020/JUL/SWG-PEL/53REV

Many of the arguments advanced by Sherley in document FISHERIES/2020/JUL/SWG-PEL/53REV are challenged, offering counter arguments on a point by point basis. An important component is a mathematical demonstration that the disaggregated data analysis approach of Sherley to provide estimates of closure effects cannot provide greater estimation precision than its aggregated data counterpart. Furthermore results in 53REV document are shown to indicate that Sherley’s choice of a random effects structure for his earlier chick condition response variable analysis is sub-optimal, and leads to pseudoreplication bias (standard errors that are too small). The approach put forward in 53REV to combine the results for an island closure effect across different response variables is also queried.