I have written code that can detect and label dropouts in my experiment. I want to automate the decisions of dropouts so that my large group experiment can continue in a reasonable way, even if dropouts occur. 

To do this, I want the decision of my dropout to be the median of the decisions of all remaining participants. I know how to select the median etc.

My issue: 

The way I am dealing with dropouts is to declare a player a dropout if a timeout happens. When this initially happens, I can easily automate the dropout's decision as described since all other player who are not dropouts will have made decisions before the timeout happened. 

In subsequent periods, I set the timeout time for dropouts to 1 second and N seconds for non-dropouts. The problem is that the dropouts are now 'making decisions' before the remaining players and so I can't select the median of players' decisions for dropouts. 

Is there a way I can make my dropouts 'wait' until all actual players make decisions and then chose the median of these decisions as the dropout's decision?

Put a wait page, and set the dropouts' decisions in after_all_players_arrive.