This is a really old feature request, but I had a thought of how we might handle
it.  Instead of doing a visual drag, what if we made use of the filtering
functionality?  So a user first sets a filter, and then there's a menu/toolbar
command along the lines of "Set Outcome of visible rows to [True|False|Rem|Imp|Con]"

This would accomplish the same functionality, but wouldn't require the
Excel-like drag-to-recode.  And it may even be more efficient (although less
intuitive), since the data set wouldn't have to be sorted properly.

Unfortunately, I don't know if this is going to be possible. The reason is that 
because we wanted persistent widgets in the outcome column (have the drop box 
always there as opposed to just when you double click the cell) we lose the 
ability to select multiple cells in this column.

I will keep looking into this.

Here's an example of where this would be useful.  In the fascist
violence data set, there's one column for "Triangle" and another
column for "North."  The observations (rows) in this data set are
incidents of fascist violence.  The North column indicates whether the
incident took place in Northern or Southern Italy.  The "Triangle"
column indicates whether the incident took place inside or outside of
Italy's "Industrial Triangle" (an area of Italy with a large number of
manufacturing industries).  Here's the thing--the Industrial Triangle
is located in Northern Italy.

Now think about what this means for the truth table.  Any row that has
False for North and True for Industrial Triangle can't exist.  This is
what the "Impossible" outcome is designed to address.  (A remainder,
in contrast, is a row that doesn't have any empirical instances but is
logically possible.)