Odd results from getCount

[AliciaMein]

I was QAQC'g my code results and found unexpected values.

My code queries point and polygon layers looking for how many negative test result are gathered before a positive.

for grid in grids:
        dates = []

        arcpy.SelectLayerByLocation_management("cwdpts4Anlys", "WITHIN", grid.shape, "", "NEW_SELECTION")
        positives = arcpy.SearchCursor("cwdpts4Anlys",' "PRION_DETECTED" = 1 ' , "", "SAMPLEDATE" )

        for positive in positives:
            dates.append(positive.SAMPLEDATE)
                
        earliest = min(dates)
        grid.setValue("FrstPosDte", earliest)

        posYear = earliest.year
        print posYear
                
        negDates = []
        
        arcpy.SelectLayerByLocation_management("cwdpts4Anlys", "WITHIN", grid.shape,"", "NEW_SELECTION")
        negatives = arcpy.SearchCursor("cwdpts4Anlys",' "PRION_DETECTED" = 0 ', "", "SAMPLEDATE")
        
        for negative in negatives:
            negDates.append(negative.SAMPLEDATE)

        if negDates:
                earliestNeg = min(negDates)
                print earliestNeg
                grid.setValue("FrstNegDte", earliestNeg)

                where = "EXTRACT(YEAR from {0}) <= {1}".format(arcpy.AddFieldDelimiters("cwdpts4Anlys", "SAMPLEDATE"), posYear)
                arcpy.SelectLayerByAttribute_management("cwdpts4Anlys", "SUBSET_SELECTION", where)
                result = int(arcpy.GetCount_management("cwdpts4Anlys").getOutput(0))
                print result
                grid.setValue("NgPriorPs", result)
        else:
                grid.setValue("NgPriorPs", 0)
                        
                
        grids.updateRow(grid)
        

It appears when checking the values for negatives prior to positive that the getCount includes the earliest positive (or it is always one more than expected.

Why?  Do I need to physically clear the selection before making the second "New_Selection"

Thanks,
Alicia