I suspect that a (amount shared) value would be higher for n > 2 if d (divorce) = 0. Otherwise it would continue to go down.

]]>In order to treat it as a continuous variable and make sense in the way you suggest, then you could change the x-axis to involve time in some way (e.g. taking the time-average) and then something like fraction of parent-hours on Facebook (or average hours spent by parents on Facebook per week, introducing a constant factor of 168 to the x-axis numeric value) would be much more appropriate.

]]>