Risk difference (or Attributable risk): The risk difference describes the actual difference in the observed risk of events between the experimental and the control interventions or between the exposed and non-exposed groups. Mathematically, it is risk difference = (Disease risk in the exposed group) − (disease risk in the unexposed group). For example, 25% risk of lung cancer in exposed vs. 15% risk in unexposed give a risk difference of 10% or 0.1
Another way to see it is this:
The attributable risk is the number of cases attributable to one risk factor. Put another way, it is the amount we would expect the incidence to decrease if a particular risk factor were removed. For example, the incidence of lung cancer in the general population is 2/100 and in smokers is 10/100. The attributable risk of smoking in causing lung cancer is 8/100, assuming that the control is properly matched otherwise. AR = 10/100 -2/100 = 8/100
Resources:
http://handbook.cochrane.org/chapter_9/9_2_2_4_measure_of_absolute_effect_the_risk_difference.htm, Accessed 01/01/2016
http://www.statsdirect.com/help/basics/attributable_risk.htm, Accessed 01/01/2016