In order to test the effectiveness of the debate of the candidates for the president, researchers randomly contacted 1000 people on the phone. An attempt was made to contact all 1000 one day before the debate and ask their voting intentions. Only 910 were actually reached by phone; 80 out of the 910 people refused to participate. One day after the debate, 810 of the 910 people were reached and again asked their voting intentions. In addition they were asked if they had seen the debate on television. The before-after change of the viewers and non-viewers were compared to conclude which candidate was more effective in the debate. What kind of experimental design is being used