I believe that the CPS result missed by four times the margin of error due to a sampling error, which is caused by the act of taking a sample. Sampling errors cause sample outcomes to be different from the outcomes of a census. In this case, 63.6% of the sample claimed that they voted in the 2008 presidential election, but actually 61.6% of the adult population voted in that election. Therefore, the inaccurate margin of error is due to a random sampling error, which is the deviation between the sample statistic and the population parameter caused by chance in selecting a random sample and the margin of error in a confidence statement includes only random sampling error. This conveys that the household survey might have taken a sample by chance
