|
Abstract:
|
8 research firms in Election 2008 conducted telephone polls using recorded voice. To be examined briefly: A) Did recorded-voice pollsters perform similarly, or dissimilarly, to each other? B) Did recorded-voice pollsters have smaller or larger errors than firms using other methods of data collection? To be examined more substantively: C) In what ways were un-weighted 2008 data more representative than un-weighted 2004 data? In what ways less representative? D) To what extent was weighting able to correct for coverage bias and non-response bias? To what extent could different weighting algorithms have further reduced error? E) In 2012, will there be a robust role, a diminished role, or no role for recorded-voice telephone surveys? F) Beyond 2012, is recorded-voice less likely, more likely, or no-more-or-less likely to survive than other methods of data collection?
|
