Charles Franklin on PPP Random Interview Deletion Method
- — Charles Franklin (@PollsAndVotes)Fri, Jul 19 2013 02:28:20Reaction to PPP's poll balancing method at Pollster http://www.huffingtonpost.com/2013/07/16/mcauliffe-cuccinelli-poll_n_3606898.html?utm_hp_ref=@pollster …ost.com/2013/07/16/mcauliffe-cuccinelli-poll_n_3606898.html?utm_hp_ref=@pollster Must be statistically inefficient but not necessarily biased. 1/2
- — Charles Franklin (@PollsAndVotes)Fri, Jul 19 2013 02:28:20Post-stratification weights in PPP poll should achieve same goal but keep all cases-Should give equivalent estimates as random delete 2/2
- — Charles Franklin (@PollsAndVotes)Fri, Jul 19 2013 02:28:20Post-stratification weights would give estimate of design effect as empirical matter. Doubt design effect can be calculated with deletions.
- — Charles Franklin (@PollsAndVotes)Fri, Jul 19 2013 02:28:21As others have said, DKos/PPP provide raw data, including deletions so this is an empirical question that can be answered. Not mere opinion
Did you find this story interesting? Be the first to like or comment.