Intelligence Squared debates favor initial loser, effect on climate debating is less clear

Intelligence Squared debates measure who wins not by which side has more votes at the end but by which side picks up the most votes compared to a pre-debate poll of the audience. Following up on a comment I left at Greenfyre that the initial loser generally picks up more votes, I put the results of the last 25 debates into an Excel file copied below.

In 16 cases, the initial loser picked up more votes, in 6 cases the reverse ocurred (written in bold below), and in 3 cases the start was nearly tied or the vote pickup was tied (shaded gray). Maybe not conclusive, but it doesn't seem like a random distribution.

The two climate debates both had "skeptics" as initial losers, both of whom picked up more votes. I didn't try to figure out if they beat the average spread, but I'd guess one was about average, and the other slightly worse than average. I don't think the results tell much about whether formal debates with skeptics is a good or bad idea.

Two tangents: first, mischief-makers on the right will notice that conservative positions picked up more votes than liberal positions, but again we hit the same issue that conservative positions tended to start with fewer votes. If you look only at the six cases where the initial winner also won more votes, there's no ideological consistency between red-shaded winners and blue-shaded winners.

Second, Intelligence Squared needs to come up with new success criteria. Their current one doesn't make sense.

(BTW, anyone who wants a copy of the Excel file should just email me.)

for art	32	55	virtual tied start
against art	30	33	virtual tied start
for cabon not worth the effort	16	42
against carbon not worth the effort	49	47
for bush worst	65	68
against bush worst	17	27
for google evil	21	47
against google evil	31	47
for guns reduce	13	27
against guns reduce	60	64
for US is winning	20	36
against US is winning	54	53
for universal health	49	58
against universal health	24	34
for legalize sale of organs	44	60
against legalize sale of organs	27	31
for islam is radical	46	73
against islam is radical	32	23
for tough interrogation necessary	46	40
against tough interrogation is necessary	35	53
for US should police	24	46
against US should police	44	48
for performance drugs	18	37
against performance drugs	63	59
for aid harmful	24	41	tie gain
against aid harmful	34	51	tie gain
for end affirmative action	34	39	liberal countertrend
against end affirmative action	44	55	liberal countertrend
for Russia becoming enemy	41	47
against Russia becoming enemy	23	41
for stop undocumented	42	60
against stop undocumented	34	37
for spread democracy is bad	46	55
against spread democracy is bad	36	40	l
for China is trouble	41	35
against China is trouble	37	59
for surveillance better than 9/11	41	40
against surveillance better than 9/11	37	55
for warming not crisis	30	46
against warming not crisis	57	42
for US too religious	67	70	liberal countertrend
against US too religious	24	24	liberal countertrend
for Hollywood is antiUS	40	35
against Hollywood is antiUS	35	59
for Hamas is terrorist	60	53
against Hamas is terrorist	19	29
for offensive expression	78	83	tie gain
against offensive expression	11	16	tie gain
for tolerate nuclear Iran	26	37
against tolerate nuclear Iran	47	52