And the faulty polls roll on...
Problems in polling generally seem to be ignored when results are presented. While the information is completely available, anyone not familiar with statistics will miss some clues as to how a poll can go poorly.
Recently, this has happened again. A new poll seems to suggest that Hillary Clinton could win the 2008 presidential election. Hmm, this deserves further research.
It becomes obvious very quickly. The question asks how likely a voter would be to vote for Hillary, but there is no mention of any competition. While the 29% who would be very likely to vote for Hillary is probably fairly certain (probably a likely group to vote for almost any other Democratic candidate), there is a group of 24% who are "somewhat likely" to vote for her. So, Hillary would win based on these numbers; Susan Page assumes that the 24% is as good as the 29% in voting for Senator Clinton? Until another candidate is added to the equation, these numbers are meaningless.
I also question the stat that 1/3 of self-described conservatives would vote for Clinton. I wish I could say that I have an answer for this one, but all I have is a hypothesis; most conservatives refused to participate in the poll, meaning a skew from the more "moderate" conservatives. For now, I'm just going to call it a hunch that that number is faulty.
This seems to be a sort of rallying call to the Democrats, in the hopes of running Clinton in 3 years. But a few things should be remembered. Bad polls never change anything (just ask John Kerry the day after Election Day). A lot can happen in that period of time. And while I don't like negative politics, Hillary Clinton has made herself an easy target getting involved in many shady situations. She may have support now, but she is known as a very partisan candidate. The Democrats need to learn from prior elections and pick a more moderate candidate, perhaps Joseph Liberman or Mark Warner, if they want a chance at taking back the Oval Office in '08.
Recently, this has happened again. A new poll seems to suggest that Hillary Clinton could win the 2008 presidential election. Hmm, this deserves further research.
It becomes obvious very quickly. The question asks how likely a voter would be to vote for Hillary, but there is no mention of any competition. While the 29% who would be very likely to vote for Hillary is probably fairly certain (probably a likely group to vote for almost any other Democratic candidate), there is a group of 24% who are "somewhat likely" to vote for her. So, Hillary would win based on these numbers; Susan Page assumes that the 24% is as good as the 29% in voting for Senator Clinton? Until another candidate is added to the equation, these numbers are meaningless.
I also question the stat that 1/3 of self-described conservatives would vote for Clinton. I wish I could say that I have an answer for this one, but all I have is a hypothesis; most conservatives refused to participate in the poll, meaning a skew from the more "moderate" conservatives. For now, I'm just going to call it a hunch that that number is faulty.
This seems to be a sort of rallying call to the Democrats, in the hopes of running Clinton in 3 years. But a few things should be remembered. Bad polls never change anything (just ask John Kerry the day after Election Day). A lot can happen in that period of time. And while I don't like negative politics, Hillary Clinton has made herself an easy target getting involved in many shady situations. She may have support now, but she is known as a very partisan candidate. The Democrats need to learn from prior elections and pick a more moderate candidate, perhaps Joseph Liberman or Mark Warner, if they want a chance at taking back the Oval Office in '08.
posted by CR UVa at 10:42 PM
<< Home