Wednesday, January 13, 2010

Divorce rates and Bixby Canyon

A few days ago, Nate Silver made some charts and graphs to help us remember that despite their incessant whinning, Liberals are just better people than conservatives.

In order to make Brian more comfortable, I'd like to add that this is also proven by the statistically valid study I just did here at PruHouse that shows liberals also like DCFC.

Sunday, January 10, 2010

Lying Liars and the Lies they tell

Yesterday I cracked on a combination of David Brooks and an Analyst from the Wall Street Journal. Both were pushing standard NeoCon / TeaParty talking points.

Brooks was claiming claiming the the Democrats were in trouble because a bipartisan cross section of America doesn't support Health Care Reform and believe Republicans have a better plan for improving America.

The Wall Street Journal Analyst was whinging about the decline of GM and the US auto industry and linking it to Unions, specifically the 30 year retirement benefit.

First, let me say that I disagree but understand that some people may have good faith conservative objections to Health Care Reform. What amazes me is that the majority of popular objectives are not in good faith. In this case, Brooks philosophical objection is that nay reform must be bipartisan and that since the current plans on the table are not it should be scrapped. He then sites polls showing that the majority of people disapprove of HCR. This is misleading as Brooks implications is that it's a binary preference: Either you support Health Care Reform or you don't. So if you disapprove of the current bills on offer you disapprove of HCR altogether. As numerous analysts with a track record of correctly interpreting data and getting measurable stuff right point out the polls actually show majority support for HCR. The high number of disapproval are in comparison to their preferred Bills. In otherwards, yes many American's would prefer a more ideallized version of the HCR bills being reconciled. But they prefer both current bills to no bill at all.

Now it's possible David Brooks doesn't understand statistical analysis. But his follow-up statement that Republicans enjoy more support than Democrats was backed up by Generic polls showing Republican approval vs. Dems trending up.

If David Brooks really beliebe what he was saying is true, his statement is very odd. Polls do show an increase in support for Republicans. But the increase is around 1-2% and Democrats still have around a 20% advantage. So Brooks statement is phrased in away that implies that Republican support has steadily increased and has passed the democrats while not actually saying that. This is indicitive to me that he knows the polls don't support his assertion but believes that if the statement is factually correct, well then he isn't lying because you drew that conclussion.

I call BS. When you say something designed to mislead someone else, but carefully parse your words so that no specific statement is actually untrue that's still willfully misleading someone.

As for the WS Journal Analyst. I'd just like to point out that all the data I could find puts the average American auto companies payroll at: 60% Blue Collar : 40% Management. I'd like to have some more specific numbers but they don't seem to be available. But that number should tell you something about whether it's the unionized workers who are the real drag on the auto industry.

Update: I wanted to provide links to the actual data. It's rather easy to find individual polls, but it's aggregaters like fivethirtyeight.com we want to watch since any individual poll can be an outlier. I'll keep looking in the event anyone actually cares. (I will say that I believe professional pundits like the ones I cap on are paid to have this data so get no slack). I did find this and wondered if maybe brooks just missed the date.

Update Update: Well look at that. RCP says I am wrong and shows the Republican generic house vote as ahead of the Dems. They rely very heavily on Rasmussen which has a significant lean Republican house effect (which isn't the same thing as saying it's wrong) but does make the trending analysis unfair since the shift in support shown in the RCP aggregate is most likely the result of Rasmussen's screen rather than a shift in actual public opinion (again, that's not to say that Rasmussen's methodology is wrong, just that it's likely that people answering "Dem" in September are likely being screened out because Rasmussen feels they are unlikely to vote).

I still stand by my post's point which is that an apples to apples comparison shows little change. In otherwards, I would argue that the polls show most people still prefer Dems to Republicans. Which was the baseline for Brook's point. What the polls are showing is likely voters seem to prefer Rebuplicans. It's not exactly the same thing, yet oddly supports one of Brooks other points, that Dems have electoral problems (which, you know, is not a secret). His lie is in trying to make a transitive statement that isn't supported.