MaxedOutMama
Saturday, October 30, 2004
Where's the rage?
In general, if I have a customer who is reporting consistent difficulties, I find it better to visit their site and study the matter by sitting with the line users. I want to watch what the people trying to run transactions on the system are experiencing, rather than talk with the managers who normally report problems.
I find this approach works wonders, not because the managers I speak with are incompetent - most of them are extremely competent - but because they report data in a way which is meaningful to them. If the problem is an odd one, all I will hear is what made sense to the manager who speaks to me. Oddly, the more competent the manager, the more likely such a screening will take place. If, for instance, the users are reporting that sometimes something works and then it doesn't, a more competent and technically knowledgeable manager may believe they are hearing reports of user error and not pass these incidents along to me. However, such a pattern may indicate a network or data line problem, in which the system doesn't always have the capacity to complete transactions.
I wrote the above to disclose my own bias, which is to go straight to the lowest level of data whenever possible. Good reporting, in my opinion, has to do the same thing. Instead what I seem to have been reading is a great deal of meta-reporting, in which reporters report on what other reporters are writing. The problem with this is that if one or two original sources are wrong, the entire story can be reported incorrectly to an astonishing degree.
Since 2000 the Democratic party has maintained that in 2000 Jeb Bush stole the election for his brother, George Bush. I suppose during the heat of the recounts this seemed reasonable, but the independent recounts did not show it. Instead, they showed an astonishingly close election in Florida, which was similar to very close elections in other states in 2000. The spoiled ballot rates in various Florida locations were not higher than the spoiled ballot rates in other urban locations, and correlated well with illiteracy rates. In fact, the spoiled ballot rate in Florida was less than the spoiled ballot rate in Georgia. Florida also has a very decentralized voting system in which local Boards of Elections are elected by their counties, design their ballots, and count and report their votes. The counties which were highly controversial had Democratic election boards. Given these conditions, it's extraordinarily difficult to envisage a scenario under which a Republican governor could influence local vote counting..
The 2002 election confirmed my original skepticism. The national Democratic party campaigned intensely on the "rage" that voters felt at being disenfranchised. They said they would target Florida, and that in the 2002 elections voters would be given a chance to express their anger. The result was that Jeb Bush was reelected by a 13 points margin and that a majority of the legislative seats up for grabs were won by Republicans. So my original presumption was confirmed, but then I began to doubt my beliefs again when the same theme came back for this election. My reasoning was that after the devastating result of the 2000 elections naturally they would have checked their facts and confirmed them.
So, as a checkpoint, I looked up the voter registration figures for Florida the other day by party. If voters in Florida were harassed and abused, I assumed that registration should show a shift to the Democratic party. Here are the results taken from the 2000 tables, the 2002 tables, and the 2004 tables:
As you can see, over the last three elections the proportion of all voters registered as Republican vs Democratic has increased slightly instead of decreasing. This is especially significant because I keep hearing about the massive increase in Democratic registrations in the national press, and one of the states that has been cited is Florida. The Horserace blog has had a lot to say about that, if you're interested.
So then I wondered, what about the allegations of suppression of the black vote? Florida keeps statistics on race and registration by party (which frankly skeezed me out a bit, I guess you check a box or something). I looked at the numbers of people listed as "Black" who registered for the Republicans, Democrats, and No Party for 2000, 2002, and 2004:
(There are a lot of different parties in Florida, but 99% of all blacks are registered under three categories)
As you can see, all the controversy has had very little effect at all. Expressed as a percentage, Democratic and Republican registrations have dropped and no party (Independent) have gained. I'd have to say that this numbers don't support the national news stories I've been reading at all. This sort of thing is why I have lost faith in papers and networks as a source of reliable news.
I find this approach works wonders, not because the managers I speak with are incompetent - most of them are extremely competent - but because they report data in a way which is meaningful to them. If the problem is an odd one, all I will hear is what made sense to the manager who speaks to me. Oddly, the more competent the manager, the more likely such a screening will take place. If, for instance, the users are reporting that sometimes something works and then it doesn't, a more competent and technically knowledgeable manager may believe they are hearing reports of user error and not pass these incidents along to me. However, such a pattern may indicate a network or data line problem, in which the system doesn't always have the capacity to complete transactions.
I wrote the above to disclose my own bias, which is to go straight to the lowest level of data whenever possible. Good reporting, in my opinion, has to do the same thing. Instead what I seem to have been reading is a great deal of meta-reporting, in which reporters report on what other reporters are writing. The problem with this is that if one or two original sources are wrong, the entire story can be reported incorrectly to an astonishing degree.
Since 2000 the Democratic party has maintained that in 2000 Jeb Bush stole the election for his brother, George Bush. I suppose during the heat of the recounts this seemed reasonable, but the independent recounts did not show it. Instead, they showed an astonishingly close election in Florida, which was similar to very close elections in other states in 2000. The spoiled ballot rates in various Florida locations were not higher than the spoiled ballot rates in other urban locations, and correlated well with illiteracy rates. In fact, the spoiled ballot rate in Florida was less than the spoiled ballot rate in Georgia. Florida also has a very decentralized voting system in which local Boards of Elections are elected by their counties, design their ballots, and count and report their votes. The counties which were highly controversial had Democratic election boards. Given these conditions, it's extraordinarily difficult to envisage a scenario under which a Republican governor could influence local vote counting..
The 2002 election confirmed my original skepticism. The national Democratic party campaigned intensely on the "rage" that voters felt at being disenfranchised. They said they would target Florida, and that in the 2002 elections voters would be given a chance to express their anger. The result was that Jeb Bush was reelected by a 13 points margin and that a majority of the legislative seats up for grabs were won by Republicans. So my original presumption was confirmed, but then I began to doubt my beliefs again when the same theme came back for this election. My reasoning was that after the devastating result of the 2000 elections naturally they would have checked their facts and confirmed them.
So, as a checkpoint, I looked up the voter registration figures for Florida the other day by party. If voters in Florida were harassed and abused, I assumed that registration should show a shift to the Democratic party. Here are the results taken from the 2000 tables, the 2002 tables, and the 2004 tables:
| Year
|
Registered Dem
|
Registered Rep
|
Dem/Rep
|
Rep/Dem
|
| 2000
|
3,803,081
|
3,430,238
|
1.109
|
.9020
|
| 2002
|
3,958,910
|
3,599,053
|
1.010
|
.9091
|
| 2003
|
4,261,249
|
3,892,492
|
1.095
|
.9135
|
As you can see, over the last three elections the proportion of all voters registered as Republican vs Democratic has increased slightly instead of decreasing. This is especially significant because I keep hearing about the massive increase in Democratic registrations in the national press, and one of the states that has been cited is Florida. The Horserace blog has had a lot to say about that, if you're interested.
So then I wondered, what about the allegations of suppression of the black vote? Florida keeps statistics on race and registration by party (which frankly skeezed me out a bit, I guess you check a box or something). I looked at the numbers of people listed as "Black" who registered for the Republicans, Democrats, and No Party for 2000, 2002, and 2004:
(There are a lot of different parties in Florida, but 99% of all blacks are registered under three categories)
| Year
|
Dem
|
Rep
|
NoParty
|
All
|
%Dem
|
%Rep
|
%NoParty
|
| 2000
|
792,168
|
49,442
|
86,225
|
934,261
|
.8479
|
.0529
|
.0923
|
| 2002
|
866,121
|
52,775
|
100,236
|
1,027,817
|
.8429
|
.0513
|
.0975
|
| 2004
|
1,008,382
|
62,512
|
186,857
|
1,270,013
|
.7940
|
.0492
|
.1471
|
As you can see, all the controversy has had very little effect at all. Expressed as a percentage, Democratic and Republican registrations have dropped and no party (Independent) have gained. I'd have to say that this numbers don't support the national news stories I've been reading at all. This sort of thing is why I have lost faith in papers and networks as a source of reliable news.
