Thursday, January 10, 2008

New Hampshire Machine Count Bias?

There has been some discussion about the difference in Hillary Clinton's vote in New Hampshire with machine counted votes versus hand counted votes -- see here. Overall, Hillary got 39.0% of the NH vote. But she received 40.1% of the machine counted vote and only 34.7% of the hand counted vote.

When my son first showed me this, I jumped to the obvious conclusion that towns in NH that have machines counting the votes are different from town that count by hand. A classic statistical problem: correlation does not prove causation, and the old omitted variables problem. Hillary probably does better with voters who live in towns that count votes by machine. If you look at the data, it appears that "machine count" is the variable determining the Hillary vote, when in fact it is an underlying variable -- wealth, race, educational levels -- that really determines the vote difference, and that we are not measuring. So "machine count" is simply "picking up" the effect of the variable(s) omitted from the analysis.

Ah, the wonders of technology. My son showed me a website that did some analysis by town size. Sure enough, Hillary does better in large towns, and large towns tend to do more machine counting. So, it appears that the relationship between Hillary's vote and METHOD (the vote counting method) is just picking up the underlying relationship between Hillary's vote and TOWN SIZE. TOWN SIZE itself is a proxy for things such as wealth, education, etc.

Even more wonders of technology. I had my son collect the voting data on the 220 towns of NH. Using PERL, he downloaded the data to Excel for me to use in about 15 minutes. So I had four variables on each town in NH: TOTAL VOTE (a measure of town size), CLINTON PERCENT, OBAMA PERCENT, and METHOD (1 for MACHINE COUNT, 0 for HAND COUNT).

I quickly ran a univariate regression of CLINTON PERCENT on METHOD: Sure enough, the regression equation is

CLINTON PERCENT = 33.68 + 5.64 METHOD

with a standard error of 1.01 on METHOD (t-statistic of 5.58).

That fits with the univariate analysis of the data as presented earlier. Sure enough, Hillary seems to do better when the vote is counted by machine!

I was sure that when I added TOTAL VOTE to the regression, the coefficient on METHOD would drop in size and in statistical significance. This does not HAVE to happen with correlated variables such as METHOD and TOTAL VOTE, but I was pretty sure it would.

Here are the multiple regression results:

CLINTON PERCENT = 33.56 +5.08 METHOD + .00028 TOTAL VOTE

with a standard error on METHOD of 1.12 and on TOTAL VOTE of .00024.

Amazing! METHOD continues to be the variable carrying the weight of the data. Town size is statistically insignificant and the method of counting accounts for most of the variation in Clinton's vote difference.

Hmmm.....developing....


UPDATE: I highly suspect, still, that METHOD is simply correlated with some underlying real determinant of the Clinton vote. If I can get more data by town, I will run those models. It is even possible that my variable TOTAL VOTE is not a good measure of town size, as it involves voter turnout as well. What would be really good is if I had exit polling data by town in NH. If I added that variable to the equation, I would think that METHOD will lose significance.

UPDATE2: I got town population data, estimated for 2006. Using that instead of TOTAL VOTE reduces the size of the METHOD variable but not by much and it is still significant. I also calculated a new variable, total vote divided by population, which is an attempt to get at a few things related to turnout and other demographics. Using this new variable in addition to popluation reduces the size of the METHOD variable a bit more, but it is still highly significant. This new variable of vote divided by population seems to be important. I think if I got better demographic data -- age and gender, income, turnout -- the METHOD variable would lose more of its significance. It must be picking up something. If somebody has good NH town data in Excel or easily parsed, let me know the source.

The other thing to note is that in the data on votes by machine vs. hand, Romney also has big differences. That would, I think, support the "underlying demographics" theory.

UPDATE3: And just for the record, I do not believe whatsoever that anything untoward or suspicious happened in the NH election. This is simply a great exercise in statistical analysis, a great example of the problem that omitted variables cause in regression analysis.

19 comments:

Anonymous said...

So for all the conspiracy theorists out there, is this an attempt by one political camp to get a leg up on another by: a)diddling with the machine hoping no one will catch it so as to get more total votes, or b) diddling with the machine actually hoping someone WILL catch it so as to implicate their opposition?

catherine said...

I have been doing the same analysis with the same results, except I have far more demographic variables for each precinct, including education (% holding bA's and % holding high school diplomas), median income, median age, down to ridiculous specificity like whether they're on municipal water supply.

In your analysis try including obama's votes as a predictor of clinton's votes, and then look at the interaction term with the diebold variable. You may also want to look at quadratic effects of obama's votes, and the interaction of that with the diebold var. The logic: if someone wanted to make sure hillary won, they might target machines in those districts where obama was stronger than average (or they may have been somewhat more clever, which is what my analysis of the quadratic interaction term suggests).

Email me for the data file chris_chatham~~~AT~~~hotmail.com (I'm using my wife's blogger acct).

Unknown said...

"The other thing to note is that in the data on votes by machine vs. hand, Romney also has big differences. That would, I think, support the "underlying demographics" theory.

Given the amount of work by Election Defense Alliance and others showing significant statistical anomalies in elections since the time computers became widely involved with election tallying, won't you consider that something may have gone awry in the New Hampshire Republican race as well?

In another post you say:

Of course, since there will be a recount, any claims of irregularities will be quickly dispensed with.

The claims may be quickly dispensed with since that is in the best interest of NH officials and politicians in general, but an uneventful recount doesn't preclude foul play. It's not impossible for government to be behind election fraud, nor for recounts to be tampered with, especially where there is a weak chain of ballot custody.

Anonymous said...

Occam's razor would suggest that there is no demographic explantion for it, that the machines were rigged. Especially when you look at the pre-election polls that heavily favored Obama, the previous corruption involving machines produced by the same company, the shoddy procedures used by officials of the company that counted the votes, the convicted felons who were helping count the votes and on and on. Unfortunately, with no transparent chain of custody for the paper ballots, no anomolies doesn't prove no fraud.

Anonymous said...

I've been making this point since the primary results came in: even if the Diebold results are verified in the hand-recount, New Hampshire's lax same-day registration requirements (see http://www.democracyfornewhampshire.com/node/view/3016) make it possible for out-of-staters to drive into New Hampshire, vote, and drive back to their home state. Yes, identification is requested at the polls, but it is not required to submit a vote. Votes submitted without an ID are followed up on weeks after the polls close (if ever), but because the results are reported the same day to CNN, ABC, and all the rest, the momentum of the campaigns has already been affected and the damage to the election has been done.

It is entirely possible that the New Hampshire primary was decided by over-zealous out-of-staters who decided to give their chosen candidate a little bump. If you examine the disctricts that defied the poll expectations, they were all in the southern part of the state, right along I-93. Easy in, easy out.

The legality of these votes that the electronic scanning machines are counting (or not counting) need to be verified before we simply recount all the possibly illegal ballots cast.

Anonymous said...

Hi, I'd like to do some analysis myself, could you send me your dataset? I'd like to run a MANOVA to see if their is no interaction effect, it is indeed possible that every single factor has insignificant effects, but that their is a significant interaction effect...

Anonymous said...

Here's what I'd like to see. I'd like to see the election commission take an actual voting machine, load it with a copy of the actual software used in the NH primary, and feed a thousand or so sample ballots through it and see if they are counted accurately.

Anonymous said...

Anonymous said:
"So for all the conspiracy theorists out there, is this an attempt by one political camp to get a leg up on another by: a)diddling with the machine hoping no one will catch it so as to get more total votes, or b) diddling with the machine actually hoping someone WILL catch it so as to implicate their opposition?"

My understanding (and please correct me if I am wrong) is that it has been demonstrated that the Diebold products can be "diddled with" without leaving a trace. Imagine for a moment a valid, public chain of custody involving physical, paper ballots in a box with a physical seal (imagine that the seal is red sealing wax). In such an imagined circumstance, a scientific, physical examination of the seal would indicate whether the seal had been tampered with; because of the public chain of custody it probably would not be tampered with. But the point is you could tell.
But the point is that with the Diebold machines you can diddle and there is no way to tell whether you diddled or not.
So it is important to recall then that there are two things in question: First, accuracy. Of course, we want to know whether the publically released tallies match the actual votes cast. But don't forge that we also have to consider the question of validity. If it can be shown that the electronic record can be changed without being detected and/or that the chain of custody has holes in it that you can drive a truck through, then the election is not valid because there is no way to demonstrate that the outcome truly reflects the ballots cast.
Of course, these two concerns interact: We worry about ballot integrity and the chain of custody precisely because if they are lacking, someone has an opportunity to diddle.
But, in theory, you could have a perfectly (and I do mean perfectly) accurate tally but an invalid election because the integrity of the ballots and the chain of custody are fatally compromised: "Candidate X did win. She (or he) got the most votes. But we have no way of showing you how we know this. Just trust us."

Anonymous said...

ugg bootsTHE SNOW STARTEDwholesale ugg boots TO FALLWholesale handbags SEVERAL HOURSwholesale clothing BEFORE HER Wholesale jewelrylabor began.wholesale clothing A few flakes first,wholesale handbags in the dullwholesale furniture gray late-afternoon sky, Furniture Wholesaleand then wind-drivenWholesale jewelry swirls andCeramic tile eddies around Micro sd cardsthe edges of Wholesale clothingtheir wide front porch. Wholesale JewelryHe stood by her Wholesale fashion jewelryside at the window, Wholesale costume jewelrywatching sharpugg boots gusts of snow billow,ugg boots then swirl andugg boots drift to the groundugg boots. All around the neighborhoodwholesale, lights came onwholesale electronics, and the naked branches of the trees turned white.

Anonymous said...

china wholesaleTHE SNOW STARTED wholesale electronicsTO FALL SEVERAL HOURSwholesale BEFORE HER labor beganwholesale. A few flakes first,china wholesale in the dull graywholesale digital cameras late-afternoon skywholesale mp3, and then wind-drivenwholesale mp3 swirls and eddies around mp4 watchesthe edges of their mp5 playerwide front porch. digital camera wholesaleHe stood by her sidewholesale mp3 players at the window, wholesale mp3 playerwatching sharp gusts of snow billow, china electronics wholesalethen swirl and drift to the ground.

Anonymous said...

「TOEFLテスト英単語3800について若干書きます。
toefl 
不動産投資
個別指導塾
幼児教室
国際協力
合宿 免許
名刺
フロアコーティング

Anonymous said...

hupoint
不動産
合宿免許
物語の世界
描く日記
愛車

Anonymous said...

gucl
shunk名刺
cona
netbol探偵
saish
town
yeawo
over
All of
raid
again
Frederick
sodas
will be set free
keeping

熱い夏
奇跡の日
桜の蕾
Radley

Unknown said...

SEO対策(けんさくエンジンさいてきか)とは、ある特定の検索エンジンを対象として検索結果でより上位に現れるようにウェブページを書き換えること。またそ の技術。サーチエンジン最適化とも言われる。英語の "Search Engine Optimization" の頭文字を取ってSEOとも言われる。最適化の対象になる検索エンジンは、Googleであることが多い。これは、海外(特にアメリカ)において Googleのシェアが高いことによる。日本ではYahoo!サーチの利用者が多いため、Yahoo!サーチ対策も重視されている。

Unknown said...

ドルチェ&ガッバーナ
DOLCE & GABBANA
ドルチェ&ガッバーナ バッグ
ドルチェ&ガッバーナ 財布
ドルチェ&ガッバーナ ネックレス
ドルチェ&ガッバーナ サングラス
ドルチェ&ガッバーナ リング
ドルチェ&ガッバーナ 香水
ドルチェ&ガッバーナ シューズ

グッチ
gucci
グッチ バッグ
グッチ 財布
グッチ ネックレス
グッチ サングラス
グッチ リング
グッチ 香水
グッチ シューズ

ディオール
Christian Dior
ディオール バッグ
ディオール 財布
ディオール ネックレス
ディオール サングラス
ディオール リング
ディオール 香水
ディオール シューズ

ヴィトン
LOUIS VUITTON
ヴィトン バッグ
ヴィトン 財布
ヴィトン ネックレス
ヴィトン サングラス
ヴィトン リング
ヴィトン 香水
ヴィトン シューズ

シャネル
CHANEL
シャネル バッグ
シャネル 財布
シャネル ネックレス
シャネル サングラス
シャネル リング
シャネル 香水
シャネル シューズ

プラダ
prada
プラダ バッグ
プラダ 財布
プラダ ネックレス
プラダ サングラス
プラダ リング
プラダ 香水
プラダ シューズ

フェラガモ
SALVATORE FERRAGAMO
フェラガモ バッグ
フェラガモ 財布
フェラガモ ネックレス
フェラガモ サングラス
フェラガモ リング
フェラガモ 香水
フェラガモ シューズ

セリーヌ
CELINE
セリーヌ バッグ
セリーヌ 財布
セリーヌ ネックレス
セリーヌ サングラス
セリーヌ リング
セリーヌ 香水
セリーヌ シューズ

ボッテガ
BOTTEGA VENETA
ボッテガ バッグ
ボッテガ 財布
ボッテガ ネックレス
ボッテガ サングラス
ボッテガ リング
ボッテガ 香水
ボッテガ シューズ

コーチ
COACH
コーチ バッグ
コーチ 財布
コーチ ネックレス
コーチ サングラス
コーチ リング
コーチ 香水
コーチ シューズ

ダンヒル
dunhill
ダンヒル バッグ
ダンヒル 財布
ダンヒル ネックレス
ダンヒル サングラス
ダンヒル リング
ダンヒル 香水
ダンヒル シューズ

ロエベ
LOEWE
ロエベ バッグ
ロエベ 財布
ロエベ ネックレス
ロエベ サングラス
ロエベ リング
ロエベ 香水
ロエベ シューズ

ディーゼル
DIESEL
ディーゼル バッグ
ディーゼル 財布
ディーゼル ネックレス
ディーゼル サングラス
ディーゼル リング
ディーゼル 香水
ディーゼル シューズ

大阪 デリヘル
仙台 デリヘル
仙台 デリヘル
仙台 デリヘル
家族葬
滋賀 賃貸
アダルト SEO
被リンク
茶道具 買取
絵画 買取
レザー
革小物
クレジットカード 現金化
現金化
ショッピング枠 現金化
クレジットカード 現金化
現金化
ショッピング枠 現金化
FX
FX 比較
FX 初心者

キャッシング
大阪 賃貸
中古車 販売
ルームウェア
大阪 マンション
賃貸マンション 神戸
中古 ゴルフクラブ
クールビズ
フィットネスクラブ
大阪府 司法書士
クレジット 申し込み
ベビードール
矯正歯科 東京
ホワイトニング 東京
大阪 ラブホテル
リサイクルショップ
不動産
カードローン
投資 信託
下着
即日 キャッシング
三井住友銀行
神戸市 中央区 税理士
FX
消費者金融
ローン
引越し
生命保険
ジェルネイル
人材派遣
ネット証券
アフィリエイト
格安航空券
ウィークリーマンション
レンタカー
SEO
オフィス家具
合宿免許
ペット用品
高速バス
デリヘル
キャバクラ
派遣
コラーゲン
化粧品
インテリア
ウェディング
結婚相談
投資物件
留学
貸事務所 大阪
経営コンサルティング
工芸品
高級品
自動車保険
ホテヘル
レストランウェディング
バイク買取
運転免許
ベビーカー
外反母趾
圧力鍋
腕時計
フェラガモ
デリヘル
キャバクラ
セレブ
プラセンタ
カルシウム
青汁
ブルーベリー
家具
脱毛クリーム
除毛クリーム
コスト削減 大阪
弁護士 大阪
車買取 大阪
バイク買取 大阪
エステ 大阪
リフォーム 大阪
大阪 歯科
派遣 大阪
アルバイト 大阪
転職 大阪
大阪 住宅
大阪 専門学校
グルメ 大阪
ホテル 大阪
一戸建て 大阪
大阪 宿泊
大阪 マンション
デリヘル 大阪
印刷 大阪
不動産 大阪
賃貸 大阪
ブライダル 大阪
リサイクル
アダルト SEO
賃貸
SEO 大阪
イベント コンパニオン 大阪
転職 大阪
大阪 ラブホ
ペット ショップ 大阪
豆腐
京都 不動産
運転免許 合宿
ヘアアイロン
ダイエット
ダイエット
デリヘル
キャバクラ
シャンパン
老人ホーム
精力剤
大阪 ラブホテル
ブランド品 買取
ワイン
京都 不動産
ペット
リサイクルショップ
歯科求人
結婚式場
バイク便
動物病院
美容整形外科
エルメス
ダイエット
ダイエット食品
腕時計
ヘアアイロン
クレイツ
アイビル
アドスト
バッグ
ネイル
アクセンツ

Anonymous said...

UGG Boots,UGGs,Jordan Shoes,Basketball Shoes ,Air Jordan Shoes,Air Jordan,Nike Shoes,Jordan Shoes,Basketball Shoes ,Air Jordan Shoes,Air Jordan,Nike Shoes,Jordan Shoes,Air Jordan
Cheap WoW Gold,Buy WoW Gold,WoW Gold,WoW Power Leveling,WoW Leveling,Cheap WoW Gold,Buy WoW Gold,WoW Gold,WoW Power Leveling,WoW Leveling

Anonymous said...

搬家 搬家服務 搬家保障 搬家網 搬家估價 搬家 搬家公司 補習班 多益 在職進修 婚紗 新娘秘書 汽車旅館 彩妝造型 新娘秘書 票貼 室內設計 室內設計 外遇 抓姦 應收帳款 徵信 徵信社 外遇 徵信 徵信社 外遇 植牙 牙齒矯正 坐月子 宜蘭民宿 婚禮佈置 宜蘭民宿推薦 催眠 派報 太陽能熱水器 Shade sail nike shoes 關鍵字廣告 租屋 搬家 搬家 買房子 花蓮民宿 花蓮民宿 花店 租房子 xo醬 房屋貸款 搬家公司 減肥 減重 床墊 創業加盟 團體服 學英文 英文 補習班 勞工體檢 資源回收 生日禮物 團體服 團體制服 班服 塑膠 日立家電 飾品批發 电动隔膜泵 自吸泵 化工泵 离心泵 磁力泵 螺杆泵 水泵 隔膜泵 气动隔膜泵 MBA 在职研究生 在职博士 搬家 搬家服務 搬家保障 搬家網 搬家估價 徵信 徵信的意義 徵信服務 徵信報導 徵信問答 徵信知識 婚禮佈置 婚禮佈置 婚禮佈置 酒店經紀 酒店經紀 班服配件 團體服配件 團體服 班服 團體服 班服 團體服 室內設計公司 室內設計公司 室內設計公司

Anonymous said...

台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司台灣汽車旅館加盟網--
台北汽車旅館
台中汽車旅館
高雄汽車旅館牛初乳保洁公司

ADMIN said...

VIPQIUQIU99.COM AGEN JUDI DOMINO ONLINE TERPERCAYA DI INDONESIA