Using Polls in your prediction:Anyone can conduct a poll and they can warp their findings to whatever end result they wish to show. It's not all that hard. In addition, even the best polls only poll a small percentage of the people and that data is assumed to be a cross-section of the community, city, state or nation.
The bottom line is that the only opinion that matters in an election is the one taken on Election Day in the voting booths.... all else is simply someone's opinion.
So you have to be careful when using polls and you only take the polls as being a general trend or a loosely based fact of how the people of any given area think. The best polls, which I have found, are: Gallup, SurveyUSA, and Rasmussen Reports. I also found that RealPolitics has an excellent site for polling data. There is one other place which is good on Polls and that is:
SurveyUSA (McCain vs. Obama 03/06/08) www.surveyusa.com/ECV_50-State_results_030608_vs_Obama.htmlI use SurveyUSA in the predictions being done in this set of messages.
Legend used in the poll extracts:Note: If you wish, check the color code for accuracy... there may be a couple of entries colored differently than the Legend shows. Maroon = McCain leads both Obama and Clinton by 10 percent or more.
Red = McCain leads over one opponent by 10 percent or more, but leads over the other by 4 to 9 percent.
Orange = McCain leads over one Dem by 4% or more, but other candidate is with in 4%.
Black = Poll Results show all Candidates with in 4% of each other.
Green = McCain wins versus one Dem, but loses against the other. (Plus or minus 4% are considered ties).
Teal = One Dem leads McCain by over 4%, but other candidate is within 4% of McCain.
Blue = One Dem leads McCain by 10 percent or more, while other Dem is 4 to 9 percent over McCain
Navy = Obama and Clinton over McCain by 10 Percent or more.
The Poll Results:Alabama: McCain (54%), Obama (40%)/McCain (51%), Clinton (41%)
Alaska: McCain (48%), Obama (43%)/McCain (56%), Clinton (34%)
Arizona: McCain (51%), Obama (39%)/McCain (54%), Clinton (39%)
Arkansas: McCain (53%), Obama (33%)/Clinton (51%), McCain (40%)
California: Obama (51%), McCain (40%)/Clinton (50%, McCain (40%)
Colorado: Obama (50%), McCain (41%)/McCain (48%), Clinton (42%)
Connecticut: Obama (55%), McCain (34%)/Clinton (50%), McCain (40%)
Delaware: Obama (50%), McCain (41%)/Clinton (46%), McCain (41%)
Florida: McCain (47%), Obama (45%)/Clinton (51%), McCain (42%)
Georgia: McCain (54%), Obama (41%)/McCain (56%), Clinton (35%)
Hawaii: Obama (61%), McCain (31%)/Clinton (43%), McCain (39%)
Idaho: McCain (52%), Obama (39%)/McCain (63%), Clinton (27%)
Illinois: Obama (60%), McCain (31%)/Clinton (48%), McCain (37%)
Indiana: McCain (50%), Obama (41%)/McCain (53%), Clinton (36%)
Iowa: Obama (50%), McCain (41%)/McCain (46%), Clinton (41%)
Kansas: McCain (50%), Obama (41%)/McCain (51%), Clinton (42%)
Kentucky: McCain (54%), Obama (33%)/McCain (50%), Clinton (41%)
Louisiana: McCain (54%), Obama (41%)/McCain (51%), Clinton (41%)
Maine: Obama (53%), McCain (39%)/Clinton (48%), McCain (42%)
Maryland: Obama (53%), McCain (40%)/Clinton (49%), McCain (40%)
Massachusetts: Obama (49%), McCain (42%)/Clinton (55%), McCain (37%)
Michigan: Obama (46%), McCain (45%)/McCain (44%), Clinton (44%)
Minnesota: Obama (49%), McCain (42%)/Clinton (47%), McCain (43%)
Mississippi: McCain (54%), Obama (41%)/McCain (51%), Clinton (42%)
Missouri: McCain (48%), Obama (42%)/McCain (48%), Clinton (44%)
Montana: McCain (47%), Obama (39%)/McCain (53%), Clinton (33%)
Nebraska: McCain (45%), Obama (42%)/McCain (57%), Clinton (30%)
Nevada Obama (46%), McCain (41%)/McCain (49%), Clinton (41%)
New Hampshire: Obama (46%), McCain (44%)/McCain (49%), Clinton (41%)
New Jersey: McCain (43%), Obama (43%)/Clinton (47%), McCain (42%)
New Mexico: Obama (50%), McCain (43%)/McCain (47%), Clinton (47%)
New York: Obama (52%), McCain (38%)/Clinton (55%), McCain (33%)
North Carolina: McCain (47%), Obama (45%)/McCain (49%), Clinton (41%)
North Dakota: Obama (46%), McCain (42%)/McCain (52%), Clinton (35%)
Ohio: Obama (50%), McCain (40%)/Clinton (50%), McCain (40%)
Oklahoma: McCain (57%), Obama (34%)/McCain (50%), Clinton (42%)
Oregon: Obama (49%), McCain 41%/McCain (48%), Clinton (43%)
Pennsylvania: McCain (47%), Obama (42%)/Clinton (47%), McCain (46%)
Rhode Island: Obama (53%), McCain (38%)/Clinton (54%), McCain (37%)
South Carolina: McCain (48%), Obama (45%)/McCain (48%), Clinton (42%)
South Dakota: McCain (47%), Obama (43%)/McCain (52%), Clinton (40%)
Tennessee: McCain (54%), Obama (38%)/McCain (46%), Clinton (46%)
Texas: McCain (47%), Obama (46%)/McCain (49%), Clinton (42%)
Utah: McCain (50%), Obama (39%)/McCain (65%), Clinton (27%)
Vermont: Obama (63%), McCain (29%)/Clinton (49%), McCain (39%)
Virginia: McCain (47%), Obama (47%)/McCain (50%), Clinton (40%)
Washington: Obama (52%), McCain (38%)/McCain (48%), Clinton (44%)
West Virginia: McCain (53%), Obama (35%)/Clinton (47%), McCain (42%)
Wisconsin: Obama (51%), McCain (40%)/Clinton (48%), McCain (44%)
Wyoming: McCain (54%), Obama (35%)/McCain (61%), Hillary (28%)
EV from Poll Summary:
Obama: 236
McCain: 163
With-in 4%: 139
If you add North Dakota to Obama and South Dakota along with South Carolina to McCain because they are at 3% and 4% difference... then you get:
Obama: 239
McCain: 179
With-in 2%: 125
The Remaining states are: Florida, Michigan, New Hampshire, New Jersey, North Carolina, Texas and Virginia.
McCain: 230
Clinton: 212
With-in 4%: 96
If you add Minnesota and Wisconsin to Hillary, and Missouri and Washington to McCain.. the new totals are:
McCain: 234
Clinton: 250
With-in 1%: 54
The remaining states are Tennessee, Pennsylvania, New Mexico, and Michigan. According to the poll these are all
Comparing the Base Prediction with Polls:The first step is to always check your work for accuracy, make sure you transpose your numbers and information correctly and then double check any math you do. For the purposes of this post, I didn't check everything as I would normally do... and I know there has to be mistakes. Take this information as a guideline for your own prediction and not as 100% fact. If you use what I have here, double check my information for accuracy. Don't be shy about pointing out mistakes for my main concern was in the points being made and not so much in the data itself.
The second step is to find a chart showing the various degrees of votes. There are sites which have a breakdown, but I like things simple -yet not too simple. My rankings go something like this:
Ranking | Poll % Difference |
Strong Dem | over 10% |
Moderate Dem | 5% - 9% |
Weak Dem | 1% - 4% |
Tied | 0 |
Weak GOP | 1% - 4% |
Moderate GOP | 5% - 9% |
Strong GOP | over 10% |
For the above chart I will see how much of a difference there is in the polls between the two opponents. In a case like the Dems have with Hillary and Obama, I will simply take the two differences they have with McCain and add them together.. then divide by two. I will then use that number as my determining factor.
For example, let's say a state says McCain would beat Hillary by 10%, but beat Obama by 8%. The difference would be 9%... That would make that state a Moderate Republican state. On the other hand if Obama is predicted to defeat McCain my 10%, but McCain is expected to defeat Hillary by 6%, I give Hillary a minus 6 and add that to Obama's number... making the spread 16%.. I then divide by 2 and get 8%.. so that would make the state a Moderate Democratic State. When doing this, please remember that when a candidate drops out, you must correct your figures.
For the historical data I like to go back to 1960, about 12 Elections. I then talley up how many times that state voted Republican or Democrat. I will add in parties that are similar to the Republican or Democratic parties to the closest party.
Ranking | Historical Data |
Strong Dem | 11 -12 |
Moderate Dem | 9 - 10 |
Weak Dem | 7 -8 |
Moderate | 6 -6 |
Weak GOP | 7 - 8 |
Moderate GOP | 9 - 10 |
Strong GOP | 11 -12 |
The third thing to do is remember to not reinvent the wheel everytime you want to revise your information. I personally think that once you set up your information then there is really no need to revise it until around September or November when it gets to be one on one... and closer to the election. At that stage you need to recheck the polls and adjust as necessary. Don't be so concerned over "exact figures", by this I mean if Barack was expected to win in April, by 5 or 10% and in September or October he is expected to win by 7 or 14%... then don't worry about it - a win is a win. He won't get any more EV than he would have before. Only be concerned if the candidate falls under the 5% or 3% margin of error you set for your prediction.
There is a tip to use here, don't make your prediction and forget about it. As you see fit and have time, make revisions as necessary. Let's say Hillary is expected to win right now in Pennsylvania by 1%. Now let's say that the figures show that only 84% of the people polled were decided voters. When May or June rolls around and you see the polls for Pennsylvania say that Hillary has a 10% or 15% lead over McCain and 94% of those polled are decided... then revise your data for Pennsylvania. In September you might see that Hillary only has a 8% Margin, but she is still well ahead of the opposition and you might not have to bother changing your data again on that particular race. This idea might save you some time in revising your prediction closer to the election.