American Research Group, Inc.
The National Economy
Electoral Vote Calculator
The New Hampshire Poll
Effective Advertising
Scintillating Grid
2014 Christmas Spending
2014 NFL Predictions
Survey Tools -
Sample Size Calculator
Margin of Error Calculator
Ballot Lead Calculator
Subscription Information


March 23, 2015

Obama Job Approvals and the National Economy


Obama Job Approval Ratings
03/20/15 Approve Disapprove Undecided
Overall 47% 49% 4%
Economy 49% 47% 4%

A total of 47% of Americans say they approve of the way Barack Obama is handling his job as president and 49% say they disapprove of the way Obama is handling his job. In February, 46% of Americans approved of the job Obama was doing and 49% disapproved.

When it comes to Obama's handling of the economy, 49% of Americans approve and 47% disapprove. In February, 45% approved and 50% disapproved.

This is the first time since September 2009 that Obama's job approval for handling the economy has been positive.

Among Americans registered to vote, 46% approve of the way Obama is handling his job as president and 50% disapprove. On Obama's handling of the economy, 49% of registered voters approve and 47% disapprove.

Details from the nationwide survey conducted March 17-20 are available at The National Economy.



March 19, 2015

2015 NCAA Men's Division I Basketball Tournament
Prediction Accuracy


Nate Silver's is back with predictions for the 2015 NCAA Men's Division I basketball tournament (see the accuracy of Nate's model for the 2014 NFL season here). Nate's model provides the odds of each team advancing in the tournament and, unlike most bracket competitions, is updated after each game.

The table below uses Brier scores to determine the accuracy of the probabilistic predictions in Nate's model. Brier scores range from 0.0 to 1.0, with 0.0 meaning the probabilities in the model perfectly match the outcomes of the games and 1.0 meaning no matches were made (so the closer to 0.0, the better the model). If a 50% probability were assigned to each team for each game, the Brier score would be 0.25.

Just as last year, Nate's model failed in the first round of the tournament.

The Brier score for the four games of 0.2640 exceeds what the score would have been by assigning each team a 50% probability of winning.

2015 NCAA Men's Division I
Round 1 First Four
Winner 538 Probability
of Win
Dayton 0.73 0.0705
Mississippi 0.56 0.1916
Robert Morris 0.49 0.2651
Hampton 0.27 0.5289
Brier score for 4 games 0.2640

Although the model failed in the first round of games, the format of the second round, where higher-ranked teams play lower-ranked teams, should lead to an improved Brier score as it did last year. The model, however, will continue to miss the upsets, which, obviously, are important in a single-elimination tournament.

As the results from the first round demonstrate, Nate's model does not solve the problem of the unrepeatable single case (100,000 game simulations do not constitute the infinite long run).

Accuracy results for each round will be posted here.



February 17, 2015

2014 NFL Prediction Accuracy - Worse Than Guessing

Updated results, including reliability tables, assessing the accuracy of Microsoft's Cortana and Nate Silver's Elo predictions for the 2014 NFL season are available here.

Both models performed worse than guessing in the second week of the season, invalidating the models.

While the Elo model finished the season with a slightly better Brier score, the Cortana model's favorites won more games and the Cortana model was less prone than the Elo model to over and under-forecasting and therefore less prone to bias.



Why doesn't most advertising work? Take a look at 10 Rules for More Effective Advertising.


American Research Group, Inc.  814 Elm Street  Manchester, NH 03101  603/624-4081
© 2015 American Research Group, Inc. All rights reserved. Reproduction in whole or in part without permission is prohibited.