Only two of the seven pollsters that published measurements between the first and second rounds anticipated the arrival of Gustavo Petro to the presidency. They all underestimated his vote and the Centro Nacional de Consultoría and Yanhaas, who guessed the winner, were the most successful and the one that farthest from the results.
With this evaluation, the La Silla Vacía electoral poll semaphore is updated, this time it does not take into account the margins of error, so as not to favor polls with smaller samples and less ambition to get it right, and removes the undecided from the figures.
Here they are, organized from the one that came closest to the results to the one that came the least.
Second round results | Gustavo Petro | Rodolfo Hernández | Blank vote | |
| 50,4% | 47,3% | 2,4% | |
| | | | |
Pollsters | Gustavo Petro | Rodolfo Hernández | Blank vote | Total error |
CNC | -0,6 | -1,8 | 3,3 | 5,7 |
Atlas Intel | -2,9 | 2,9 | 0,0 | 5,8 |
Invamer | -3,2 | 0,9 | 2,3 | 6,4 |
GAD3 | -3,3 | 0,6 | 2,6 | 6,5 |
Guarumo | -3,9 | 0,9 | 2,9 | 7,7 |
Massive Caller | -5,8 | 8,1 | N/A | 13,9 |
Yanhaas | -2,3 | -9,9 | 11,5 | 23,7 |
Numbers in green mean that the pollster overestimated the vote for the candidate; in red that she underestimated her.
THE POLLS IN THE SECOND ROUND OF 2022
Centro Nacional de Consultoría:
His last poll was ten days before the election, not the Saturday before the week of the closure like the other pollsters. However, it was the poll that was closest to the results, once the undecided were removed from the base. Without the 2.5% margin of error (within which all of his results fell), his biggest fault was having underestimated Rodolfo Hernández's results by less than 2 percentage points. He not only predicted Petro's victory but only underestimated him by 0.6%.
Atlas Intel:
He hit exactly the 2.4% who got the blank vote but his last measurement gave Hernández the victory with 50.2 percent, 2.9 points above what he got, and underestimated Petro by the same 2.9%. It had the smallest margin of error, a single percentage point.
Invamer:
He underestimated Petro by 3.2% (the only overestimation above the margin of error of 2.7 percentage points), he overestimated the blank vote by 2.3% and Hernández by less than 1%.
GAD3:
He underestimated Petro's result by 3.3 percentage points and estimated that Rodolfo Hernández would win with 47.9% of the vote on Sunday, just 0.6% behind his result. His latest projection of the blank vote (of 5%) was 2.6 percentage points higher than the results.
Guarumo:
He underestimated Gustavo Petro by almost four points and was the pollster that put him the lowest, with 46.5% of the vote. Her last measurement put the blank vote at 5%, almost three points above the actual result. He overestimated Hernández by less than a percentage point.
Massive Caller:
He had a big disappointment with Hernández, giving him 55.4%, when he got 8.1 points less. He projected Petro with 5.8% less of his result. He had the highest margin of error: 3.4%.
Yanhaas:
The poll furthest from the results, beat the winner even though it underestimated him by 2.3%. He projected Hernández with 37.5%, almost ten percentage points lower than his vote, and he missed a lot in the blank vote, forecasting it at 14% when it was 2.4%.
Article taken from lasillavacia.com
Comments