Pro Touch Hair Straighteners

surveys: kicking with alternate numbers

Cristián Bofill, director of The Third , feared, hated and respected. Mateo has a reputation for being a workaholic boss and demanding: they have hundreds of stories from arbitrary dismissal, for long hours in the drafting of the paper, countless challenges to his subordinates for lack of thoroughness or not having any of the information. But all its rigor has not served absolutely nothing when it comes to publishing statistics reliably, it would be time to start taking the matter seriously.

.

statistical errors The Third are repeated , which is more serious considering for some time that the newspaper decided to build its own center of surveys. What helps them produce their own data if they are using them so badly? Let only a couple of mistakes.

.

2 + 2 = 5 (for extremely high values \u200b\u200bof 2 ...).

.
The last survey was published in two shifts: Saturday and Sunday. On Saturday, he said on the first page that " Alvear is installed as the best presidential rival Piñera." This is the chart that accompanied the note:

The tie between Soledad Alvear and Piñera by 45% is evident. However, to presume from these data that actually Piñera beats Lagos and Insulza a little more complicated. The first thing to keep in mind (which is obvious, but the paper seems to consider minor) is that survey data are always estimates and the actual value of each number moves between certain margins : this is a confidence interval and corresponds to the value of the estimator + / - the value of the estimation error.

Another way to see if Lagos wins Piñera e Insulza is a difference of proportions test for independent samples . According to the calculations of this blogger, the statistic test (z) of the comparison between Piñera and Lagos da and between 1.4845 and Insulza Piñera da 1.5011, considering a confidence level of 95% and a sample size of 800 cases. Both test statistics are within the acceptance region: no history to say that one (Piñera) is greater than the other (Lagos or Insulza). Piñera That is not only tied with Soledad Alvear. Lagos also tied with Insulza.

The editorial staff of interpreting the data prefer a rather more free, "however, the former Minister Alvear lost ground in their dispute with Piñera. In July two-Insulza and the helmsman of the DC-got a draw against Piñera (...). Now, although the former minister close the gap with Piñera (42% vs. 47%) is second Alvear . The text continues by stating that "despite having suffered their pole position in the presidential race RN the flagship is not surpassed by any of the figures in the Coalition: Alvear and tied with Lagos and Insulza exceeds five points .

In the fall of the same article says that "the standard bearer RN barrier does not exceed 50% compared to the top 3 cards of the ruling party ": reality, however, is that support for Piñera among the population, according to the calculation of confidence interval should be between 46.6% and 53.4%, ie may well be over 50% .

Soledad Alvear, the best ever.

On Sunday published the results of the evaluation of public figures. The title page 8 reads: "Alvear leading political ranking and outperforms the PS presidential." Also presented the following chart:

thumbs shows the variation of the estimator since the previous survey, carried in July. Whether the difference is significant would require further calculation we will not do here. I'll settle for actually show what is said in the headline is not true : an estimate of confidence intervals shows that the value for the people of the positive evaluation of Soledad Alvear could be among 65.6% and 72.4% . The Insulza could be among 63.6% and 70.4% and Lagos between 56.6% and 63.4% . Intervals as Soledad Alvear and Insulza overlap, you can not say that the president of the DC is heading a ranking: there is actually a statistical tie with Insulza.

Knowledge and news.

There is nothing problematic (at least from my perspective, a journalist may think otherwise) use surveys as a tool of construction news, provided they are used well. The problem surely arises when the imperative for 'hit' or have an interesting story conflicts with what the data say: understand that there is nothing more boring and exciting than a headline saying "TIE! ". I imagine that the news judgments requires that there be a winner, even at the expense of discrediting little a little to their polling center.

anyone else inclined to conspiracy theories that I could draw the following conclusion: The Third is right and is interested in building a climate of opinion unfavorable to the possible pre-socialist candidates . This is difficult to prove. What we do know, thanks to Elisabeth Noelle-Neumann is that the climate of opinion is very important for how individuals decide to vote. Anyway, I would not go so far as to say that the suspicious are wrong: on the contrary, it is the responsibility of the newspaper ensure public confidence in writing and keep a good image . The burden of proof lies with the newspaper, not the public. If you are served to the public to distrust and arguments presented in large print and color cartoons, the fault of La Tercera.

Recommended Reading:
Public Opinion Does Not Exist of Pierre Bourdieu.
reporting How survey of Eduardo Arriagada .

Sources: the graphics were removed from the editions of Saturday 1 and Sunday September 2, La Tercera. Extracts of the texts also originate there.

.