Official News Discussion

[CFHollister]

Seeing as the Official News Thread is closed (as it should be), I though it might be nice to have a thread for discussing the official news. Particularly since I had a question to throw out there and also thought this situation my come up again in future.

Official News item:
"-Alizée was mentioned yesterday on the TF1 show, "Attention à la marche!". The question was: "Out of 100 teens, how many think that Alizée is still trendy?" Answer: 40%."

Question:
Do you think we should view this 40% figure as: 1) a positive sign that Alizée has mainted this level of popularity despite her long break, 2) pessimistically that it is only 40%, or 3) do you think the small sample (which may not be representantive) doesn't really give us enough useful information to make a judgement?

[RMJ]

Well, I don't know the error margin but 40% sounds a lot.

40% of all teens would be millions of people.

And 40% is a lot if you consider how much she's been on public lately.

[RadioactiveMan]

40% is very considerable. It's not like it would have been 100% even when she was popular. I was impressed when I saw the figure.

[SupaKrupa]

I think it's a positive sign that people think something more than 4 years ago is still trendy

But at the same time, the sample isn't big enough. For all we know the 40% were gay, or mostly girls etc.

[Drake498]

yeah, the survery is way too small if they just asked 100 teens, and =like Kupra said, we don't know the specifics either

[Zack -Alizee Lover-]

Originally Posted by Drake498

yeah, the survery is way too small if they just asked 100 teens, and =like Kupra said, we don't know the specifics either

You dont need hundreds of people to do that kind of surveys... 100 teens is just enough.

[Drake498]

Originally Posted by Zack -Alizee Lover-

You dont need hundreds of people to do that kind of surveys... 100 teens is just enough.

100 is a really small survey, idk how you consider only 100 participants to give any conclusive results, maybe general, but it would be too generalized for anything specific or more accurate

[brad]

you also have to keep in mind that if they only asked teenagers today, a lot of the people that were really fans of Alizée aren't teenagers anymore. I mean Gourmandises came out 7 years ago ..

That said, getting 40% of people to agree on anything is still very respectable.

[theit77]

For those of you who don't want to read the following statistical jargon, here's the jist: Somewhere between 30% and 50% of French teenagers still consider Alizée trendy.




"Alizée was mentioned yesterday on the TF1 show, "Attention à la marche!". The question was: "Out of 100 teens, how many think that Alizée is still trendy?" Answer: 40%.""


The first thing to confirm is if the show actually performed a survey. If they actually did question 100 french teenagers, then we can continue with the analysis. If the number is just an estimate... well, then it's just the show's estimate.

The population that the statistic is measuring is all teenagers in France. To confidently analyze the data, the survey must have been an SRS or Simple Random Sample and the population size must have been at least 10 times the sample size. The number of French teenagers is certainly greater than 100(10) or 1000. But an SRS is less certain. To get a Simple Random Sample, the survey would have had to randomly survey 100 teenagers out the entire population of French teenagers. This is essentially impossible to do without getting all teenagers' social security numbers, randomly choosing 100, and then forcing those 100 to answer the question. So, without a certain SRS, we can use a confidence interval, but we cannot be certain of its accuracy.

Calculation of a confidence interval: (p-hat)±z*(((p-hat)(1-(p-hat))/n)^(1/2) Where (p-hat) is the statistic of 0.4, n is 100, and z* is determined by the desired confidence interval: InvNorm((1-.95)/2)

Crunching the numbers, we find there is a 95% confidence interval of (0.30398, 0.49602). This means that we are 95% confident that the true percentage of French teenagers who still consider Alizée trendy is somewhere between 30.398% and 49.602%. A 90% confidence interval would yeild (0.31942, 0.48058), and a 99% confidence interval would yeild (0.27381, 0.52619).

[CFHollister]

Originally Posted by theit77

For those of you who don't want to read the following statistical jargon, here's the jist: Somewhere between 30% and 50% of French teenagers still consider Alizée trendy.

"Alizée was mentioned yesterday on the TF1 show, "Attention à la marche!". The question was: "Out of 100 teens, how many think that Alizée is still trendy?" Answer: 40%.""

The first thing to confirm is if the show actually performed a survey. If they actually did question 100 french teenagers, then we can continue with the analysis. If the number is just an estimate... well, then it's just the show's estimate.

The population that the statistic is measuring is all teenagers in France. To confidently analyze the data, the survey must have been an SRS or Simple Random Sample and the population size must have been at least 10 times the sample size. The number of French teenagers is certainly greater than 100(10) or 1000. But an SRS is less certain. To get a Simple Random Sample, the survey would have had to randomly survey 100 teenagers out the entire population of French teenagers. This is essentially impossible to do without getting all teenagers' social security numbers, randomly choosing 100, and then forcing those 100 to answer the question. So, without a certain SRS, we can use a confidence interval, but we cannot be certain of its accuracy.

Calculation of a confidence interval: (p-hat)±z*(((p-hat)(1-(p-hat))/n)^(1/2) Where (p-hat) is the statistic of 0.4, n is 100, and z* is determined by the desired confidence interval: InvNorm((1-.95)/2)

Crunching the numbers, we find there is a 95% confidence interval of (0.30398, 0.49602). This means that we are 95% confident that the true percentage of French teenagers who still consider Alizée trendy is somewhere between 30.398% and 49.602%. A 90% confidence interval would yeild (0.31942, 0.48058), and a 99% confidence interval would yeild (0.27381, 0.52619).

I really respect people that can understand statistics. It's so important yet its just one of those things I can't get to stick with me. I could remember what I needed long enough for the test.

As to answering my own question, I think the significance of the figure depends a lot on the sample and the sampling methods used. If they had any semblance of decent random sampling, then I think 40% after such a break is pretty darn good. If not, then it doesn't really tell us anything useful. As has been already mentioned, the significance of the result also depends on what is meant by "trendy" and how the respondants interpretted the term/question.

PS-Thanks to whoever stickied the thread. The actual news thread is rightfully locked so that it doesn't become a place where anyone's rumors are thrown around and the same questions get asked over and over. I would just ask for everyone to not allow this thread to turn into that sort of thing either