|
|
Hong Kong9154 Posts
On April 26 2014 10:30 m4ini wrote:Show nested quote +On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is.
Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No.
It samples a slice of people that are not representative of the population.
I'll not even get into asking how they even formed the list of numbers to call.
https://en.wikipedia.org/wiki/Selection_bias https://en.wikipedia.org/wiki/Accidental_sampling
|
On April 26 2014 10:34 itsjustatank wrote:Show nested quote +On April 26 2014 10:30 m4ini wrote:On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is. Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No.
You missed the in-person interviews. You don't need a phone there. Not to mention, if random person A doesn't receive the call, random person B does. It doesn't make it less random.
It samples a slice of people that are not representative of the population.
Funny, the last time i checked, these methods are widely used for election opinion surveys, for example.
Ah, there you're coming from. Well that's easy.
From here.
Obviously, if you call certain groups like family members and certain political groups where you can predict the outcome, randomness goes down the drain. Unless you can prove that in this case though, it's as random as it gets. It's pretty much good enough for every single survey.
Edit: no, you're just plain wrong, they even state that they use RDD. That's as random as it gets. First five paragraphs explain how they worked. I can't make sense out of all of it since i have to google translate, but the RDD thing is extremely obvious.
|
Hong Kong9154 Posts
On April 26 2014 10:44 m4ini wrote:Show nested quote +On April 26 2014 10:34 itsjustatank wrote:On April 26 2014 10:30 m4ini wrote:On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is. Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No. You missed the in-person interviews. You don't need a phone there. Not to mention, if random person A doesn't receive the call, random person B does. It doesn't make it less random. Funny, the last time i checked, these methods are widely used for election opinion surveys, for example. Ah, there you're coming from. Well that's easy. From here.Obviously, if you call certain groups like family members and certain political groups where you can predict the outcome, randomness goes down the drain. Unless you can prove that in this case though, it's as random as it gets. It's pretty much good enough for every single survey.
You still miss the point. Phone samples, no matter how they are conducted, only sample the portion of the population with phones, who are able to answer phones at the time you call, and who care strongly enough either way about the issue to want to respond (or you are otherwise coercing them to respond) . They are all absolutely worthless.
The methodology section of phone surveys can be translated as follows: "This is bullshit. Our methodology is bullshit. Some of you know it is bullshit, but our intended audience won't even read this explanation that the survey is bullshit. Even if they did read that it was bullshit, they wouldn't begin to understand why it was bullshit. They wouldn't care it was bullshit anyways because it is conducted by people they either already agree or disagree with."
|
On April 26 2014 10:52 itsjustatank wrote:Show nested quote +On April 26 2014 10:44 m4ini wrote:On April 26 2014 10:34 itsjustatank wrote:On April 26 2014 10:30 m4ini wrote:On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is. Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No. You missed the in-person interviews. You don't need a phone there. Not to mention, if random person A doesn't receive the call, random person B does. It doesn't make it less random. It samples a slice of people that are not representative of the population.
Funny, the last time i checked, these methods are widely used for election opinion surveys, for example. Ah, there you're coming from. Well that's easy. From here.Obviously, if you call certain groups like family members and certain political groups where you can predict the outcome, randomness goes down the drain. Unless you can prove that in this case though, it's as random as it gets. It's pretty much good enough for every single survey. You still miss the point. Phone samples, no matter how they are conducted, only sample the portion of the population with phones, who are able to answer phones at the time you call, and who care strongly enough either way about the issue to want to respond (or you are otherwise coercing them to respond) . They are all absolutely worthless.
That might be your opinion, even if you're wrong there (you can't always ask every single person in a country). That does not make it less random. It might make it not representative enough for you, which is totally fine - but not "not random". Especially since they not only did RDD, but also face-to-face surveying. If a person doesn't care, he has no opinion on that matter and completely falls out of the picture.
|
Hong Kong9154 Posts
On April 26 2014 10:57 m4ini wrote:Show nested quote +On April 26 2014 10:52 itsjustatank wrote:On April 26 2014 10:44 m4ini wrote:On April 26 2014 10:34 itsjustatank wrote:On April 26 2014 10:30 m4ini wrote:On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is. Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No. You missed the in-person interviews. You don't need a phone there. Not to mention, if random person A doesn't receive the call, random person B does. It doesn't make it less random. It samples a slice of people that are not representative of the population.
Funny, the last time i checked, these methods are widely used for election opinion surveys, for example. Ah, there you're coming from. Well that's easy. From here.Obviously, if you call certain groups like family members and certain political groups where you can predict the outcome, randomness goes down the drain. Unless you can prove that in this case though, it's as random as it gets. It's pretty much good enough for every single survey. You still miss the point. Phone samples, no matter how they are conducted, only sample the portion of the population with phones, who are able to answer phones at the time you call, and who care strongly enough either way about the issue to want to respond (or you are otherwise coercing them to respond) . They are all absolutely worthless. That might be your opinion, even if you're wrong there (you can't always ask every single person in a country). That does not make it less random. It might make it not representative enough for you, which is totally fine - but not "not random". Especially since they not only did RDD, but also face-to-face surveying. If a person doesn't care, he has no opinion on that matter and completely falls out of the picture.
You can ask everyone in a country. That is what a census is. You can also develop proper stratified random samples using methods that are not phone sampling.
Just because it has "random" in it doesn't mean it is statistically random, a state that allows you to make inferences about the true population.
|
On April 26 2014 11:00 itsjustatank wrote:Show nested quote +On April 26 2014 10:57 m4ini wrote:On April 26 2014 10:52 itsjustatank wrote:On April 26 2014 10:44 m4ini wrote:On April 26 2014 10:34 itsjustatank wrote:On April 26 2014 10:30 m4ini wrote:On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is. Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No. You missed the in-person interviews. You don't need a phone there. Not to mention, if random person A doesn't receive the call, random person B does. It doesn't make it less random. It samples a slice of people that are not representative of the population.
Funny, the last time i checked, these methods are widely used for election opinion surveys, for example. Ah, there you're coming from. Well that's easy. From here.Obviously, if you call certain groups like family members and certain political groups where you can predict the outcome, randomness goes down the drain. Unless you can prove that in this case though, it's as random as it gets. It's pretty much good enough for every single survey. You still miss the point. Phone samples, no matter how they are conducted, only sample the portion of the population with phones, who are able to answer phones at the time you call, and who care strongly enough either way about the issue to want to respond (or you are otherwise coercing them to respond) . They are all absolutely worthless. That might be your opinion, even if you're wrong there (you can't always ask every single person in a country). That does not make it less random. It might make it not representative enough for you, which is totally fine - but not "not random". Especially since they not only did RDD, but also face-to-face surveying. If a person doesn't care, he has no opinion on that matter and completely falls out of the picture. You can ask everyone in a country. That is what a census is. You can also develop proper stratified random samples using methods that are not phone sampling. Just because it has "random" in it doesn't mean it is statistically random, a state that allows you to make inferences about the true population.
No. A census is extremely different from a survey, that's so farfetched, i don't even know where to begin. You know that in a census, you're obliged to answer, yes? By law? Not to mention, that even the census is only "accurate" once every ten years, since in between those years, they use basically a survey (microcensus)?
I'm pretty sure you can develop proper random samples without phone sampling, but phone sampling by itself doesn't make a survey biased. It can, yes. But it's kinda moot to discuss this, since you won't have the means to prove that they did indeed worked biased.
The only gripe i would have with this survey (which i btw still didn't bother to read), is that the sample size is too small. By, for my taste, quite a margin.
To make it a bit clearer: Phone interviews in large cities with random digit dialing, in-person interviews in smaller cities with randomly chosen addresses.
Thanks, russian to german in googletranslate kinda.. borks everything a bit.
|
United Kingdom13775 Posts
To make it a bit clearer: Phone interviews in large cities with random digit dialing, in-person interviews in smaller cities with randomly chosen addresses.
Whether or not you like this poll is up to you. Personally, I feel that they ask a whole lot of loaded questions.
|
Just by overlooking the questions, they're pretty loaded, yes. But, the answers are more interesting, they're more or less in line with what people think here anyway.
Yanukovich can go to hell, the acting government is widely not considered legitimate, etc. Just by judging the couple of questions you translated though, since questions like "what today is, because you fear rather" are a bit.. well.
|
United Kingdom13775 Posts
A few other questions translated:
3. The Maidan was a: Protest against Yanukovych corruption/ Government takeover by the western opposition party/ something else/ don't know / no answer
6. Who is responsible for the Maidan: Yanukovych / regional party / Berkut / protestors / opposition leaders / the West (US/EU) / Russia / Other / Don't Know / No Answer
Yes/ No / Don't Know / No Answer 16.1 Are Ukraine and Russia at war? 16.2 Is Russia unlawfully meddling in Ukraine? 16.3 Did Russia organize the separatists? 16.4 Is Russia protecting the interests of Russian speakers in Ukraine? 16.5 True/False: Does Russia have no control over events in Ukraine? 16.6 True/False: Russia wants a united Ukraine?
20.1 What relations should Ukraine and Russia have? Same as most other countries / Independent but closely allied (such as NAFTA) / Annexation / Don't Know / No Answer
22. If Russia invades, you plan to: Fight against them / Welcome them / Join Russian army / Stay home and not get involved / Don't Know / No Answer
|
So Putin has cut off all communication to the White House. Sooner or later they will cut off access to the ISS, which would piss off the Eu, Japan.
This is becoming NASA's wet dream.
|
On April 26 2014 10:57 m4ini wrote:Show nested quote +On April 26 2014 10:52 itsjustatank wrote:On April 26 2014 10:44 m4ini wrote:On April 26 2014 10:34 itsjustatank wrote:On April 26 2014 10:30 m4ini wrote:On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is. Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No. You missed the in-person interviews. You don't need a phone there. Not to mention, if random person A doesn't receive the call, random person B does. It doesn't make it less random. It samples a slice of people that are not representative of the population.
Funny, the last time i checked, these methods are widely used for election opinion surveys, for example. Ah, there you're coming from. Well that's easy. From here.Obviously, if you call certain groups like family members and certain political groups where you can predict the outcome, randomness goes down the drain. Unless you can prove that in this case though, it's as random as it gets. It's pretty much good enough for every single survey. You still miss the point. Phone samples, no matter how they are conducted, only sample the portion of the population with phones, who are able to answer phones at the time you call, and who care strongly enough either way about the issue to want to respond (or you are otherwise coercing them to respond) . They are all absolutely worthless. That might be your opinion, even if you're wrong there (you can't always ask every single person in a country). That does not make it less random. It might make it not representative enough for you, which is totally fine - but not "not random". Especially since they not only did RDD, but also face-to-face surveying. If a person doesn't care, he has no opinion on that matter and completely falls out of the picture.
What he's saying is not an opinion. If you were to take a course in stats or ask any statistician they would say the same thing, statistically phone surveys are not real random sampling because not everyone has a phone and is available at the time of the call.
|
Thanks, Legallord, for preliminary translations!
|
On April 26 2014 12:55 LegalLord wrote: A few other questions translated:
3. The Maidan was a: Protest against Yanukovych corruption/ Government takeover by the western opposition party/ something else/ don't know / no answer
6. Who is responsible for the Maidan: Yanukovych / regional party / Berkut / protestors / opposition leaders / the West (US/EU) / Russia / Other / Don't Know / No Answer
Yes/ No / Don't Know / No Answer 16.1 Are Ukraine and Russia at war? 16.2 Is Russia unlawfully meddling in Ukraine? 16.3 Did Russia organize the separatists? 16.4 Is Russia protecting the interests of Russian speakers in Ukraine? 16.5 True/False: Does Russia have no control over events in Ukraine? 16.6 True/False: Russia wants a united Ukraine?
20.1 What relations should Ukraine and Russia have? Same as most other countries / Independent but closely allied (such as NAFTA) / Annexation / Don't Know / No Answer
22. If Russia invades, you plan to: Fight against them / Welcome them / Join Russian army / Stay home and not get involved / Don't Know / No Answer Imagine living in eastern Ukraine and someone from Kiev calls you on your home phone number and asks 'do you agree with the president? Or are you against him?' Before you know it you're on a list.
|
On April 26 2014 17:02 zeo wrote:Show nested quote +On April 26 2014 12:55 LegalLord wrote: A few other questions translated:
3. The Maidan was a: Protest against Yanukovych corruption/ Government takeover by the western opposition party/ something else/ don't know / no answer
6. Who is responsible for the Maidan: Yanukovych / regional party / Berkut / protestors / opposition leaders / the West (US/EU) / Russia / Other / Don't Know / No Answer
Yes/ No / Don't Know / No Answer 16.1 Are Ukraine and Russia at war? 16.2 Is Russia unlawfully meddling in Ukraine? 16.3 Did Russia organize the separatists? 16.4 Is Russia protecting the interests of Russian speakers in Ukraine? 16.5 True/False: Does Russia have no control over events in Ukraine? 16.6 True/False: Russia wants a united Ukraine?
20.1 What relations should Ukraine and Russia have? Same as most other countries / Independent but closely allied (such as NAFTA) / Annexation / Don't Know / No Answer
22. If Russia invades, you plan to: Fight against them / Welcome them / Join Russian army / Stay home and not get involved / Don't Know / No Answer Imagine living in eastern Ukraine and someone from Kiev calls you on your home phone number and asks 'do you agree with the president? Or are you against him?' Before you know it you're on a list.
OR
Imagine living in the eastern Ukraine and someone claiming to be from Kiev calls on your home phone number and ask "do you agree with the president? Or are you against him?" Before you know your're on a Russian intelligence's agency's list.
To be safest as possible is to stay put and offer no comment to anyone. As long as armed men roam around.
|
G7 statementWe, the leaders of Canada, France, Germany, Italy, Japan, the United Kingdom, the United States, the President of the European Council and the President of the European Commission, join in expressing our deep concern at the continued efforts by separatists backed by Russia to destabilize eastern Ukraine and our commitment to taking further steps to ensure a peaceful and stable environment for the May 25 presidential election. We welcomed the positive steps taken by Ukraine to meet its commitments under the Geneva accord of April 17 by Ukraine, Russia, the European Union, and the United States. These actions include working towards constitutional reform and decentralization, proposing an amnesty law for those who will peacefully leave the buildings they have seized in eastern Ukraine, and supporting the work of the Organization for Security and Cooperation in Europe (OSCE). We also note that the government of Ukraine has acted with restraint in dealing with the armed bands illegally occupying government buildings and forming illegal checkpoints. In contrast, Russia has taken no concrete actions in support of the Geneva accord. It has not publicly supported the accord, nor condemned the acts of pro-separatists seeking to destabilize Ukraine, nor called on armed militants to leave peacefully the government buildings they’ve occupied and put down their arms. Instead, it has continued to escalate tensions by increasingly concerning rhetoric and ongoing threatening military manoeuvers on Ukraine’s border. We reiterate our strong condemnation of Russia’s illegal attempt to annex Crimea and Sevastopol, which we do not recognize. We will now follow through on the full legal and practical consequences of this illegal annexation, including but not limited to the economic, trade and financial areas. We have now agreed that we will move swiftly to impose additional sanctions on Russia. Given the urgency of securing the opportunity for a successful and peaceful democratic vote next month in Ukraine’s presidential elections, we have committed to act urgently to intensify targeted sanctions and measures to increase the costs of Russia’s actions. Russia’s actions in Ukraine and the response from the international community already have imposed significant costs on its economy. While we continue to prepare to move to broader, coordinated sanctions, including sectoral measures should circumstances warrant, as we committed to in The Hague on March 24, we underscore that the door remains open to a diplomatic resolution of this crisis, on the basis of the Geneva accord. We urge Russia to join us in committing to that path. Source.
|
On April 26 2014 16:04 hunts wrote:Show nested quote +On April 26 2014 10:57 m4ini wrote:On April 26 2014 10:52 itsjustatank wrote:On April 26 2014 10:44 m4ini wrote:On April 26 2014 10:34 itsjustatank wrote:On April 26 2014 10:30 m4ini wrote:On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is. Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No. You missed the in-person interviews. You don't need a phone there. Not to mention, if random person A doesn't receive the call, random person B does. It doesn't make it less random. It samples a slice of people that are not representative of the population.
Funny, the last time i checked, these methods are widely used for election opinion surveys, for example. Ah, there you're coming from. Well that's easy. From here.Obviously, if you call certain groups like family members and certain political groups where you can predict the outcome, randomness goes down the drain. Unless you can prove that in this case though, it's as random as it gets. It's pretty much good enough for every single survey. You still miss the point. Phone samples, no matter how they are conducted, only sample the portion of the population with phones, who are able to answer phones at the time you call, and who care strongly enough either way about the issue to want to respond (or you are otherwise coercing them to respond) . They are all absolutely worthless. That might be your opinion, even if you're wrong there (you can't always ask every single person in a country). That does not make it less random. It might make it not representative enough for you, which is totally fine - but not "not random". Especially since they not only did RDD, but also face-to-face surveying. If a person doesn't care, he has no opinion on that matter and completely falls out of the picture. What he's saying is not an opinion. If you were to take a course in stats or ask any statistician they would say the same thing, statistically phone surveys are not real random sampling because not everyone has a phone and is available at the time of the call.
The limitations of survey data are well known. However carefully designed pre-tested surveys with response rates over 10% do allow you to make inferences and are certainly not worthless. There may be no reason to suspect that people with phones that are available at the time of calling are biased compared to the population.
That said, this survey does seem poorly designed.
|
On April 26 2014 17:32 Saihv wrote:Show nested quote +On April 26 2014 17:02 zeo wrote:On April 26 2014 12:55 LegalLord wrote: A few other questions translated:
3. The Maidan was a: Protest against Yanukovych corruption/ Government takeover by the western opposition party/ something else/ don't know / no answer
6. Who is responsible for the Maidan: Yanukovych / regional party / Berkut / protestors / opposition leaders / the West (US/EU) / Russia / Other / Don't Know / No Answer
Yes/ No / Don't Know / No Answer 16.1 Are Ukraine and Russia at war? 16.2 Is Russia unlawfully meddling in Ukraine? 16.3 Did Russia organize the separatists? 16.4 Is Russia protecting the interests of Russian speakers in Ukraine? 16.5 True/False: Does Russia have no control over events in Ukraine? 16.6 True/False: Russia wants a united Ukraine?
20.1 What relations should Ukraine and Russia have? Same as most other countries / Independent but closely allied (such as NAFTA) / Annexation / Don't Know / No Answer
22. If Russia invades, you plan to: Fight against them / Welcome them / Join Russian army / Stay home and not get involved / Don't Know / No Answer Imagine living in eastern Ukraine and someone from Kiev calls you on your home phone number and asks 'do you agree with the president? Or are you against him?' Before you know it you're on a list. OR Imagine living in the eastern Ukraine and someone claiming to be from Kiev calls on your home phone number and ask "do you agree with the president? Or are you against him?" Before you know your're on a Russian intelligence's agency's list. To be safest as possible is to stay put and offer no comment to anyone. As long as armed men roam around. Which is why I stated a hundred pages ago that such polls should not be taken seriously.
|
On April 26 2014 16:04 hunts wrote:Show nested quote +On April 26 2014 10:57 m4ini wrote:On April 26 2014 10:52 itsjustatank wrote:On April 26 2014 10:44 m4ini wrote:On April 26 2014 10:34 itsjustatank wrote:On April 26 2014 10:30 m4ini wrote:On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is. Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No. You missed the in-person interviews. You don't need a phone there. Not to mention, if random person A doesn't receive the call, random person B does. It doesn't make it less random. It samples a slice of people that are not representative of the population.
Funny, the last time i checked, these methods are widely used for election opinion surveys, for example. Ah, there you're coming from. Well that's easy. From here.Obviously, if you call certain groups like family members and certain political groups where you can predict the outcome, randomness goes down the drain. Unless you can prove that in this case though, it's as random as it gets. It's pretty much good enough for every single survey. You still miss the point. Phone samples, no matter how they are conducted, only sample the portion of the population with phones, who are able to answer phones at the time you call, and who care strongly enough either way about the issue to want to respond (or you are otherwise coercing them to respond) . They are all absolutely worthless. That might be your opinion, even if you're wrong there (you can't always ask every single person in a country). That does not make it less random. It might make it not representative enough for you, which is totally fine - but not "not random". Especially since they not only did RDD, but also face-to-face surveying. If a person doesn't care, he has no opinion on that matter and completely falls out of the picture. What he's saying is not an opinion. If you were to take a course in stats or ask any statistician they would say the same thing, statistically phone surveys are not real random sampling because not everyone has a phone and is available at the time of the call.
Feel free to ask a statistician, but make sure to not leave out the random face to face interviews. That survey was not only conducted by phone. It's moot anyway since 3200 people (+-) are nowhere near a big enough sample size - but the method by itself does not make it biased. Certainly not to a point where you could disregard the statistic based on that (if, obviously, the sample size is realistic).
|
On April 26 2014 22:07 m4ini wrote:Show nested quote +On April 26 2014 16:04 hunts wrote:On April 26 2014 10:57 m4ini wrote:On April 26 2014 10:52 itsjustatank wrote:On April 26 2014 10:44 m4ini wrote:On April 26 2014 10:34 itsjustatank wrote:On April 26 2014 10:30 m4ini wrote:On April 26 2014 10:19 itsjustatank wrote:On April 26 2014 09:53 LegalLord wrote:Opening few paragraphs: a little discussion of statistical methods, nothing too meaningful. They did phone and in-person interviews with 3200 random people. The words "phone" and "in-person interviews" do not go with "random" in terms of statistical sampling, because they aren't random. How is phonesampling not random? I didn't follow this particular instance now, but in general, it pretty much is. Does everyone in the population own a phone? Is everyone in the population in the position to answer the phone at the time they call? No. You missed the in-person interviews. You don't need a phone there. Not to mention, if random person A doesn't receive the call, random person B does. It doesn't make it less random. It samples a slice of people that are not representative of the population.
Funny, the last time i checked, these methods are widely used for election opinion surveys, for example. Ah, there you're coming from. Well that's easy. From here.Obviously, if you call certain groups like family members and certain political groups where you can predict the outcome, randomness goes down the drain. Unless you can prove that in this case though, it's as random as it gets. It's pretty much good enough for every single survey. You still miss the point. Phone samples, no matter how they are conducted, only sample the portion of the population with phones, who are able to answer phones at the time you call, and who care strongly enough either way about the issue to want to respond (or you are otherwise coercing them to respond) . They are all absolutely worthless. That might be your opinion, even if you're wrong there (you can't always ask every single person in a country). That does not make it less random. It might make it not representative enough for you, which is totally fine - but not "not random". Especially since they not only did RDD, but also face-to-face surveying. If a person doesn't care, he has no opinion on that matter and completely falls out of the picture. What he's saying is not an opinion. If you were to take a course in stats or ask any statistician they would say the same thing, statistically phone surveys are not real random sampling because not everyone has a phone and is available at the time of the call. Feel free to ask a statistician, but make sure to not leave out the random face to face interviews. That survey was not only conducted by phone. It's moot anyway since 3200 people (+-) are nowhere near a big enough sample size - but the method by itself does not make it biased. Certainly not to a point where you could disregard the statistic based on that (if, obviously, the sample size is realistic).
3200 quite a large sample.
|
* Companies find it tough to issue bonds or obtain loans * Corporate bond sales four-fifths down on last year * Analyst calls short-term debt 'Russia's soft underbelly' By Sujata Rao LONDON, April 25 (Reuters) - Usually prolific borrowers on global markets, Russian companies are finding their funding lifeblood cut off by banks and asset managers who fear their investments will get caught up in the standoff between Moscow and the West. With military tensions running high between Russia and Ukraine and a looming threat of tough Western economic sanctions, it is getting harder for companies to issue bonds or obtain loans - a situation that could eventually threaten some of them with default. Corporate bond sales from Russia have ground to a standstill, amounting to just $4.6 billion so far this year. That is a fifth of year-ago levels and under 5 percent of what emerging market firms have raised, according to Thomson Reuters data. Worse still, it is a tiny proportion of the $150 billion or so of the hard currency debt due to be repaid by Russian companies in 2014 Some of that debt was taken out via loans, but the picture there is no better than on bond markets. Russian loans worth $34 billion mature this year but just $4 billion of this has been raised, according to data compiled by Thomson Reuters Loan Pricing Corp. For years, Russia has been a favourite credit for investors, boasting an investment grade credit rating, half a trillion dollars in central bank reserves, big cash-generating companies and a debt ratio of just a third of annual economic output. But in the minds of investors, all that is now overshadowed by the threat of sanctions. They could include freezes on assets or money transfers, restrictions on exports or even a bar on holding Russian companies' stocks and bonds, they fear. 'POLITICAL CALL' "The first call one has to make is the political one. The political turmoil and the possibility of sanctions make everyone cautious on holding Russia. The companies' fundamentals and cash balances are a secondary issue," said Hakan Enoksson, head of fixed income at RBC Wealth Management. Enoksson says he has had no exposure to Russian debt since early March and does not plan any in the near future, given the fear of more losses as more investors exit. Already, U.S. sanctions on businessmen such as Russian Railways boss Vladimir Yakunin and oil tycoon Gennady Timchenko have prompted a selloff in the dollar bonds issued by their companies as investors fear reputational damage. Gas company Novatek for instance, where Timchenko has a stake, has seen the yield on its 2016 dollar bond rise to 5.36 percent, the highest since December 2011, almost double where it was at the end of last year. Yield spreads on Russian companies' dollar bonds have spiralled across the board this year. The implication of all this? A major funding crunch, or even a series of defaults, unless markets pick up soon. Standard Bank analyst Tim Ash terms Russia's short-term debt - what's maturing in the coming year - its "soft underbelly". "I think this debt will be increasingly difficult to refinance if, as seems likely, relations with the West further deteriorate over Ukraine, as the sanctions list is expanded and few foreign institutions are willing to finance Russian entities," Ash said. END OF THE AFFAIR All this marks a major hiatus in a decade-long love affair between Western investors and Russia's corporate sector, which has seen companies - oil and mining firms as well as telecoms and banks - chalk up overseas debt of around $650 billion. They now account for over a tenth of the main emerging corporate bond index, JPMorgan's CEMBI. Even now, the default threat is not immediate - data compiled by various banks shows most big Russian companies have cash balances that are up to three times the value of short-term debt repayments. Prominent exceptions include diamond miner Alrosa and oil major Rosneft, research from Barclays shows. There are also expectations that Russia's government will help companies avoid default, as it did during the 2008-2009 global crisis. Crucially though, it could be less willing this time to ensure Western investors get paid. "I'm looking at all the Russian companies we hold in terms of their total debt, how much short-term debt is maturing and what is their free cash position," said Okan Akin, an investment strategist for fixed income at U.S. asset manager AllianceBernstein. "At the end of the day, if sanctions are imposed it is not clear if the Russian government would step in to help companies to repay debt." Fund managers such as Samantha Lamb at Standard Life in London are also reluctant to dump Russia entirely - she holds bonds from Gazprom and Vimpelcom for instance, noting they will not need external funding for the next year or so. But with lenders reluctant to engage, companies may increasingly have to run down their coffers. State development bank VEB was this week forced to repay maturing debt worth $2.5 billion, rather than refinance it with a fresh loan. Another bank, VTB, will be forced to repay a $3.13 billion due in July, bankers said, describing a rollover of the loan as "out of reach" for the company. Meanwhile Standard & Poor's decision on Friday to cut Russia's rating - placing it on the brink of becoming a junk credit - will make borrowing even more difficult and costly. "If Russia goes to junk, that would have consequences for the private sector," said Cristian Maggio, an analyst at TD Securities. "Every company rated above would be cut to the same rating. That could trigger pretty sharp capital outflows." (Additional reporting by Carolyn Cohn; Editing by Mark Trevelyan) Source.
|
|
|
|