I do not think I have done one of these since I moved to Marmots Hole, but ever so often I like to take a look at the numbers behind the numbers. I do this since some of the reporting about Korea is never quite as it seems, and can be proven by a quick and dirty look at the numbers. To give you an idea you can look at my rarely maintained blog Dram Man.
Last week many around the Korean Blogo-sphere referred to this article:
As many as 70.2 percent of housewives aged 30 to 40 replied that they would not buy or were disinclined to buy U.S. beef, Korea Gallup said in the survey. Only 10.6 percent answered that they would buy American beef.
Now that was the rather incendiary part of the article, particularly given the current flogging of the FTA dead horse. However it was the very next paragraph that hit me:
Korea Gallup conducted the survey-commissioned by the Ministry of Agriculture and Forestry-on consumption sentiment for American beef on 1,213 housewives from across the country.
So given that information about the poll is Korea as anti-US beef as the previous paragraph suggests?
Note the slight disconnect in reporting. The “70.2 of housewives aged 30 to 40” versus the “survey…on 1,213 housewives”. It does not explicitly state that the 1,213 housewives are “30 to 40”. Assuming the marriages are normally distributed the same as age (i.e. the percent of population of an age group is the same in the survey), this leads to a “30 to 40” a figure of 266 housewives*.
Meaning in a country of 48 million people, 266 of those are supposed to be representative. In other words a sample size is only .0000055% of the population!
Once you get past the quantitative results of the survey you have the qualitative aspects, mainly a big sample bias. In surveying “housewives” it assumes that such a group is the key decision maker in the purchase/consumption of beef. The first problem, perhaps over looked, is “What about the men?”. A bit more higher up is “What about restaurants and institutional purchasers?” which I recall reading somewhere is the largest purchaser of beef by a very high margin.
Even if you want to limit it to females in a household, you have problems. A recent ILO study says Koreans pay the most in the world for beef. Now I ask you given prices like that, who do you think is deciding to purchase beef on a regular basis? I think we can agree that older couples make more than younger ones, and more obviously dual income couples make more than single income ones. Therefore can you honestly tell me at 30-40 year old single income households are the best gauges of beef consumption?
* Figure extrapolated from 2000 data found on the Korean National Statistics Office website. Women ages 0-20 were excluded to reflect they are not eligible for marriage normally.
15 Comments
Well, I believe in the idea of statistics and random sampling, and I believe Gallup would probably do a good job in getting the sample. I think this is more of a problem with interpretation. As you point out, they probably looked at each group that showed up in their sample, adn chose the group that responded with the most sentiment not to buy U.S. beef. Then they used this as their main headline grabber, instead of really being representative and saying what percent of total respondents said they would not buy beef. I also saw the news story on meant being the most expensive in Korea of anywhere in the world. Ridiculous! I refuse to buy and eat beef that costs that much because of import taxes!
More monkey business with the same numbers is found
here — sixth paragraph. Same survey cited, but this time it’s 70.2% of the 1,213 housewives (apparently of all age groups) that won’t buy American beef.
I reckon US beef producers wouldn’t be displeased with 10.6% of the housewives’ market.
BTW, any word of the missing 19.2%?
seouldout, the irony of it all is that the Korean media likes to give the impression that Hyundai’s 5% share of the North American auto market is a clearly shows that Korean-made cars are superior products.
Here’s an interesting report, in the Hanky of all places, about the ILO’s findings that beef prices in Korea, at US$56.44/kg, are 6 times US prices and nearly 30% higher than Japanese prices:
http://english.hani.co.kr/arti.....85906.html
The article says the ILO report attributes the differential to the high cost of land in Korea and the (resulting?) inability to effect economies of scale in production.
Korea has similarly higher prices for pork
The explanation doesn’t seem entirely plausible to me because rural land in Korea is pretty cheap as long as it isn’t in the path of urban development.
Moreover, Korea also has disproportionately high prices for seafood, which points up the real culprit here - prices of not just food but all manner of consumer goods are grotesquely high in Korea because of the tariff and non-tariff barriers to entry of foreign produced goods.
Sorry… is a clear indication that Korean-made…
do Korean consumers not realize this? FTA = cheaper goods across the board?
with proper random sampling, 266 is a statistically significant representation. But I agree with the other points raised.
266 people would give you a margin of error of +/- 6.1%, 19 times out of 20.
Two of my favorite quotations about statistics and numbers…
1) Figures don’t lie, but liars figure — Samuel Clemens
2) Facts are stubborn things, but statistics are more pliable — Mark Twain
266 is statistically significant representation for that particular group (again, assuming random sampling techniques were valid), but that says nothing about women outside the 30-40 year old range. That is the point of the post.
They are being suspiciously picky about which group they choose to report on. Why not just report on all age groups? Obviously, the numbers for other age groups didn’t pan out the way they wanted.
In the other article referred to by Seoulout above, they are briefly referring back to the originally study (the article is mostly on the price of beef in Korea). I suspect that this reporter didn’t read the original reports carefully and just assumed that 70% referred to all housewives.
By the way, Dramman, what happened to the post you put up a few weeks ago about Oriental medicine?
I am aware that 266 is approching a “large” sample size. However that holds if they are randomly selected, which they are not. They are expressly 30-40 year old houswives which are being extrapolated by many to be representive of Korea at large, or at best key beef buyers.
PP>Don’t ask about Quasimoto.
In Canada, most papers will include information about how the survey is “plus or minus 1 in 19 of 20 cases,” but leave out the much more important bits, like how many people were included and what they were actually asked.
I can just imagine the question they asked the Korean housewives:
“Taking into consideration all of the possible health risks associated with American beef, would you risk your children’s health by buying it?”
“Do you want to expose your family to mad cow disease by purchasing American beef?”
“Has the alarming increase in reported cases of mad cow disease in American beef made you more or less inclined to purchase non-Korean beef?”
“Korean beef is better than disease-filled American beef. How does this fact affect your shopping habits?”
“Only traitors and people who hate their own children buy American beef. How does this fact affect your shopping habits?”
“You have been specially selected to help represent the opinions of the fatherland. Do you prefer our beef, or beef from the people who cut our country in half?”
JD, is right about the biggest cause of the heavily skewed responses: question bias. The latter half of his hypothetical questions were probably used.
On the KTX to Puson a couple weeks ago they repeatedly showed clips of cows staggering around and bellowing. Farmers wearing masks were shaking their heads sadly.
I think you are over-emphasizing these skewed figures. The journalists here are simply ham-handed. I truly believe there is a pervasive sentiment against US beef and there is alot of propaganda responsible. Main-stay journalism here is as manufactured as The Enquirer back home.
Do they know how many people in the history of man have died from this spogiform thing? My guess is less than ten.
@Maddlew
FWIW, Wikipedia says 170 have contracted it.