| View previous topic :: View next topic |
| Author |
Message |
PatrickGHBusan
Joined: 24 Jun 2008
Location: Busan (1997-2008) Canada 2008 -
|
 Posted: Thu Jul 11, 2013 4:23 am Post subject: |
 |
|
| OneWayTraffic wrote: |
| PatrickGHBusan wrote: |
| Quote: |
| The number of respondents of the survey, however, was only 307 |
This should send up a red flag on any statistical significance the data produced by this survey has.... |
Not at all. Assuming that the data was random we get a standard normal error of 0.65*0.35/SQRT(307) roughly 1%.
Taking a 95% confidence interval we can estimate that the real proportion of couples experiencing domestic violence is highly likely to be in the 63%-67% range. I don't know the sample size of the interracial couples, but if it's smaller then the confidence interval will be wider, but still centered on the 69% quoted.
The only obvious fly in the ointment, is the possibility of bias from the method of sampling, but I would guess underreporting is more likely than overreporting in this case.
If you haven't guessed I now teach High School mathematics. |
That is well said and I agree. Based on this, it seems highly unlikely this poll was designed and delivered accurately or with a high degree of attention towards sample relevance. |
|
| Back to top |
|
 |
Julius
Joined: 27 Jul 2006
|
|
| Back to top |
|
|
Captain Corea
Joined: 28 Feb 2005
Location: Seoul
|
| Posted: Thu Jul 11, 2013 4:59 am Post subject: |
|
|
That's horribly high. Really bad, if accurate.
I'm curious though - is it the same criteria used for that US 20%? I mean, some of the definitions mentioned for abuse seemed pretty specific. |
|
| Back to top |
|
|
PatrickGHBusan
Joined: 24 Jun 2008
Location: Busan (1997-2008) Canada 2008 -
|
| Posted: Thu Jul 11, 2013 6:41 am Post subject: |
|
|
| That number is indeed very high. |
|
| Back to top |
|
|
bigverne
Joined: 12 May 2004
|
| Posted: Thu Jul 11, 2013 5:14 pm Post subject: |
|
|
According to the report, "When multiple replies were allowed, 13.4 percent of the victims in multiracial families said they were physically beaten"
How does this square with the figure that 69% were victims of 'domestic violence'? |
|
| Back to top |
|
|
Captain Corea
Joined: 28 Feb 2005
Location: Seoul
|
| Posted: Thu Jul 11, 2013 5:43 pm Post subject: |
|
|
| Yeah, while I'm not a fan of this type of spouse, "always wanting to know where you are" doesn't necessarily mean Domestic Violence in my book. That's why I was curious as to the numbers Julius put up for the US in comparison. |
|
| Back to top |
|
|
Seoulman69
Joined: 14 Dec 2009
|
| Posted: Thu Jul 11, 2013 6:26 pm Post subject: |
|
|
| So the findings show that the majority of Korean men are abusive husbands. Interesting. |
|
| Back to top |
|
|
OneWayTraffic
Joined: 14 Mar 2005
|
| Posted: Fri Jul 12, 2013 1:15 am Post subject: |
|
|
| PatrickGHBusan wrote: |
| OneWayTraffic wrote: |
| PatrickGHBusan wrote: |
| Quote: |
| The number of respondents of the survey, however, was only 307 |
This should send up a red flag on any statistical significance the data produced by this survey has.... |
Not at all. Assuming that the data was random we get a standard normal error of 0.65*0.35/SQRT(307) roughly 1%.
Taking a 95% confidence interval we can estimate that the real proportion of couples experiencing domestic violence is highly likely to be in the 63%-67% range. I don't know the sample size of the interracial couples, but if it's smaller then the confidence interval will be wider, but still centered on the 69% quoted.
The only obvious fly in the ointment, is the possibility of bias from the method of sampling, but I would guess underreporting is more likely than overreporting in this case.
If you haven't guessed I now teach High School mathematics. |
That is well said and I agree. Based on this, it seems highly unlikely this poll was designed and delivered accurately or with a high degree of attention towards sample relevance. |
Polls tend to reflect the sample population that they came from. There is the small chance of getting sampling error so large that wrong conclusions are drawn, but it does happen at times. Non sampling error is a bigger problem. Before I mentioned that polls reflect the actual population that they come from, which might not be the population one is drawing conclusions about.
My guess is that much of it is to do with different definitions of domestic violence, together with (in the case of foreign spouses) certain power dynamics and correlations to do with the types of men who buy a foreign wife.
At least once in the last year would cover a lot of temporary tiffs, when the real problem is ongoing behaviour. |
|
| Back to top |
|
|
Julius
Joined: 27 Jul 2006
|
| Posted: Fri Jul 12, 2013 3:44 am Post subject: |
|
|
| OneWayTraffic wrote: |
My guess is that much of it is to do with different definitions of domestic violence. |
In the case of the US, the 20% figure concerns the proportion of marriages wherein the couples "slap, shove, hit or otherwise assault eachother". |
|
| Back to top |
|
|
aq8knyus
Joined: 28 Jul 2010
Location: London
|
| Posted: Fri Jul 12, 2013 4:07 am Post subject: |
|
|
The US usually has a higher bar for what it considers violence. For example in its crime statistics they only count 5 types of the most serious violent crimes.
That is why US violent crime rates are lower than many other developed countries that include much broader definitions of violent crime. |
|
| Back to top |
|
|
PatrickGHBusan
Joined: 24 Jun 2008
Location: Busan (1997-2008) Canada 2008 -
|
| Posted: Fri Jul 12, 2013 4:24 am Post subject: |
|
|
The problem of definition is a huge one because it can vary wildly from one country to another and then from one survey to another.
I agree with what Captain said for example: a spouse that wants to know where his or her spouse is all the time may not qualify as domestic violence at all in many surveys, yet in this one it did.
Still, the number if high. |
|
| Back to top |
|
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
