[问题] 请教一道统计题目. 感谢感谢.

[如果时光可以倒流]

Re: Not really...

比方说第一道题，在失业率升高1%的时候，公式里的percentage of female and percentage of non-academics 不会有变化吗？这几个变量之间是有关系的吧。这是我的考虑。我原先也是认为1573。

最初由 如果时光可以倒流 发布
能具体说说吗，
我知道写出来肯定比较麻烦，
不过实在是很难理解，
大概思路就好了，
多谢了。

 

[如果时光可以倒流]

This is great, makes great sense!!

But the only thing is that the data you have referred are the observed, rather than the predicted,which I believe is what's been defined in the principle:

"The fact that no case falls more than 3 Std. Deviations from the PREDICTED values indicates that there are no extreme outliners."

最初由 spsun1 发布
the second question could be answered as follows:

the mean for the unemployment rate is 1.525 and sd is 1.273, so once the unemployment rate is greater than 2.58 or 3.??? times sd , the result would be not reliable. check the book for the value of the confidence level at 95 % and 99%.

 

[如果时光可以倒流]

这就是文科题,没看到是SOCIAL Statistics吗.

这么多人关心我,感动.

最初由 恐龙姐姐 发布


望倒流三思啊！

 

[如果时光可以倒流]

Re: Re: 凡间精灵,你好象对我满了解的样子.

说到听说,凡间精灵,我可是真听说,
你一朋友因为想要嫁外国人这两天都歇斯底里了,到处咬人,
也不知道是谁给她煽风点火,灌输的这种想法.
你说现在全球化了,她怎么还有这么深的种族意识呀,
不能平等的看待别人,非得觉得外国人和中国人就不一样,
是不是比较悲哀.
好好劝劝你朋友,老大不小的人了,
珍惜自己的家庭和老公,
要是真不满意了,就去追求,
别丫酸溜溜的指责这个指责那个.

最初由 凡间精灵 发布
恩，听说了。

 

[jackiezhe]

最初由 spsun1 发布
the second question could be answered as follows:

the mean for the unemployment rate is 1.525 and sd is 1.273, so once the unemployment rate is greater than 2.58 or 3.??? times sd , the result would be not reliable. check the book for the value of the confidence level at 95 % and 99%.

Sorry, I am afraid u cannot do like that. The mean and std of umemployment rate is the data of the observations. The 95% or 99% CI is just the range of variation of that series of observations. When the actual employment rate exceeds that CI, we may conclude that it is not from the same population within which the observations are drawn. But the regression model may be still useful to predicit the correct earning when putting the actual unemployement rate into the equation. So, for question 2, we want to exam the maximum actual umemployment rate which may make our regression model invalid.

 

[spsun1]

yes. you need to state: other things being equal.. everything percentage point change in unemployment rate, the income penalty varies 1,573...

the mean of unemployment rate is 1.525 and sd is 1.273, I think these would give you the information on the range of the data..

 

[spsun1]

you are fine: keep on thinking. you need to understand the theory and assumption etc behind the regression analysis.

typing error: "everything should be ....every one percentage point..

 

[如果时光可以倒流]

可问题是对于Multiple Regression 来说,

并不是我们定义说other things being equal.., 他们就能equal的,因为这几个变量之间是互相影响的,当其中一个增长1%的时候,我们并不能假设其他的变量没有变化,你认为呢?

还有,1.525和1.273都是observed, 不是predicted,在这题中,我们需要的是predicted的数据,不是吗?

谢谢.

最初由 spsun1 发布
yes. you need to state: other things being equal.. everything percentage point change in unemployment rate, the income penalty varies 1,573...

the mean of unemployment rate is 1.525 and sd is 1.273, I think these would give you the information on the range of the data..

 

[功夫]

Re: Re: Re: 凡间精灵,你好象对我满了解的样子.

最初由 如果时光可以倒流 发布
说到听说,凡间精灵,我可是真听说,
你一朋友因为想要嫁外国人这两天都歇斯底里了,到处咬人,
也不知道是谁给她煽风点火,灌输的这种想法.
你说现在全球化了,她怎么还有这么深的种族意识呀,
不能平等的看待别人,非得觉得外国人和中国人就不一样,
是不是比较悲哀.
好好劝劝你朋友,老大不小的人了,
珍惜自己的家庭和老公,
要是真不满意了,就去追求,
别丫酸溜溜的指责这个指责那个.

如果时光可以倒流和凡间精灵，谁更有魅力，我支持谁

 

[如果时光可以倒流]

Re: Re: Re: Re: 凡间精灵,你好象对我满了解的样子.

你这不挺有工夫的吗,功夫,帮我想想题吧.

最初由 功夫 发布
如果时光可以倒流和凡间精灵，谁更有魅力，我支持谁

 

[如果时光可以倒流]

Re: Re: Re: Re: Re: 凡间精灵,你好象对我满了解的样子.

我可比不上凡间精灵,你不用支持我.
只要你帮我看看题就可以了.

最初由 如果时光可以倒流 发布
你这不挺有工夫的吗,功夫,帮我想想题吧.

 

[如果时光可以倒流]

spsun1,这样你连Model2 都没用呀. 怎么会这样?

最初由 spsun1 发布
The maximum unemployment rate should be 22 the constant 36,701/1573 = 22, please do the detailed calculation yourself. At this point the income will be negative..
do these make sense?

 

[如果时光可以倒流]

（54.118+2.86*18.923*100）里的2.86*18.923*100是怎么来的呢?

最初由 jackiezhe 发布
第1问 感觉题目不通顺，不知是否漏打。
第2问 看TABLE 3， Model 2。 设upper limit 为x, 排除另一个负系数变量“Pfemale”时，x可取最大值,使之与其 系数（-1573.282）的积达最大绝对值，此时收入为0。 另一方面，正系数在“Pnonac ”达最大（100 既100%）时，正数 35164+（54.118+2.86*18.923*100） =45988 达最大。最后，x 的的系数绝对值最大可达 -1573.282-2.934*536.181=-3148. 要使 -3148X+45988>=0，此时,x=14.6. 所以失业率最高只能 14.6。超过此值，收入有出现 负数的可能，因此 该公式无效。

以上只是 我的粗略的思路和计算，你应该自己好好想想 怎样陈述和是否正确。

 

[spsun1]

I am still thinking about the second question, as I think 22 is not correct, or we need to say that all other independent variables have to equal to zero. This is far beyond the mean plus 3 standard deviations.

From statistics, 95% of the data should fall within mean + 2.86 x SD. Any estimtion beyond this range would be not reliable or should not be used. make sense?

I suggest that if unemployment rate is greater than its mean value plus 2.86 time SD, then the result will exceed the 95% confidence level. Is 2.86 the coefficient at 95% of confidence level? You need to check the book. This way you will use the information provided in one of the table.

My computer sometime does not display the attachments automatically.

 

[如果时光可以倒流]

spsun1, do you know how should I use table ANOVA for my analysis by chance?

Thanks.