[请求] 请教哪位高人一道Social Statistics的题目。着急。

[如果时光可以倒流]

Refusing to be counted:

Steeh (1981) analyzed refusal rates for surveys conducted by the University of Michigan’s Survey Research Center and comes up with the following result: R = 0.198Q + 5

Where R = predicted refusal rate, measured as the percentage of potential respondents who refused to be interviewed;
and Q = number of quarter years, with the first quarter year beginning in Jan. 1952.

The r2 for this least squares regression equation was 0.89.

Source: Steeh, C. G. (1981). “Trends in nonresponse rates, 1952-1979.” Public Opinion Quarterly 45: 40-57

a) What is the predicted refusal rate for surveys conducted in Jan. 1977? Please show your work.

b) Why would you not trust this equation for studies conducted in 1995?

我更想知道第二个问题的答案，提示也好，应该从哪里下手呀？多谢了。

 

[jinghuacao]

倒！

Q=（1977-1952）* 4 = 100
R = 0。198*100+5 = 24。8

为什么不能TRUST，是因为整个统计方程的抽样期是1952 - 1979，1995已经在这个区间外面了。REGRESSION的预测模式当且仅当下一次样本参数落在方程有效区间内有效。这里拿时间做参数，所以有点晕。

 

[如果时光可以倒流]

感谢感谢。

第二个题目的答案真的只有这么简单吗？没有其他任何的因素了吗？“REGRESSION的预测模式当且仅当下一次样本参数落在方程有效区间内有效”这是一条定律吗？或者你能帮我解释一下为什么落在区间外就没有效了的原因吗？对不起也许问题比较弱智。多谢你了。:thanks:

最初由 jinghuacao 发布
倒！

Q=（1977-1952）* 4 = 100
R = 0。198*100+5 = 24。8

为什么不能TRUST，是因为整个统计方程的抽样期是1952 - 1979，1995已经在这个区间外面了。REGRESSION的预测模式当且仅当下一次样本参数落在方程有效区间内有效。这里拿时间做参数，所以有点晕。

 

[spsun1]

There is a specification problem too. We have to understand or whether the theorm is correct. Why the refusal rate is a function of time.

I do not think there is a reason to believe/support the refusal rate should increase according time. Even the R2 is high, it is only a statistical result or maybe other reasons cause the refusal rate to increase through the observation period, which may not hold in 1995.

In addition to R2, the standard deviations of the estimation is also very important. If it is very large,even the estimated coefficient is positive, it is not significant from zero, etc....

 

[如果时光可以倒流]

太感谢了楼上的两位。

让我好好想象。多谢你们的提示。

 

[如果时光可以倒流]

还有一道题。

The prestige of gender segregated occupations

Conflicting findings have been reported concerning the relationship between the prestige of an occupation and its gender composition. Some researchers report that the higher the percentage of males in a given occupation, the higher is its occupational prestige. Others, such as Glick (1995: 579), fail to find such a relationship of this form. They note that “all of the highly prestigious jobs…were male-dominated jobs” but that there was an “even greater male dominance in very low prestige, blue collar work. In short, there is a relationship between sex and prestige, but it is not a linear one; the percentage of male incumbents is high in both the lowest and highest status jobs.”

Draw hypothetical scatter plots that depict the two sets of findings. Provide appropriate titles and labels for the two plots. Let occupational prestige be measured on a scale from 0 to 100, where 100 represents the highest prestige level. Gender segregation of occupations is usually measured as the percentage of males in the various occupations.
(Source: Glick, P., K. Wilk, et al. (1995). “Images of occupations: Components of gender and status in occupational stereotypes.” Sex Roles 32(9/10): 565-582.)

这道题什么数字也没有给出，除了知道两个极限值100和100%。那么我该怎么画这个scatter plots呀，没有slope， 也没有交叉值，难道我就大概画个位置和形状就可以了吗？

 

[spsun1]

画个位置和形状就可以.

Let set the Y axis is the degree of occupational prestige and X is the % of male.

The first one graph would be with most of the dotes forming a straight band upwards toward right (positive relationship).

The second one would be th same on the top but the band will come down toward right, as low prestige jobs also have a high percentage of males.

Do you agree?

 

[如果时光可以倒流]

真的画个大概就可以吗？will that work？你以前这样做过是吗？

首先表示感谢。

其次呢，我想第二个图我跟你想得有点不一样，因为照你说的话，那描述就变成了，在low prestige， 分别有最多和最少的males(%)。 不成一个assumption.

所以能如果照题中所说的，在high and low prestige，males都占有很高的比例的话，图形的形状应该是一个口向右开的钟形。

可是我觉得这样很不方便，就把independent variable 换成prestige了，这样的话，图形是一个口向上的钟形。而且我认为在题中更强调的是从male-dominated (也就是比例) 推出high and low prestige, respectively.

不值这样可不可以，你认为呢？

而且，我是不是不能把点们都画在直线上呀，还要装作把点散布在线周围的样子？

最初由 spsun1 发布
画个位置和形状就可以.

Let set the Y axis is the degree of occupational prestige and X is the % of male.

The first one graph would be with most of the dotes forming a straight band upwards toward right (positive relationship).

The second one would be th same on the top but the band will come down toward right, as low prestige jobs also have a high percentage of males.

Do you agree?

 

[如果时光可以倒流]

有点不明白你的答案。

求Q时，你为什么要乘以4呀，这是从那跑出来的一个数字呀？

我认为应该是：Q = (1977-1967)/15 + 1 = 5/3
R = 0.198 * 5/3 +5 = 5.33

你怎么看？

最初由 jinghuacao 发布
倒！

Q=（1977-1952）* 4 = 100
R = 0。198*100+5 = 24。8

为什么不能TRUST，是因为整个统计方程的抽样期是1952 - 1979，1995已经在这个区间外面了。REGRESSION的预测模式当且仅当下一次样本参数落在方程有效区间内有效。这里拿时间做参数，所以有点晕。

 

[spsun1]

还要装作把点散布在线周围的样子. If all the points on the line, it is would be a perfect fit. In reality, this will not be the case.

Yes. It will be an U shape toward right. This is the reason that the theory is not conclusive, as noted by the second author or study or the relationship is not linear. At the right side, e.g., 90% male dominated point, the Y values are scattered at both high and low.

I think it is more appropriate for the prestige to be the dependent variable, as it is not "right" to estimate the % of male from the level of prestige.

do you agree?



Times 4 is because it is on a quaterly basis. When you run the regression the value of Q is 1, 2,3, 4, for the first year and then 5,6,7,and 8 for the second year....

 

[spsun1]

1977 Jan should be the first quarter of 1977, so the value of Q is 101 not 100.

Please use a speadsheet to continue the count to see whether it shuld be 101 or 100.

 

[如果时光可以倒流]

谢谢，让我先想想在回复你。多谢了。

最初由 spsun1 发布
1977 Jan should be the first quarter of 1977, so the value of Q is 101 not 100.

Please use a speadsheet to continue the count to see whether it shuld be 101 or 100.

 

[jinghuacao]

Re: 有点不明白你的答案。

最初由 如果时光可以倒流 发布
求Q时，你为什么要乘以4呀，这是从那跑出来的一个数字呀？

我认为应该是：Q = (1977-1967)/15 + 1 = 5/3
R = 0.198 * 5/3 +5 = 5.33

你怎么看？

Q = number of quarter years, with the first quarter year beginning in Jan. 1952.


不知道你1967这个数字是从哪里拿出来的，我没看懂，汗

 

[jinghuacao]

最初由 spsun1 发布
1977 Jan should be the first quarter of 1977, so the value of Q is 101 not 100.

Please use a speadsheet to continue the count to see whether it shuld be 101 or 100.

这。。。这。。。这。。。这。。。这是我崇拜了多年的陈景润同志啊，我对您的景仰犹如滚滚长江，滔滔不绝。

 

[如果时光可以倒流]

Re: Re: 有点不明白你的答案。

我错了，没好好看题，看了一个quarter不知怎么就跟15挂上钩了，全理解错了。:blowzy:

最初由 jinghuacao 发布


Q = number of quarter years, with the first quarter year beginning in Jan. 1952.


不知道你1967这个数字是从哪里拿出来的，我没看懂，汗