112 年 國立政治大學經濟學系碩士班《統計學》
第 1 題25 分
There are four restaurants in the Northtown area. The numbers of burgers sold at the respective restaurants for each of the last 6 weeks are shown below. At the .05 significance level, what is the value of the test statistic and is there a difference in the mean number sold among the four restaurants when the factor of week is considered?
Restaurant
| Week | Metro | Interstate | University | River |
|---|---|---|---|---|
| 1 | 124 | 160 | 320 | 190 |
| 2 | 234 | 220 | 340 | 230 |
| 3 | 430 | 290 | 290 | 240 |
| 4 | 105 | 245 | 310 | 170 |
| 5 | 240 | 205 | 280 | 180 |
| 6 | 310 | 260 | 270 | 205 |
登入後即可作答並保存紀錄。
本題考驗單因子變異數分析(One-way ANOVA)的應用。題目要求檢定不同餐廳(或說不同種類的餐廳)的平均漢堡銷售量是否存在差異,並計算檢定統計量。
核心觀念: 單因子變異數分析用於比較三個或三個以上獨立樣本的平均數。
步驟:
-
設定虛無與對立假設:
- : 四家餐廳的平均漢堡銷售量均相等。()
- : 至少有一家餐廳的平均漢堡銷售量與其他餐廳不同。
-
資料整理與計算:
我們有 4 個處理(餐廳種類)和 6 個觀察值(週數)。- 總觀察數 。
- 總自由度 。
各組(餐廳)的統計量:
- Metro:
- Sum () =
- Mean () =
- Sum of Squares () =
- Interstate:
- Sum () =
- Mean () =
- Sum of Squares () =
- University:
- Sum () =
- Mean () =
- Sum of Squares () =
- River:
- Sum () =
- Mean () =
- Sum of Squares () =
總體平均數:
- 總和 () =
第 2 題25 分
For many years, TV executives used the guideline that 30% of the audience were watching each of the traditional big three prime-time networks and 10% were watching cable stations on a weekday night. A random sample of 500 viewers in the Tampa-St. Petersburg, Florida, area last Monday night showed that 165 homes were tuned in to the ABC affiliate, 140 to the CBS affiliate, and 125 to the NBC affiliate, with the remainder viewing a cable station. What is the value of the test statistic? At the .05 significance level, can we conclude that the guideline is still reasonable? What is the value of the test statistic?
登入後即可作答並保存紀錄。
核心觀念
本題考察「卡方適合度檢定(chi-square goodness-of-fit test)」,用來判斷樣本觀察到的各類別比例,是否符合既定的母體比例。
假設:
- :觀眾收視比例仍為 ABC、CBS、NBC 各 ,有線電視 。
- :實際收視比例不符合上述指導比例。
卡方檢定統計量為
其中 為觀察次數, 為在 成立下的期望次數。
解題方法
樣本數為 。依據既定比例,各類別的期望次數為:
| 頻道類別 | 觀察次數 | 比例 | 期望次數 |
|---|---|---|---|
| ABC | 165 | ||
| CBS | 140 | ||
| NBC | 125 | ||
| 有線電視 |
因此,
第 3 題20 分
An investigation of the effectiveness of a training program to improve customer relationships included a pre-training and a post-training customer survey. To compare the differences, they computed (post-training survey score - pre-training survey score). Seven customers were randomly selected and completed both surveys. The results are shown below. For a 0.05 significance level, what is the value of the test statistic and the decision regarding the hypothesis that the training was effective in improving customer relationships?
The information of question 3.
| Customer | Pre-training Survey | Post-training Survey |
|---|---|---|
| A | 6 | 8 |
| B | 5 | 5 |
| C | 10 | 10 |
| D | 7 | 10 |
| E | 6 | 8 |
| F | 5 | 6 |
| G | 2 | 8 |
登入後即可作答並保存紀錄。
本題考驗配對樣本 t 檢定(Paired Samples t-test),用於檢定兩個相關樣本(在此為同一受試者前後測的得分)的平均數是否存在顯著差異。題目要求判斷訓練是否有效,這意味著我們需要檢定訓練後的平均分數是否顯著高於訓練前的平均分數。
核心觀念: 配對樣本 t 檢定用於分析成對觀測值之間的差異。
步驟:
-
計算差異分數 (Difference Scores, ):
題目已說明計算方式為 (post-training survey score - pre-training survey score)。- A:
- B:
- C:
- D:
- E:
- F:
- G:
-
計算差異分數的平均數 () 與標準差 ():
- 樣本數 。
- 差異分數列表:
- 平均差異 。
- 計算差異分數的變異數 ():
第 4 題30 分
Suppose that the sales manager of a large automotive parts distributor wants to estimate the total annual sales for each of the company's regions. Five factors appear to be related to regional sales: the number of retail outlets in the region, the number of automobiles in the region registered as of April 1, the total personal income recorded in the first quarter of the year, the average age of the automobiles (years), and the number of sales supervisors in the region. The data for each region were gathered for last year. For example, see the following table. In region 1 there were 1,739 retail outlets stocking the company's automotive parts, there were 9,270,000 registered automobiles in the region as of April 1, and so on. The region's sales for that year were $37,702,000.
(1) (10%) The output for all five variables is shown below. What percent of the variation is explained by the regression equations?
(2) (20%) Some result is shown below. Conduct a global test of hypothesis to determine whether any of the regression coefficients are not zero. State null and alternate hypotheses. Identify and compute the test statistic, and formulate a decision rule. Use the .05 significance level.
| Annual Sales ($millions), y | Number of Retail Outlets, x₁ | Number of Automobiles Registered (millions), x₂ | Personal Income ($billions), x₃ | Average Age of Automobiles (years), x₄ | Number of Supervisors, x₅ |
|---|---|---|---|---|---|
| 37.702 | 1,739 | 9.27 | 85.4 | 3.5 | 9.0 |
| 24.196 | 1,221 | 5.86 | 60.7 | 5.0 | 5.0 |
| 32.055 | 1,846 | 8.81 | 68.1 | 4.4 | 7.0 |
| 3.611 | 120 | 3.81 | 20.2 | 4.0 | 5.0 |
| 17.625 | 1,096 | 10.31 | 33.8 | 3.5 | 7.0 |
| 45.919 | 2,290 | 11.62 | 95.1 | 4.1 | 13.0 |
| 29.600 | 1,687 | 8.96 | 69.3 | 4.1 | 15.0 |
| 8.114 | 241 | 6.28 | 16.3 | 5.9 | 11.0 |
| 20.116 | 649 | 7.77 | 34.9 | 5.5 | 16.0 |
| 12.994 | 1,427 | 10.92 | 15.1 | 4.1 | 10.0 |
| Predictor | Coef | SE Coef | T | P |
|---|---|---|---|---|
| Constant | -19.672 | 5.422 | -3.63 | 0.022 |
| Outlets | -0.0000629 | 0.002638 | -0.24 | 0.823 |
| Automobiles | 1.7399 | 0.55530 | 3.15 | 0.035 |
| Income | 0.40994 | 0.04385 | 9.35 | 0.001 |
| age | 2.0357 | 0.8779 | 2.32 | 0.081 |
| bosses | -0.0344 | 0.1880 | -0.18 | 0.864 |
Analysis of Variance
| Source | DF | SS | MS | F | P |
|---|---|---|---|---|---|
| Regression | 1593.81 | ||||
| Error | 9.08 | ||||
| Total |
登入後即可作答並保存紀錄。
核心觀念
本題考多元線性迴歸的整體配適度與整體顯著性檢定。
決定係數 表示應變數總變異中,能由迴歸模型解釋的比例:
整體 檢定則檢驗五個斜率係數是否同時為零。若拒絕虛無假設,代表至少有一個解釋變數與年銷售額有線性關係。
解題方法
(1)計算迴歸解釋的變異百分比
題目給出迴歸平方和 、誤差平方和 ,所以總平方和為:
因此:
換算成百分比:
答:迴歸模型解釋了約 的年銷售額變異。
(2)進行整體 檢定
令 為五個解釋變數的母體迴歸係數,檢定假設為:
共有 筆資料、 個解釋變數,因此自由度為: