Anda di halaman 1dari 9

LFS PROJECT I – ANSWERS FOR MASTER SAMPLE.

XLS

Graphical Descriptive Techniques


Q1

Pg 27 K&W 5th ed

Distribution of Ages Bin Freq.


10 0
20 3
40
30 23
35
40 34
30 50 25
60 8
Frequency

25

20 70 6
80 1
15
90 0
10
100 0
5 More 0
0
10 20 30 40 50 60 70 80 90 100 More Max 34
Bin

The maximum frequency value is 34 (There are 34 observations whose age is between 30 (exclusive)
and 40 (Inclusive).

WebCT 1 : 34

Q2

Pg 46 K&W 5th ed

African/Black (1) 79
Coloured (2) 10
9%
Indian/Asian (3) 2
2%
African/Black (1) White (4) 9
10%
Sum 100
Coloured (2)

Indian/Asian (3)

79%
White (4)
The largest population group represented in the pie-chart is the African/Black population with 79%

WebCT 2 : 79

Numerical Descriptive Measures


Q3, Q4, Q5

Pg 97 K&W 5th ed

Tools… > Data Analysis… > Descriptive Statistics…

yearsed

Mean 7.46
Standard Error 0.412291
Median 8
Mode 10
Standard
Deviation 4.12291
Sample Variance 16.99838
Kurtosis -0.77938
Skewness -0.48005
Range 15
Minimum 0
Maximum 15
Sum 746
Count 100

WebCT 3: 7.46

WebCT 4: 8

WebCT 5: 15

Q6

Pg 130 K&W 5th ed

Correlation Matrix:

annualsal yearsed
annualsal 1
yearsed 0.452155 1

The coefficient of correlation using the years of education (yearsed) and annual salary (annualsal)
variables is 0.452155 (positive relationship between the two variables)

WebCT 6 : 0.4522
Probability (Pivot Tables)
EXCEL 2007
Insert … Pivot Table
Drag Gender to left column, drag Type to top row, drag Type to body of table. Change
the body of the table from Sum to Count.

EXCEL 2003
Data … Pivot Table and Pivot Chart Wizard
Step 2: Select your data … Next
Step 3: Select Layout (bottom left) …
Drag Gender to left column, drag Type to top row and drag Type (from the full list of
headings) to body of table. Change the body of the table from Sum to Count (double
click on ‘sum of type’) … OK
Click FINISH

Sum of C_Gender Type


C_Gender 1 2 Grand Total
1 41 22 63
2 19 18 37
Grand Total 60 40 100

Q7 What is P( M )? 0.63

Q8 What is P( U )? 0.6

Q9 What is P( M or U )? 0.81

Q10 What is P( M and U )? 0.18

Q11 What is P( M | U )? 0.41/0.6 = 0.683


Introduction to Hypothesis Testing
Q12

Pg 323 K&W 5th ed

Excel gives us the following population figures on the logannualsal variable: (Please note that this
table was calculated using the entire dataset)

logannualsal

Mean 9.322461
Standard Error 0.008055
Median 9.357207
Mode 8.188689
Standard Deviation 1.162719

From the sample we calculate the following values:

logannualsal

Mean 9.334814
Standard Error 0.126136
Median 9.287301
Mode 9.798127
Standard Deviation 1.261362

TEST:

Tools… > Data Analysis CC… > One Sample Inference > Inference about a Mean (Sigma known)…

H0: μ = 9.32 (the population mean is equal to 9.32)

H1: μ ≠ 9.32 (the population mean is not equal to 9.32)

Test Statistic:

x−μ 9.3348 − 9.32


z= = = 0.1276
σ 1.16
n 100
From Excel:

One Sample z-test

H0: Mu = 9.32
H1: Mu <> 9.32

Sample Mean 9.3348


Sigma 1.16
Z-statistic 0.1277
Sample Size 100
p-value 0.8984

95% Confidence Interval


(9.1075,9.5622)

0.8984 > 0.05, hence we cannot reject the null hypothesis at the 5% significance level.

Q12 WebCT 7: 0.8984

Q13

WebCT 9: Type II

Q14

One Sample z-test

H0: Mu = 9.1
H1: Mu <> 9.1

Sample Mean 9.3348


Sigma 1.16
Z-statistic 2.0243
Sample Size 100
p-value 0.0429

WebCT 10: 0.0429

Q15

0.0429 < 0.05, Now we can reject the null hypothesis at the 5% significance level and conclude that
there is sufficient evidence to infer that the population mean is not equal to 9.1

WebCT 11: Yes

Q16

WebCT 12 : Type I
Inference about the Description of a Single Population
Q17

Pg 356 K&W 5th ed

Tools… > Data Analysis CC… > One Sample Inference > Inference about a Mean (Sigma
unknown)…

H0: μ = 40 (the population mean is equal to 40)


H1: μ ≠ 40 (the population mean is not equal to 40)

Test Statistic:
x − μ 38.98 − 40
t= = = -0.8816
s 11.5697
n 100

From Excel:

One Sample t-test

H0: Mu = 40
H1: Mu <> 40

Sample Mean 38.98


Sample Standard Deviation 11.56971
t-statistic -0.8816
Sample Size 100
p-value 0.3801

95% Confidence Interval


(36.6843,41.2757)

0.3801 > 0.05, hence we cannot reject the null hypothesis at the 5% significance level. There is
insufficient evidence to infer that the population mean is not equal to 40.

WebCT 13: 0.3801


Q18

Tools… > Data Analysis CC… > One Sample Inference > Inference about a Mean (Sigma
unknown)…

H0: μ = 40 (the population mean is equal to 40)

H1: μ < 40 (the population mean is less than 40)

From Excel:

One Sample t-test

H0: Mu = 40
H1: Mu < 40

Sample Mean 38.98


Sample Standard Deviation 11.56971
t-statistic -0.8816
Sample Size 100
p-value 0.1901

0.1901 > 0.05, hence we cannot reject the null hypothesis at the 5% significance level. There is
insufficient evidence to infer that the population mean is less than 40.

WebCT 14: 0.1901

Q19

Pg 359 K&W 5th ed

0.95 Confidence Interval Estimate of MU (SIGMA Unknown)

Sample mean = 38.98


Sample standard deviation = 11.5697
Lower confidence limit = 36.6843
Upper confidence limit = 41.2757

WebCT 15: 41.2757


Additional tests for Qualitative Data

Q20

Pg 548 K&W 5th ed

Expected Values based on previous year’s figures:

Number of observations Proportion Expected Value


100 0.75 75
100 0.12 12
100 0.08 8
100 0.05 5

Observed Values from sample data:

Observed Values
79
10
9
2

H0: p1 = 0.75, p2 = 0.12, p3 = 0.08, p4 = 0.05

H1: at least one of the proportions is not equal to its specified value

Test Statistic

Race Obs. Freq. ( f i ) Expected Freq. ( ei ) ( f i - ei ) ( f i − ei ) 2


ei
Black/African 79 75 4 0.2133
Coloured 10 12 -2 0.3333
White 9 8 1 0.1250
Indian/Asian 2 5 -3 1.8000
TOTAL 100 100 2.4713

χ 2 = 2.4713
THE IMPORTANT BIT!!!
Set up a table as below & use CHITEST.

From Excel: Using the CHITEST function in Excel, the following p-value can be calculated:

Obs Exp

79 75
10 12
9 8
2 5

P-value 0.480433

Rejection Region

χ 2 > χ α2 ,k −1 = χ 02.05,3 = 7.815

Conclusion

2.4713 < 7.815, hence we cannot reject the null hypothesis at the 5% significance level. The
proportions are a good fit for the data.

WebCT 23: 0.4804

Anda mungkin juga menyukai