BA 275
Fall 2006
Statgraphics
Instruction 2
Statgraphics Instruction 2
Confidence Interval Estimation
1. To estimate  (population mean):
A. With Raw Data (helpful in analyzing the projects):
Describe / Numeric Data / One-Variable Analysis. Enter the variable (e.g., SALARY) to
be analyzed in the Data box of the pop-up dialog box. Click OK. Click the Tabular
Options icon and check Confidence Intervals.
B. With Summary Statistics (helpful in doing your practice problems):
Describe / Numeric Data / Hypothesis Tests... . Click on Normal Mean and fill in Sample
Mean, Sample Sigma (i.e., sample standard deviation) and Sample Size. Ignore the Null
Hypothesis box.
2. To estimate p (population proportion):
A. With Raw Data (helpful in analyzing the projects):
Step 1. Describe / Categorical Data / Tabulation. Enter the categorical variable (e.g.
Gender) to be analyzed in the Data box of the pop-up dialog box. Click OK. Write down
the Frequency (n) and Relative Frequency ( p̂ ) of the class of interest (e.g., Female CEOs)
from the Frequency Table.
Step 2. Describe / Hypothesis Tests... . Click on Binomial Proportion and fill in Sample
Proportion (i.e., p̂ ) and Sample Size (n). Ignore the Null Hypothesis box.
B. With Summary Statistics (helpful in doing your practice problems):
Follow Step 2. in 2-A.
3. To estimate the difference between two population means, 1 – 2:
A. With Raw Data (helpful in analyzing the projects):
Compare / Two Samples / Independent Samples. Change the Input option in the pop-up
dialog box if necessary. In our CEO example, we need to select the Data and Code
Columns option. Enter the numerical variable of interest (e.g., SALARY) in the Data box
and the code variable (e..g, GENDER) in the Sample Code box. Click OK.
B. With Summary Statistics (helpful in doing your practice problems):
Compare / Two Samples / Hypothesis Tests.... Click on Normal Means and fill in Sample
Mean, Sample Sigma and Sample Size for each sample. Ignore the Null Hypothesis box.
4. To estimate the difference between two population proportions, p1 – p2:
A. With Raw Data (helpful in analyzing the projects):
Step 1. Describe / Categorical Data / Crosstabulation. Enter the categorical variable (e.g.
Gender) to be analyzed in the Data box of the pop-up dialog box. Click OK. Write down
the Frequencies (n1 and n2) and Relative Frequencies ( p̂1 and p̂ 2 ) of the two classes of
interest (e.g., Female and Male CEOs) from the Frequency Table.
Step 2. Compare / Two Samples / Hypothesis Tests.... Click on Binomial Proportions
and fill in Sample Proportions ( p̂1 and p̂ 2 ) and Sample Sizes (n1 and n2). Ignore the Null
Hypothesis box.
B. With Summary Statistics (helpful in doing your practice problems):
Follow Step 2 in 4-A.
Hsieh, P.-H.
1
BA 275
Fall 2006
Statgraphics
Instruction 2
Confidence Interval Estimation: CEO Data
Means and 95.0 Percent LSD Intervals
500
SALARY
400
300
200
100
0
Bachelors Doctorate Masters
None
EDUCATION
Hsieh, P.-H.
2
BA 275
Fall 2006
Statgraphics
Instruction 2
Hypothesis Testing
1. To test H0:  = 0 (population mean):
A. With Raw Data (helpful in analyzing the projects):
Describe / Numeric Data / One-Variable Analysis. Enter the variable (e.g., SALARY) to
be analyzed in the Data box of the pop-up dialog box. Click OK. Click the Tabular
Options icon and check Hypothesis Tests. Right click to change Ha and .
B. With Summary Statistics (helpful in doing your practice problems):
Describe / Numeric Data / Hypothesis Tests... . Click on Normal Mean and fill in Sample
Mean, Sample Sigma (i.e., sample standard deviation) and Sample Size. Enter 0 in the Null
Hypothesis box.
2. To test H0: p = p0 (population proportion):
A. With Raw Data (helpful in analyzing the projects):
Step 1. Describe / Categorical Data / Tabulation. Enter the categorical variable (e.g.
Gender) to be analyzed in the Data box of the pop-up dialog box. Click OK. Write down
the Frequency (n) and Relative Frequency ( p̂ ) of the class of interest (e.g., Female CEOs)
from the Frequency Table.
Step 2. Describe / Hypothesis Tests... . Click on Binomial Proportion and fill in Sample
Proportion (i.e., p̂ ) and Sample Size (n). Enter p0 in the Null Hypothesis box.
B. With Summary Statistics (helpful in doing your practice problems):
Follow Step 2. in 2-A.
3. To test H0: 1 – 2 = D0:
A. With Raw Data (helpful in analyzing the projects):
Compare / Two Samples / Independent Samples. Change the Input option in the pop-up
dialog box if necessary. In our CEO example, we need to select the Data and Code
Columns option. Enter the numerical variable of interest (e.g., SALARY) in the Data box
and the code variable (e..g, GENDER) in the Sample Code box. Click OK.
B. With Summary Statistics (helpful in doing your practice problems):
Compare / Two Samples / Hypothesis Tests.... Click on Normal Means and fill in Sample
Mean, Sample Sigma and Sample Size for each sample. Enter D0 in the Null Hypothesis box.
4. To test H0: p1 – p2 = D0:
A. With Raw Data (helpful in analyzing the projects):
Step 1. Describe / Categorical Data / Crosstabulation. Enter the categorical variable (e.g.
Gender) to be analyzed in the Data box of the pop-up dialog box. Click OK. Write down
the Frequencies (n1 and n2) and Relative Frequencies ( p̂1 and p̂ 2 ) of the two classes of
interest (e.g., Female and Male CEOs) from the Frequency Table.
Step 2. Compare / Two Samples / Hypothesis Tests.... Click on Binomial Proportions
and fill in Sample Proportions ( p̂1 and p̂ 2 ) and Sample Sizes (n1 and n2). Enter D0 in the
Null Hypothesis box.
B. With Summary Statistics (helpful in doing your practice problems):
Follow Step 2 in 4-A.
Hsieh, P.-H.
3
BA 275
Fall 2006
Statgraphics
Instruction 2
Hypothesis Testing: CEO Data
Hsieh, P.-H.
4