1. Make a frequency table about Educational level.

Requirements: (i)
number of observations for each level and (ii) specific percentage for each
level. Which level of education accounts for the highest percentage?
Which level of education accounts for the lowest percentage?

Educational Level
Cumulative
Frequency Percent Valid Percent Percent
Valid High School 23 15.2 15.2 15.2
Bachelor 88 58.3 58.3 73.5
Master 26 17.2 17.2 90.7
Doctor 14 9.3 9.3 100.0
Total 151 100.0 100.0

Highest percentage: bachelor

Lowest percentage: Doctor

2. Draw a pie chart showing the percentage of observations by Gender

(male, female). Show the specific percentages on the graph. In the sample,
males or females accounted for a higher proportion?

It is clearly that the percentage of female is 66,33%, which is higher than the proportion of male, at
about above 44%.

3. Compare the mean of Income of the 2 groups of gender. Is there any

statistically significant difference between the 2 groups of gender in terms
of Income? ( T-test)

1
 Independent Samples Test
Levene's Test
for Equality of
Variances t-test for Equality of Means
95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper
Monthly Equal 1.063 .304 - 149 .197 -1.968 1.519 -4.970 1.034
Income variances 1.295
assumed
Equal - 129.222 .206 -1.968 1.548 -5.032 1.095
variances 1.271
not assumed

B1: Levene Test:

If Sig value in Levene test <0.05=> the variance are different => ko cần làm ttest
If Sig value in Levene test >0.05=> the variance are not different => làm t test
In example: Sig 0.304>0.05 => use the T-test
B2: T-test
Mean < 0.05 => There is a significance difference in the means of variable (A) between group
Mean > 0.05 => There is no significance difference in the means of variable (A) between group
In example: 0.197> 0.05 => so it is conclude that there is no significant difference in means of
the variable “income” between 2 group (male and female)
4. Compare the mean of Income among different educational levels. Is there
any statistically significant difference among educational levels in terms
of Income? (ANOVA)
Cách làm: Analyze -> compare means -> one way -> overall sat (dependent) and status la factor -> click
chon option -> descriptive + homo + welch

2
 Test of Homogeneity of Variances
Levene Statistic df1 df2 Sig.
Monthly Income Based on Mean 4.824 3 147 .003
Based on Median 3.383 3 147 .020
Based on Median and with 3.383 3 116.718 .021
adjusted df
Based on trimmed mean 4.122 3 147 .008

ANOVA
Monthly Income
Sum of Squares df Mean Square F Sig.
Between Groups 3724.315 3 1241.438 19.751 .000
Within Groups 9239.473 147 62.854
Total 12963.788 150

Robust Tests of Equality of Means

Monthly Income
Statistica df1 df2 Sig.
Welch 59.439 3 42.811 .000
a. Asymptotically F distributed.

B1: Levene Test:

0.003<0.005=> dùng robust tests
B2:
Xài robust test sig=0.000<0.005=> mean hai nhóm khác nhau
The model has an F-value of 9.963, which yields a p-value of 0.00 (less than
0.05), suggesting that at least two of the three groups differ significantly with
regard to the mean of overall price/performance satisfaction. (Có sự khác biệt có
ý nghĩa thống kê giữa ít nhất 2 nhóm khách hàng về giá trị trung bình của sự hài
lòng về giá.)

3
 5. Check whether there is multicollinearity among the variables: Age,
Gender, Education, Marital status, Doing exercises?
Multicollinearity (slide 100 regression)
Analyze-correlations- bivariate- file colleges and universities- 4 cái đầu

Correlations
Educational
Age Gender Level Marital Status Excercise
** **
Age Pearson Correlation 1 .012 .606 .666 .102
Sig. (2-tailed) .884 .000 .000 .215
N 151 151 151 151 151
Gender Pearson Correlation .012 1 .070 .007 .289**
Sig. (2-tailed) .884 .393 .929 .000
N 151 151 151 151 151
Educational Level Pearson Correlation .606** .070 1 .594** .036
Sig. (2-tailed) .000 .393 .000 .661
N 151 151 151 151 151
** **
Marital Status Pearson Correlation .666 .007 .594 1 .057
Sig. (2-tailed) .000 .929 .000 .487
N 151 151 151 151 151
**
Excercise Pearson Correlation .102 .289 .036 .057 1
Sig. (2-tailed) .215 .000 .661 .487
N 151 151 151 151 151
**. Correlation is significant at the 0.01 level (2-tailed).

4
Because no exceed 0.7=> there is no problem multicollinearity
6. Use multiple linear regression to analyze the impact of the variables Age,
Gender, Education, Marital status, Doing exercises on the variable
Income?
Model Summary
Adjusted R Std. Error of the
Model R R Square Square Estimate
a
1 .658 .432 .417 7.099
a. Predictors: (Constant), Excercise, Marital Status, Gender, Age

ANOVAa
Model Sum of Squares df Mean Square F Sig.
1 Regression 5606.676 4 1401.669 27.816 .000b
Residual 7357.112 146 50.391
Total 12963.788 150
a. Dependent Variable: Monthly Income
b. Predictors: (Constant), Excercise, Marital Status, Gender, Age

Coefficientsa
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 2.007 1.564 1.283 .201
Age .176 .058 .253 3.012 .003
Gender 1.333 1.215 .071 1.097 .274
Marital Status 8.674 1.654 .438 5.244 .000
Excercise 1.784 1.214 .096 1.470 .144
a. Dependent Variable: Monthly Income
COEFFICIENTS TABLE

ANALYZE->REGRESSION->LINEAR -> GRAD (DEPEND) CÒN LẠI LÀ INDEPEND -> OK

Used to test the overall significance of a regression model

Step 1: Sig (anova) <0.005 => The overall model is statistically significant at the 5% significance level

Step 2: Coefficent sign (= 0.176) is positive so age has positive impact on income

5
If age increase 1% then graduation will increase by 0.176%, keeping all the factors
unchanged
The sig of age =0.03<0.05 và hệ số hồi quy mang dấu (+) => positive impact

Step 1: Sig.compared to 0.05

If sig. < 0.05 -> The overall model is statistically significant at the 5% significance level (có ý nghĩa
thống kê và phù hợp)

If sig. > 0.05 -> The overall model is not statistically significant at the 5% significance level (mô hình
không có ý nghĩa thống kê và không phù hợp

Step 2: Is the impact positive or negative?

If the coefficient sign is + : X has a positive impact on Y. For example, SAT
scores have a positive impact on graduation rates
If the coefficient sign is -: X has negative /negative impact on Y
Notes:
The signs of Understandardized and Standardized coefficients are always the
same
Understandardized Coeffients: Coeffocients that still retain the original units
of variable
Standardized Coefficients: Coefficients that have been standardized, no longer
retain the original units of the variable
To explain the economic meaning, Unstandardized Coefficients will be used

7. Create an interaction variable between Age and Doing exercises. Analyze the
moderating effect of Doing exercises on the relationship between Age and
Income?

Step 2: Is the impact positive or negative?

6
 If the coefficient sign is + : X has a positive impact on Y. For example, SAT
scores have a positive impact on graduation rates
If the coefficient sign is -: X has negative /negative impact on Y
Notes:
The signs of Understandardized and Standardized coefficients are always the
same
Understandardized Coeffients: Coeffocients that still retain the original units
of variable
Standardized Coefficients: Coefficients that have been standardized, no longer
retain the original units of the variable
To explain the economic meaning, Unstandardized Coefficients will be used

Coefficientsa
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 893.588 1824.575 .490 .628
Age 1044.146 42.141 .975 24.777 .000
MBA 14767.232 1351.802 .430 10.924 .000
a. Dependent Variable: Salary

7
 Coefficientsa
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 3902.509 1336.398 2.920 .006
Age 971.309 31.069 .907 31.263 .000
MBA -2971.080 3026.242 -.086 -.982 .334
Age_MBA 501.848 81.552 .531 6.154 .000
a. Dependent Variable: Salary

Note:

1. When regressing with interaction variables, it is necessary to include component variables

(age, mba)
2. When regressing with interaction variables, do not read the results of components variables
(age,mba)