EDRS 811 Quantitative Methods in Educational Research Assignments George Mason University Spring 2008

2 Way ANOVA Corrected

2 way ANOVA

Calculate          MS_w              and     MS_B 

MS_{B   =}   

MS_{B   =   =}

  MS_{w =}           MS_{w =}

= 6         = 30              = 12            = 16                   k =3

s=         s²  =

25 + 9 + 36 + 4 =

= o + 256 + 100 + 36 =

 100 + 64 + 64+ 100 = 

MS_{w =}

MS_{B   =   =}

                                     Descriptive Statistics

Dependent Variable: Y

                                                      Tests of Between-Subjects Effects

Dependent Variable: Y

a  R Squared = .611 (Adjusted R Squared = .525)

                                                                               Multiple Comparisons

Dependent Variable: Y

Tukey HSD

Based on observed means.

*  The mean difference is significant at the .05 level.

                                                           Y

Tukey HSD

Means for groups in homogeneous subsets are displayed.

 Based on Type III Sum of Squares

 The error term is Mean Square(Error) = 88.222.

a  Uses Harmonic Mean Sample Size = 4.000.

b  Alpha = .05.

2 way ANOVA

Calculate          MS_w              and     MS_B 

MS_{B   =}

MS_{w =}

Corrolation

Variables Entered/Removed(b)                                                                                                                                                               2

a  All requested variables entered.

b  Dependent Variable: Mental health

                                             Model Summary(c)

a  Predictors: (Constant), Physical functioning, Vitality, Social functioning

b  Predictors: (Constant), Physical functioning, Vitality, Social functioning, Body pain, Health perception

c  Dependent Variable: Mental health

                                                                                ANOVA(c)

a  Predictors: (Constant), Physical functioning, Vitality, Social functioning

b  Predictors: (Constant), Physical functioning, Vitality, Social functioning, Body pain, Health perception

c  Dependent Variable: Mental health

                                                                                   Coefficients(a)

a  Dependent Variable: Mental health

                                                                                                                                                                                                                   2

                                                                            Excluded Variables(b)

a  Predictors in the Model: (Constant), Physical functioning, Vitality, Social functioning

b  Dependent Variable: Mental health

                                                                       Residuals Statistics(a)

a  Dependent Variable: Mental health

Description This time I did mental health and recognized from other homework that the significance for body pain and health perception were not significant

I put in the first 4 variables 1^st then chose next, then put in the 2 that were not shown to be significant

Regression

Prediction of Current Salary from Beginning Salary.

    ^                                   ^                                                  ^

       Y= (3910 x 20,000) – 18331      Y = $78,181,669.00

Homework for 1-12-08                                               

 1.      We will test the hypothesis that the 0.05 level.

The Null Hypothesis is that “There is no difference in beginning salaries when comparing people in the non-minority group with people who are in a minority group.”

H_o: μ 1= μ2

We will seek the alternative hypothesis that: “The mean beginning salary for the non-minority group is greater than the mean for the beginning salary minority group.”

H_a: μ 1>μ2

The ρ value, is zero which is smaller than 0.05 level.  The results show that the population variances are not equal so we reject the Null Hypothesis. Further, the results also show that there is a statistically significant difference in the salaries between non-minority and minority groups.       t(281)=5.0, ρ = 0.000

There is 95 % confidence that the mean is greater than 1700 because all numbers are positive and the lower number is 1700. Therefore:

μ₁ > μ₂  by at least 1700, but, not more than 4287.

In addition, the 95% confidence interval of the difference shows that non-minority beginning salaries exceed minority beginning salaries by more than $1701 and possibly up to $4287.

The results support the claim that there is a salary bias in favor of the non-minority employees.

ES see the SPSS page.

2.      We will test the hypothesis that the 0.05 level.

The Null Hypothesis is that “There is no gender bias in current salaries.”

H_o: μ 1= μ2

The Alternative Hypothesis is that: “The mean current salaries for the male employees is greater than the mean current salaries for the female employees.”

H_a: μ 1>μ2

Since the ρ value, is zero, the results show that the population variances are not equal, so we reject the Null Hypothesis. Further, the results also show that there is a statistically significant difference in the current salaries between males and females.  

t(344) =12,   ρ = 0.000

There is 95 % confidence that the mean is greater than 12817 because all numbers are positive and the lower number is 12817. Therefore:

μ₁ > μ₂  by at least 12817  but not more than 18003.

In addition, the 95% confidence interval of the difference shows that male current salaries exceed female current salaries by more than $12817 and possibly up to $18003.

The results support the claim that there is a gender bias for current salaries in favor of the male employees.

Effect Size -                         

2-6-08

Homework for 1-12-08                                                Susan Kenney

3.      We will test the hypothesis that the 0.05 level.

The Null Hypothesis is that “There is no gender bias in beginning salaries.”

H_o: μ 1= μ2

The alternative hypothesis is that: “The mean beginning salaries for male employees is greater than the mean beginning salaries for the female employees.”

H_a: μ 1>μ2

Since the ρ value is zero, the results show that the population variances are not equal, so we reject the Null Hypothesis. Further, the results also show that there is a statistically significant difference in the salaries between males and females.

t(319)=12,   ρ = 0.000

There is 95 % confidence that the mean is greater than 6026 because all numbers are positive and the lower number is 6026 Therefore:

μ₁ > μ₂  by at least 6026 but not more than 8393

In addition, the 95% confidence interval of the difference shows that male beginning salaries exceed female beginning salaries by more than $6026 and possibly up to $8393.

The results support the claim that there is a gender bias in beginning salaries in favor of the male employees.

ES see the SPSS page.

 Back to EDRS 811

Back to Courses

More Courses

Home