Upload
syed-asfar-ali-kazmi
View
111
Download
2
Tags:
Embed Size (px)
DESCRIPTION
QTIA Report 2012 IQRA UNIVERSITYDedicated to my respected teachers
Citation preview
IQRA UNIVERSITY (IU)
Quantitative Techniques In Analysis
Report of about
Tests of Regression Analysis and Differences Between Groups
Submitted To: Syed Hammad Ali
Submitted By: Syed Asfar Ali Kazmi (14807)
Date of Submitted: 13-05-2012
CONTENTS:
Other Test
1. Test of Normality ………………………… 3 - 5
2. Test of linearity ………………………… 6 – 7
Tests of Regression Analysis
3. Linear Regression ………………………… 8 - 11
4. Multiple Regression ………………………… 12 – 15
Differences Between Groups
5. One Sample T Test ………………………… 16 – 18
6. Independent Sample T Test ………………. 18 – 21
7. Paired Sample T Test ………………………… 22 – 23
8. One-Way ANOVA ………………………… 23 – 25
9. Two-Way ANOVA …………………………. 25 – 27
10. SPANOVA ………………………… 27 – 29
11. MANOVA …………………………. 30 - 40
OTHER TESTS
Test of Normality:
Normality tests are used to determine whether a data set is well-modeled by a normal distribution or not, or to compute how likely an underlying random variable is to be normally distributed.
Question:Is the variable ‘’ AGE OF RESPONDENT ’’ of GSS2000R.sav is normally distributed or not?
Hypothesis:
HO: The null hypothesis states that the data is normally distributedHA: The alternative hypothesis states that the data is not normally distributed
Interpretation:Case Processing Summary
Cases
Valid Missing Total
N Percent N Percent N Percent
AGE OF RESPONDENT 270 100.0% 0 .0% 270 100.0%
This Table show us that sample size caese N=270, there is no missing cases and the data
is 100% valid.
Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
AGE OF RESPONDENT .083 270 .000 .957 270 .000
a. Lilliefors Significance Correction
Decision:
The test statistics are shown in the third table. Here two tests for normality are run. In our case, since we have N= 270 elements sample size,for AGE OF RESPONDENT data set which is greather than 51 elements, we use the Kolmogorov-Smirnov test .From A, the p-value is 0.000 is less than 0.05.
So we can reject the null hypothesis and conclude the sample is not normally distributed.
The histogram plot indicates the shape of the distributed data have a “ bell” shape These data are clearly Not normally distributed.
From this graph we can conclude that the data appears is not to be normally distributed as it follows the diagonal line closely and does appear to have a non-linear pattern.
Test of Linearity:
Linearity means that the amount of change, or rate of change, between scores on two variables are constant for the entire range of scores for the variables.when one variable( X ) increases and the other variable(Y) is also be increased in the same way.we check through linearity test that there relationships are linear not linear.
Question:
hypothesis test of the correlation coefficient, the relationship between "total hours spent on the chat" and "total hours spent on the Internet" is not linear. However, the square transformation of the independent variable "total hours per spent on the Internet" does result in a relationship that is linear.
Interpretation:
The value of R2 Linear (0.656) suggests that the relationship between “total time spent on the internet” and “the total hours spent on chat” is strong. R= 0.8099
Tests for Regression Analysis
Regression analysis is used to measure the re la t ionship be tweent w o o r m o r e variables. One variable is called dependent (response, or outcome) variable and the other is called Independent (explanatory or predictor) variables.It is used to check that due to one unit change in the independent variable(s) how much change occurs in dependent variable.
Or
The use of regression to make quantitative predictions of one variable from the values of another variable is called regression analysis. There are following several types of regression, which may be used by the researcher.
Linear regression Multiple linear regression Quadratic / Curvilinear regression Logistic / Binary logistic regression Multivariate logistic regression
Linear Regression:
When one dependent variable depends on single independent variable then their dependency called linear regression it is a measure of how strongly the independent variable predicts the dependent variable and its model is given by
y = a + bx
Assumptions:
Variables are measured at the interval or ratio level (continuous). Variables are approximately normally distributed There is a linear relationship between the two variables.
Question:
Can we predict math achievement from grades in high school?
Variables:
D.V= Math achievement
I.V = grades in high school
Hypothesis:
H0: there is the relationship b/w V1 and V2HA: there is no relationship b/w V1 and V2
Interpretation:
Variables Entered/Removedb
Model Variables
Entered
Variables
Removed Method
dim
ensi
on0
1 grades in h.s.a . Enter
a. All requested variables entered.
b. Dependent Variable: math achievement test
The above table tells us about the independent variable and the regression method used.
Here we see that the independent variable i.e. grads in high school is entered for the
analysis as we selected the Enter method.
Model Summary
Model
R R Square
Adjusted R
Square
Std. Error of the
Estimate
dime
nsio
1 .504a .254 .244 5.80018
a. Predictors: (Constant), grades in h.s.
This table gives us the R-value, which represents the correlation between the
observed values and predicted values of the dependent variable. R-Square is called
the coefficient of determination and it gives the adequacy of the model. Here the
value of R-Square is 0.504 that means the independent variable in the model can
predict 50% of the variance in dependent variable. Adjusted R-Square gives the
more accurate information about the model fitness if one can further adjust the
model by his own.
ANOVAb
Model Sum of Squares df Mean Square F Sig.
1 Regression 836.606 1 836.606 24.868 .000a
Residual 2455.875 73 33.642
Total 3292.481 74
a. Predictors: (Constant), grades in h.s.
b. Dependent Variable: math achievement test
The above table gives the test results for the analysis of one-way ANOVA. The results are given in three rows. The first row labeled Regression gives the variability in the model due to known reasons. The second row labeled Residual gives the variability due to random error or unknown reasons. F-value in this case is 24.868 and the p-value is given by 0.000 which is less that 0.05, so we reject our null hypothesis and conclude that there is no relationship between math achievement and grades in high school.
Coefficientsa
Model
Unstandardized Coefficients
Standardized
Coefficients
t Sig.B Std. Error Beta
1 (Constant) .397 2.530 .157 .876
grades in h.s. 2.142 .430 .504 4.987 .000
a. Dependent Variable: math achievement test
The above table gives the regression constant and coefficient and their significance. These regression coefficient and constant can be used to construct an ordinary least squares (OLS) equation and also to test the hypothesis of the independent variable. Using the regression coefficient and the constant term given under the column labeled B; one can construct the OLS equation for predicting the math achievement i.e.
Math achievement = .397 + (2.1242) (grades in h.s)
Multiple Regression (Hierarchical Method)
Multiple regression is the most commonly used technique to assess the relationship between one dependent variable and several independent variables. There are three major types of multiple regression i.e.
Simultaneous regression. Hierarchical or Sequential regression. Stepwise or statistical regression.
Assumptions:
1) Dependent variables should be scale.2) The relationship between the predictor variable and the dependent variable in linear.3) The error/residual4) Multicollinearity should not be exist5) Homogenity should be exist
Question :
How well can we predict current salary from a combination of three variables Beginning salary, Educational Level, and Month since hire?
Variables:
Dependent variable = Current Salary
Indendepent Variable = (1) Beginning salary
(2) Educational Level
(3) Month since hire
Interpretation:
Descriptive Statistics
Mean Std. Deviation N
Current Salary $34,419.57 $17,075.661 474
Beginning Salary $17,016.09 $7,870.638 474
Months since Hire 81.11 10.061 474
Educational Level (years) 13.49 2.885 474
Correlations
Current Salary
Beginning
Salary
Months since
Hire
Educational
Level (years)
Pearson Correlation Current Salary 1.000 .880 .084 .661
Beginning Salary .880 1.000 -.020 .633
Months since Hire .084 -.020 1.000 .047
Educational Level (years) .661 .633 .047 1.000
Sig. (1-tailed) Current Salary . .000 .034 .000
Beginning Salary .000 . .334 .000
Months since Hire .034 .334 . .152
Educational Level (years) .000 .000 .152 .
N Current Salary 474 474 474 474
Beginning Salary 474 474 474 474
Months since Hire 474 474 474 474
Educational Level (years) 474 474 474 474
Variables Entered/Removedb
Model Variables
Entered
Variables
Removed Method
dim
ensi
on0
1 Educational
Level (years),
Months since
Hire, Beginning
Salarya
. Enter
a. All requested variables entered.
b. Dependent Variable: Current Salary
Model Summary
Model
R R Square
Adjusted R
Square
Std. Error of the
Estimate
dim
ensi
on0
1 .895a .801 .800 $7,645.998
a. Predictors: (Constant), Educational Level (years), Months since Hire,
Beginning Salary
ANOVAb
Model Sum of Squares df Mean Square F Sig.
1 Regression 1.104E11 3 3.681E10 629.703 .000a
Residual 2.748E10 470 5.846E7
Total 1.379E11 473
a. Predictors: (Constant), Educational Level (years), Months since Hire, Beginning Salary
b. Dependent Variable: Current Salary
Coefficientsa
Model Unstandardized
Coefficients
Standardized
Coefficients
t Sig.
Collinearity
Statistics
B Std. Error Beta Tolerance VIF
1 (Constant) -19986.502 3236.616 -6.175 .000
Beginning Salary 1.689 .058 .779 29.209 .000 .597 1.676
Months since Hire 155.701 35.055 .092 4.442 .000 .994 1.006
Educational Level
(years)
966.107 157.924 .163 6.118 .000 .595 1.679
a. Dependent Variable: Current Salary
Collinearity Diagnosticsa
Model Dimension
Eigenvalue
Condition
Index
Variance Proportions
(Constant)
Beginning
Salary
Months since
Hire
Educational
Level (years)
dimension0
1
dimension1
1 3.847 1.000 .00 .01 .00 .00
2 .125 5.542 .01 .57 .02 .00
3 .020 13.734 .02 .42 .12 .92
4 .007 23.392 .97 .00 .86 .08
a. Dependent Variable: Current Salary
Result:Simultaneously multiple regression was conducted to investigate the best predictors of current salary. The means, standard deviation and inter correlations can be found in table. The combination of variables to predict current salary from Beginning salary,Educational Level ,& Month since hire was statistically significant, F =629.703., p<0.05 . The be ta coef f ic ien ts a re presented in las t t ab le . Note tha t a l l indenpent var iab les Beginning salary,Educational Level ,& Month since hire are significantly predicts on current salary when all three variables are included. The adjusted R2 value was 0.800. This indicatesthat 80 % of the variance in current salary is a large effect.
( Differences Between Groups )
T-TEST Statistics:
The t test is used to compare to groups to answer the differential research questions. Its values determines the difference by comparing means.
Hypothesis for T-test:
HO: there is no difference between variable 1 and variable 2 (Accept when the significant value is greater than 0.05)
H1:There is difference between variable 1 and variable 2 (Accept when the significant value is less than 0.05)
Types of T-testThere are three types of T-tests.
1) One sample t-test.2) Independent sample t-test.3) Paired sample t-test
1) ONE SAMPLE T-TEST:
One sample t-test is used to determine if there is difference between population mean(Test value) and the sample mean (X)
Assumptions and conditions of sample t-test:1.The dependent variable should be normally distributed within the population2.The data are independent.(scores of one participant are not depend on scores of the other:participant are independent of one another )
Question:
Is the average salary of employee in the ( employee data.sav ) is equal or more than $ 30000 ( per month ) in US?
The hypotheses are:
1)The null hypothesis states that the average salary of the employee is equal to “30000”
H0: 30000
2) The alternative hypothesis states that the average salary of the employee is not equal to “30000”
HA: 30000
Interpretation:0
One-Sample Statistics
N Mean Std. Deviation Std. Error Mean
Current Salary 474 $34,419.57 $17,075.661 $784.311
In above table “N” shows the total number of observation.(sample Size N=474 employees) The
average salary of total employees is “34,419.57”.
The standard deviation of the data is “17,075.661”and the standard error of the mean is
“784.311”.
One-Sample Test
Test Value = 30000
T df Sig. (2-tailed) Mean Difference
95% Confidence Interval of the
Difference
Lower Upper
Current Salary 5.635 473 .000 $4,419.568 $2,878.40 $5,960.73
Through above table we can observe that,
i. “T” value(5.635) is positive which show that our estimated mean value is less than actual value of mean.
ii. Degree of freedom is (N – 1) = 473.
iii. `````The “P-value” is “0.000” which is less than “0.05”.
iv. The difference between the estimated & actual mean is “4,419.568”.
v. Confidence interval has the lower & upper limit 2,878.40 & 5,960.73respectively. The confidence interval limits does not contains zero.
Decision:-On the basis of following observation I reject my “Null hypothesis” and accept the “Alternative hypothesis”. I am almost “100%” sure on my decision.
i. The “P-value” is “0.000” which is less than “0.05”.ii. The confidence interval limits does not contains zero.
Therefore the average salary of employees is not equal to “30000”.
2) INDEPENDENT SAMPLE T-TEST:
Independent sample T-test is used to compare two independent groups (Male and Female) with respect to there effect on same dependent variable.
Assumptions and conditions of Independent T-test:
1.Variance of the dependent variable for two categories of the independent variableshould be equal to each other.
2.Dependent variable should be scale
3.Data on dependent variable should be independent.
Question:
Are the mean differences average salaries of male & female employees differ significantly or equal to their current salary in US?
( employee data.sav)
The hypotheses are:
1) The null hypothesis states that the average salary of the male employee is equal to average salary of the female employee.
H0 :
2) The alternative hypothesis states that the average salary of the male employee is not equal to average salary of the female employee.
HA :
Interpretation:-
Group Statistics
Gender N Mean Std. Deviation Std. Error Mean
Current Salary Female 216 $26,031.92 $7,558.021 $514.258
Male 258 $41,441.78 $19,499.214 $1,213.968
Through above table we can observe that,
i. Total number of male is “258” and the female is “216”.ii. The mean value of salaries of male employee is 41,441.78 & the female employee is
26,031.92.
iii. Standard deviation of salaries of male employee is 19,449.214 & the female employee is 7,558.021.
iv. Standard error of mean of salaries of male employees is 1,213.968 & the Standard error of mean of salaries of female employees is 514.258.
Independent Samples Test
Levene's Test
for Equality of
Variances t-test for Equality of Means
F Sig. t df
Sig.
(2-
tailed)
Mean
Difference
Std. Error
Difference
95% Confidence Interval
of the Difference
Lower Upper
Current
Salary
Equal
variance
s
assumed
119.669 .000 10.94
5
472 .000 $15,409.86
2
$1,407.90
6
$12,643.32
2
$18,176.40
1
Equal
variance
s not
assumed
11.68
8
344.26
2
.000 $15,409.86
2
$1,318.40
0
$12,816.72
8
$18,002.99
6
Interpretation:-
In above table we have two parts (a) f-test, (b) t-test, through which we can observe that,
i. “F” value is “119.669” with significant value of “0.00” which is less than “0.05”.ii. On the basis of P-value of F-test part we assume that the variance of the two populations
is not equal. iii. “T” value is positive which show that the mean value of salaries of male employees is
greater than the mean value of salaries of female employeesiv. Degree of freedom is “344.262”.v. The “P-value” is “0.000” which is less than “0.05”.
vi. The difference between the two population mean is “15,409.862”.vii. The standard error difference between the two population mean is “1,318.400”.
viii. Confidence interval has the lower & upper limit “12,816.728” & “18,002.996” respectively. The confidence interval limits does not contains zero.
Decision:-
On the basis of following observation I reject my “Null hypothesis” and accept the “Alternative hypothesis”. I am almost “100%” sure on my decision.
i. The “P-value” is “0.000” which is less than “0.05”.ii. The confidence interval limits does not contains zero.
The average salaries of male & female employees are not equal.
3) Paired t-test : A paired (samples) t-test is used when you have two related observations (i.e., two observations per subject) and you want to see if the means on these two normally distributed interval variables differ from one another.
Assumptions and conditions of Paired sample T-test :
1)The independent variable is dichotomous and its levels (or groups) are paired, or matched, in some way (husband-wife, pre-post etc)
2) The dependent variable is normally distributed in the two conditions
Question:
The mean difference of the two paired variables ( current and beginning salary) is significant or equal?
Interpretation:
Paired Samples Statistics
Mean N Std. Deviation Std. Error Mean
Pair 1 Current Salary $34,419.57 474 $17,075.661 $784.311
Beginning Salary $17,016.09 474 $7,870.638 $361.510
Through above table we can observe that,
i. The mean value of current & beginning salary is “34,419.57” & “17,016.09” respectively.
ii. Total number of both groups is “474” individually.
iii. The standard deviation of current & beginning salary is “17,075.661” & “7,870.638” respectively
iv. The standard error mean of current & beginning salary is “784.331” & “361.510” respectively
Paired Samples Correlations
N Correlation Sig.
Pair 1 Current Salary & Beginning
Salary
474 .880 .000
Through above table we can observe that,
i. The total number of pair is “474”.ii. “0.88” show that the both values of group are highly co-related, which indicate that the
employees who has greater begging salary has also greater current salary.
iii. The P-value is “0.00” which is less than “0.05”
Paired Samples Test
Paired Differences
t df
Sig. (2-
tailed)Mean
Std.
Deviation
Std. Error
Mean
95% Confidence Interval of
the Difference
Lower Upper
Pair
1
Current Salary -
Beginning
Salary
$17,403.481 $10,814.620 $496.732 $16,427.407 $18,379.555 35.036 473 .000
Interpretation:-
In above table we have two parts (a) f-test, (b) t-test, through which we can observe that,
i. The mean value of pair is “17,403.481”.ii. The standard deviation of pair is “10,814.620”.
iii. The standard error mean of pair is “496.732”.
iv. Confidence interval has the lower & upper limit “16,427.407” & “18,379.555” respectively. The confidence interval limits does not contains zero.
v. T- Value is “35.036”.
vi. Degree of freedom is (N-1) = “473”.
vii. P-vale is “0.00” which is less than “0.05”.
Decision:-
On the basis of following observation I reject my “Null hypothesis” and accept the “Alternative hypothesis”. I am almost “100%” sure on my decision.
iii. The “P-value” is “0.000” which is less than “0.05”.iv. The confidence interval limits does not contains zero.
The mean difference of the two paired variables i.e. current and beginning salary is
significant or not same.
One-way ANOVA:A one-way analysis of variance (ANOVA) is used when you have a categorical independent variable (with two or more categories) and a normally distributed interval dependent variable and you wish to test for differences in the means of the dependent variable broken down by the levels of the independent variable.
Assumptions:
Independent variable consists of two or more categorical independent groups.
Dependent variable is either interval or ratio (continuous).
Dependent variable is approximately normally distributed for each category of the independent variable.
Equality of variances between the independent groups (homogeneity of variances).
Observations are independent.
Question:
Are there difference among the ethnicity groups(Euro-American, African-American, Latino-American, Asian-American ) on competence scale?
Dependent variable : Competence scaleIndependent variable : Ethnicity groups
Interpretation:
Descriptives
competence scale
N Mean
Std.
Deviation Std. Error
95% Confidence Interval for
Mean
Minimum MaximumLower Bound Upper Bound
Euro-Amer 40 3.4000 .54243 .08577 3.2265 3.5735 1.75 4.00
African-
Amer
14 3.1786 .90101 .24080 2.6583 3.6988 1.00 4.00
Latino-Amer 10 2.8750 .72887 .23049 2.3536 3.3964 1.00 3.50
Asian-Amer 7 3.3214 .51467 .19453 2.8454 3.7974 2.50 4.00
Total 71 3.2746 .66300 .07868 3.1177 3.4316 1.00 4.00
The descriptives table provides us the descriptive statistics.
mean, standard deviation and 95% confidence intervals for the dependent variable
(competence scale ) for each separate groups of ethnicity ( Euro-American,
African-American, Latino-American, Asian-American ) as well as when all groups
are combined (Total ).
Test of Homogeneity of Variances
competence scale
Levene Statistic df1 df2 Sig.
1.323 3 67 .274
The assumption of equal variances (Homogeneity ) has been met,because the sig.
value =.274 is more than 0.05 .(we accept the Null hypothesis )
ANOVA
competence scale
Sum of Squares df Mean Square F Sig.
Between Groups 2.370 3 .790 1.864 .144
Within Groups 28.399 67 .424
Total 30.769 70
The above table gives the test results for the analysis of one-way ANOVA. The results are given in three rows. The first row labeled between groups gives the variability due to the different designations of the ethnicity groups (known reasons). The second row labeled within groups gives the variability due to random error (unknown reasons), and the third row gives the total variability. In this case, F-value is 1.864, and the corresponding p-value=0.144 is greather than 0.05. Therefore we accept the null hypothesis and conclude that the no difference among the ethnicity groups are not the same in all four categories
There is no difference between the ethnicity groups and the competence scale.
Two-way ANOVA:The two-way ANOVA compares the mean differences between groups that have been split on two independent variables (called factors). You need two independent, categorical variables and one continuous, dependent variable
Assumptions to use two-way ANOVA:
As with other parametric tests, we make the following assumptions when using two-way ANOVA:
Dependent variable is either interval or ratio (continuous).
Dependent variable is approximately normally distributed for each category of the independent variable.
The variances among populations must be equal (homogeneity).
Data are interval or nominal.
Interpretation:
Between-Subjects Factors
Value Label N
math grades 0 less A-B 43
1 most A-B 30
father's educ
revised
1.00 HS grad or
less
38
2.00 Some College 16
3.00 BS or More 19
This table is provide us the sample size N of our independent variables ( math grades and
Father’s educations revised groups ).
Descriptive Statistics
Dependent Variable:math achievement test
math grades father's educ revised Mean Std. Deviation N
dimension1
less A-B
dimension2
HS grad or less 9.8261 5.03708 23
Some College 12.8149 5.05553 9
BS or More 12.3636 7.18407 11
Total 11.1008 5.69068 43
most A-B
dimension2
HS grad or less 10.4889 6.56574 15
Some College 16.4284 3.43059 7
BS or More 21.8335 2.84518 8
Total 14.9000 7.00644 30
Total
dimension2
HS grad or less 10.0877 5.61297 38
Some College 14.3958 4.66544 16
BS or More 16.3509 7.40918 19
Total 12.6621 6.49659 73
This table is very useful as it provides the mean and standard deviation for the groups that have been split by both independent variables. In addition, the table also provides "Total" rows, which allows means and standard deviations for groups only split by one independent variable or none at all to be knownThere are six-cell means will be shown in the plot.
Levene's Test of Equality of Error Variancesa
Dependent Variable:math achievement test
F df1 df2 Sig.
2.548 5 67 .036
Tests the null hypothesis that the error variance
of the dependent variable is equal across groups.
a. Design: Intercept + mathgr + faedRevis +
mathgr * faedRevis
The Levene Statistic p-value = 0.36 is greater than α = 0.05 ,so we fail toreject the null hypothesis that the variances are all equal. Since the variances appear to be equal (and we have random/independent samples), the assumption of homogeneity has been met we may continue with ANOVA .
One-way repeated measures ANOVA:one-way repeated measures analysis of variance if you had onecategorical independent variable and a normally distributed interval dependentvariable that was repeated at least twice for each subject. This is the equivalent of the paired samples t-test, but allows for two or more levels of the categorical variable.This tests whether the mean of the dependent variable differs by the categoricalvariable.
SPANOVA / MIXED ANOVA: (split plot ANOVA)
Robust: if the assumptions are not met even though we run the test.
Anova = difference between the groups (boys & Girls) (couples, non couples)Spanova = difference with in the groups. (1st class to last class anxiety level of boys)(time intervals)
Question: which intervention develop maths skill or develop confidence building in more effective in reducing students fear of statistics test score across the three time period:(pre-intervention, post intervention and follow up(three month later)
Assumptions:
One variable between the group One variable within the group Continuous variable Homogeniety of variance co-variance (will find through box m test) More assumptions like the ANOVA Homogeniety of variance co-variance:
Changes between the two groups are the same in the three time intervals.(through genereal liner model)
Procedure:Analyze>general linera model> repeated measeuretype: Time, factor 3shift with in group Fear 1, fear 2, fear 3intervals add 03 and between subject: types of classOptions: Check Estimation, effect size, homogenietyplots: groups = separate line, time = horizontal
Interpretation:
Between-Subjects Factors
Value Label N
type of
class
1 maths skills 15
2 confidence
building
15
Three dependent variables are there.
Descriptive Statistics
type of class
Mean
Std.
Deviation N
fear of stats
time1
maths skills 39.87 4.596 15
confidence
building
40.47 5.817 15
Total 40.17 5.160 30
fear of stats
time2
maths skills 37.67 4.515 15
confidence
building
37.33 5.876 15
Total 37.50 5.151 30
fear of stats
time3
maths skills 36.07 5.431 15
confidence
building
34.40 6.631 15
Total 35.23 6.015 30
Range is 20-60, 60 is the highst fear of mean we see here the aveerage of the means and which says that in Time1: they are very
much fear Time2: they are not very much fear Time3: they are very much fear
Box's Test of
Equality of
Covariance
Matricesa
Box's M 1.520
F .224
df1 6
df2 5680.302
Sig. .969
Tests the null
hypothesis that
the observed
covariance
matrices of the
dependent
variables are
equal across
groups.
a. Design:
Intercept + group
Within Subjects
Design: time
If it is > 0.05 Null hypotheses is equal, assumpition of co-variance is met
Multivariate Testsc
Effect
Value F
Hypothesi
s df
Error
df Sig.
Partial
Eta
Squared
Noncent.
Paramete
r
Observed
Powerb
time Pillai's
Trace
.663 26.59
3a
2.000 27.00
0
.00
0
.663 53.185 1.000
Wilks'
Lambda
.337 26.59
3a
2.000 27.00
0
.00
0
.663 53.185 1.000
Hotelling's
Trace
1.970 26.59
3a
2.000 27.00
0
.00
0
.663 53.185 1.000
Roy's
Largest
Root
1.970 26.59
3a
2.000 27.00
0
.00
0
.663 53.185 1.000
time *
group
Pillai's
Trace
.131 2.034a 2.000 27.00
0
.15
0
.131 4.067 .382
Wilks'
Lambda
.869 2.034a 2.000 27.00
0
.15
0
.131 4.067 .382
Hotelling's
Trace
.151 2.034a 2.000 27.00
0
.15
0
.131 4.067 .382
Roy's
Largest
Root
.151 2.034a 2.000 27.00
0
.15
0
.131 4.067 .382
a. Exact statistic
b. Computed using alpha = .05
c. Design: Intercept + group
Within Subjects Design: time
Time: Wilks Lambda: Significance is < than 0.05 and which means there are
differences in the fear scale of 03 dimension / time intervals.
Then we seee time * groups(time depends on groups): in here significane value is >
than 0.05 which means there is no differences. If the changes same over time for the
two different groups.
Partial Eta Squared(it is predicting dependent variable in time interval)(it tells us the
effect size of differences / relationship):we square root 0.663 = 0.81is is largely
effective and give the very large differences b/w the time interval for the dependent
variable FEAR. it will explain the 0.663 the time.
Mauchly's Test of Sphericityb
Measure:MEASURE_1
Within Subjects
Effect Mauchly's
W
Approx. Chi-
Square df Sig.
Epsilona
Greenhouse-
Geisser
Huynh-
Feldt
Lower-
bound
dimension1 time .348 28.517 2 .000 .605 .640 .500
Tests the null hypothesis that the error covariance matrix of the orthonormalized transformed dependent
variables is proportional to an identity matrix.
a. May be used to adjust the degrees of freedom for the averaged tests of significance. Corrected tests are
displayed in the Tests of Within-Subjects Effects table.
b. Design: Intercept + group
Within Subjects Design: time
Are the variance of the student remain same during the interval 1st time student get the marks (like = 1st student is 10, 2nd is 20, 3rd is 30, 4th is 40, 5th is 50)
Are the variance of the student remain same during the interval 2st time student get the marks (like = 1st student is 20, 2nd is 30, 3rd is 40, 4th is 50, 5th is 60)
Are the variance of the student remain same during the interval 3rd time student get the marks (like = 1st student is 30, 2nd is 40, 3rd is 50, 4th is 60, 5th is 70)
Here above variance are the same than we met the assumption of spersity (if the intervals have same variance / range it is called spercity)
Tests of Within-Subjects Effects
Measure:MEASURE_1
Source Type III
Sum of
Square
s df
Mean
Square F Sig.
Partial
Eta
Square
d
Noncent.
Paramete
r
Observe
d Powera
time Sphericity
Assumed
365.86
7
2 182.93
3
43.28
6
.00
0
.607 86.571 1.000
Greenhouse
-Geisser
365.86
7
1.210 302.24
7
43.28
6
.00
0
.607 52.397 1.000
Huynh-Feldt 365.86
7
1.281 285.63
2
43.28
6
.00
0
.607 55.445 1.000
Lower-
bound
365.86
7
1.000 365.86
7
43.28
6
.00
0
.607 43.286 1.000
time *
group
Sphericity
Assumed
19.467 2 9.733 2.303 .10
9
.076 4.606 .449
Greenhouse
-Geisser
19.467 1.210 16.082 2.303 .13
4
.076 2.788 .342
Huynh-Feldt 19.467 1.281 15.198 2.303 .13
2
.076 2.950 .352
Lower-
bound
19.467 1.000 19.467 2.303 .14
0
.076 2.303 .311
Error(tim
e)
Sphericity
Assumed
236.66
7
56 4.226
Greenhouse
-Geisser
236.66
7
33.89
4
6.983
Huynh-Feldt 236.66
7
35.86
5
6.599
Lower-
bound
236.66
7
28.00
0
8.452
a. Computed using alpha = .05
There are difference b/w faer scale when effecting time.
Tests of Within-Subjects Contrasts
Measure:MEASURE_1
Source Time Type III
Sum of
Squares df
Mean
Square F Sig.
Partial
Eta
Squared
Noncent.
Paramete
r
Observed
Powera
time Linear 365.067 1 365.06
7
49.22
2
.000 .637 49.222 1.000
Quadratic .800 1 .800 .772 .387 .027 .772 .136
time *
group
Linear 19.267 1 19.267 2.598 .118 .085 2.598 .344
Quadratic .200 1 .200 .193 .664 .007 .193 .071
Error(tim
e)
Linear 207.667 28 7.417
Quadratic 29.000 28 1.036
a. Computed using alpha = .05
Levene's Test of Equality of Error Variancesa
F df1 df2 Sig.
fear of stats
time1
.893 1 28 .353
fear of stats
time2
.767 1 28 .389
fear of stats
time3
.770 1 28 .388
Tests the null hypothesis that the error variance of the
dependent variable is equal across groups.
a. Design: Intercept + group
Within Subjects Design: time
Here assumption are met that the variances are same for 1st, 2nd & 3rd time interval
Tests of Between-Subjects Effects
Measure:MEASURE_1
Transformed Variable:Average
Source Type III
Sum of
Squares df
Mean
Square F Sig.
Partial Eta
Squared
Noncent.
Parameter
Observed
Powera
Intercept 127464.10
0
1 127464.100 1531.757 .000 .982 1531.757 1.000
Group 4.900 1 4.900 .059 .810 .002 .059 .056
Error 2330.000 28 83.214
a. Computed using alpha = .05
Here we see signifiacnce level which is >0.05 says that with in the group both are
effecting the fear level in the same way and have no differences
Its partial ETA iafter squaring is 0.04 which is very small hich says that there is no
efect of group on the fear
Here lines are in same direction in all 3rd intervals which says that there are no differences
One-way MANOVA:
Multivariate Analysis of Variance ( Manova ) is used to model two or more dependent variables that are continuous with one or more categorical predictor variables.
Assumptions:
One independent variable consists of two or more categorical independent groups. Two or more dependent variables that are either interval or ratio (continuous) Multivariate Normality Equality of variances between the independent groups (homogeneity of variances). Independence of cases.
Question:
Do male and female differ in terms of overall well being in other words are males better adjusted than female in term of their positive and negative mood stats and levels of perceived stress ?
Interpretion:
Between-Subjects Factors
Value Label N
sex 1 MALES 184
2 FEMALES 248
This table show us the independent variable which gender sex
the sample size cases of males are N= 184
and females cases areN= 248
Descriptive Statistics
sex Mean Std. Deviation N
Total positive affect MALES 33.62 6.985 184
FEMALES 33.69 7.439 248
Total 33.66 7.241 432
Total negative affect MALES 18.71 6.901 184
FEMALES 19.98 7.178 248
Total 19.44 7.082 432
Total perceived stress MALES 25.79 5.414 184
FEMALES 27.42 6.078 248
Total 26.72 5.854 432
the Descriptive Statistics table show us the samples sizes of all dependent variables
means and standard deviations.
Box's Test of Equality of
Covariance Matricesa
Box's M 6.942
F 1.148
df1 6
df2 1074771.869
Sig. .331
Tests the null hypothesis that
the observed covariance
matrices of the dependent
variables are equal across
groups.
a. Design: Intercept + sex
One of the assumptions of the MANOVA is homogeneity of covariances, which is
tested for by Box's Test of Equality of Covariance Matrices. If the "Sig." value
is less than .005 (P < 0.05) then the assumption of homogeneity of covariances was
violated. Then we can say this the assumption of homogeneity of covaiances has
been met (P = .331).
Multivariate Testsc
Effect
Value F
Hypothesis
df Error df Sig.
Partial Eta
Squared
Noncent.
Parameter
Observed
Powerb
Intercept Pillai's Trace .987 10841.625a 3.000 428.000 .000 .987 32524.875 1.000
Wilks'
Lambda
.013 10841.625a 3.000 428.000 .000 .987 32524.875 1.000
Hotelling's
Trace
75.993 10841.625a 3.000 428.000 .000 .987 32524.875 1.000
Roy's Largest
Root
75.993 10841.625a 3.000 428.000 .000 .987 32524.875 1.000
sex Pillai's Trace .024 3.569a 3.000 428.000 .014 .024 10.707 .788
Wilks'
Lambda
.976 3.569a 3.000 428.000 .014 .024 10.707 .788
Hotelling's
Trace
.025 3.569a 3.000 428.000 .014 .024 10.707 .788
Roy's Largest
Root
.025 3.569a 3.000 428.000 .014 .024 10.707 .788
a. Exact statistic
b. Computed using alpha = .05
c. Design: Intercept + sex
The Multivariate Tests table is where we find the actual result of the one-way MANOVA.
You need to look at the second Effect, labelled "Sex", and the Wilks' Lambda row
(highlighted in red). To determine whether the one-way MANOVA was statistically significant
you need to look at the "Sig." column. We can see from the table that we have a "Sig."
value of .014, which means P < 0.0166. Therefore, we can conclude that this sex gender
impact was significantly dependent on which prior behaviours they had attended (P <
0.05/3= 0.01666).
Levene's Test of Equality of Error Variancesa
F df1 df2 Sig.
Total positive affect 1.065 1 430 .303
Total negative affect 1.251 1 430 .264
Total perceived stress 2.074 1 430 .151
Tests the null hypothesis that the error variance of the dependent variable is equal
across groups.
a. Design: Intercept + sex
Levene's Test of Equality of Error Variances Table, as shown
We can see from the table above that all dependent variables have homogeneity of
variances (P > .05)
Tests of Between-Subjects Effects
Source Dependent
Variable Type III Sum
of Squares df
Mean
Square F Sig.
Partial
Eta
Squared
Noncent.
Parameter
Observed
Powerb
Corrected
Model
Total
positive
affect
.440a 1 .440 .008 .927 .000 .008 .051
Total
negative
affect
172.348c 1 172.348 3.456 .064 .008 3.456 .458
Total
perceived
stress
281.099d 1 281.099 8.342 .004 .019 8.342 .822
Intercept Total
positive
affect
478633.634 1 478633.634 9108.270 .000 .955 9108.270 1.000
Total
negative
affect
158121.903 1 158121.903 3170.979 .000 .881 3170.979 1.000
Total
perceived
stress
299040.358 1 299040.358 8874.752 .000 .954 8874.752 1.000
sex Total
positive
affect
.440 1 .440 .008 .927 .000 .008 .051
Total
negative
affect
172.348 1 172.348 3.456 .064 .008 3.456 .458
Total
perceived
stress
281.099 1 281.099 8.342 .004 .019 8.342 .822
Error Total
positive
affect
22596.218 430 52.549
Total
negative
affect
21442.088 430 49.865
Total
perceived
stress
14489.121 430 33.696
Total Total
positive
affect
512110.000 432
Total
negative
affect
184870.000 432
Total
perceived
stress
323305.000 432
Corrected
Total
Total
positive
affect
22596.657 431
Total
negative
affect
21614.435 431
Total
perceived
stress
14770.220 431
a. R Squared = .000 (Adjusted R Squared = -.002)
b. Computed using alpha = .05
c. R Squared = .008 (Adjusted R Squared = .006)
d. R Squared = .019 (Adjusted R Squared = .017)