QADM Case Study

Saqib Ismail Khatri (Student ID: 23616)

Instructor: Miss Hina Samreen
Date: 30 December 2018
1. Which variables are the participant variables? (They act as independent variables in this study.)

Ans. Following variables are independent variables

o Gender
o Sports
o Anger-out
o Anger-in
o Control-out
o Control-In

2. What are the dependent variables?

Ans. Expression is a dependent variable as it has been calculated by equation: (Anger-Out) + (Anger-
In) - (Control-Out) - (Control-In) + 48

3. Is Anger-Out a quantitative or qualitative variable?

Ans. It’s a qualitative variable because it is a categorical variable.

4. Does Anger-Out have a positive skew, a negative skew, or no skew?

Ans. Anger-out is positively skewed because it has positive coefficient of skewness = 0.586
Statistics
Anger-Out
N Valid 78
Missing 0
Skewness .586
Std. Error of Skewness .272

5. What are the mean and standard deviation of the Anger-Out scores?

Ans:

Statistics
Anger-Out
N Valid 78
Missing 0
Mean 16.08
Std. Deviation 4.217

6. Is there a difference in how much males and females use aggressive behavior to improve an angry
mood? For the variable "Anger-Out" scores do the following questions:
 a. Create box plots.
b. Create stem and leaf diagrams

Ans. Yes, there is a difference in males and females who use aggressive behavior to improve angry
mood because they have difference in mean:

Descriptive Statisticsa
N Mean
Anger-Out 30 16.57
Valid N (listwise) 30
a. Gender = Male

Descriptive Statisticsa
N Mean
Anger-Out 48 15.77
Valid N (listwise) 48
a. Gender = Female

Box plot and stem and leaf diagram is saved in data file.

7. Is the variable anger out score is normally distributed if so answer the following questions

Ans:

Ho: anger-out is normally distributed

Ha: anger-out is not normally distributed

α = 0.05 CI=95%

Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic Df Sig. Statistic df Sig.
Anger-Out .119 78 .008 .960 78 .015
a. Lilliefors Significance Correction

P-value < α = 0.05

Therefore, reject Ho. Hence, anger-out is not normally distributed.

8. At the 5% level of sig. do the data provide evidence to conclude that there is no difference between
mean anger out score between male and female?

Ans:

Assumption 01: Data has independent samples (Male/Female)

Assumption 02: Data is normally distributed.

Checking normal distribution

Ho: Both data sets are normally distributed
Ha: Both data sets are not normally distributed

α = 0.05

Tests of Normalitya
Kolmogorov-Smirnovb Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
Anger out .138 30 .151 .951 30 .183
a. Gender of respondent = Male
b. Lilliefors Significance Correction

Tests of Normalitya
Kolmogorov-Smirnovb Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
*
Anger out .108 48 .200 .960 48 .100
*. This is a lower bound of the true significance.
a. Gender of respondent = Female
b. Lilliefors Significance Correction

From Kolmogrov and Shapiro, P-value > .05

Do not reject Ho; that is both data sets are normally distributed.

Conclusion: At α= 0.05 data do provide evidence that differences are normally distributed.

Assumption 03: Equality of variance.

Check equality of variance through F-test

Ho: Two variances are equal

Ha: Two variances are not equal

 α = 0.05

Conclusion: P value of F-Test: 0.952 > α, so we do not reject Ho, hence equality of variance is
proven. We will go for pooled T-Test

Performing pool-t-test

Let µ1 = mean anger out score of male

µ2 = mean anger out score of male

Ho: µ1= µ2

Ha: µ1≠ µ2

α = 0.05

P value of pooled T-Test: 0.421 > α, so we do not reject Ho.

Conclusion: Therefore, at a 5% level of significance data do provide evidence to conclude that there
is no difference between mean anger out score between male and female.

9. Compute a 95% confidence interval on the difference between mean anger out score between male
and female.
Ans: 95% confidence interval on the difference between mean anger out score between male and
female is shown in upper row of the far right column of the above table.

10. What is the range of the Anger-In scores? What is the interquartile range?

(a) Range = Max-Min

31-10 = 21 units.
(b) Interquartile range = Q3 – Q1
IQR: 22-15 = 7 units.
Statistics
Anger in
Valid 78
N
Missing 0
Range 21
Minimum 10
Maximum 31
25 15.00
Percentiles 50 18.50
75 22.00

11. Create box plots for the Anger-In scores by sports participation.

Ans: Please see the out-put of Q 11 for Box Plot.

12. Is the variable Anger in score is normally distributed if so answer the following questions

Ans: Ho: anger-in is normally distributed

Ha: anger-in is not normally distributed

α = 0.05 CI=95%

Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic Df Sig. Statistic df Sig.
Anger-In .067 78 .200* .979 78 .235
a. Lilliefors Significance Correction
*. This is a lower bound of the true significance.

P-value > α = 0.05

Therefore, do not reject Ho. Hence, anger-in score is normally distributed.

13. Calculate the 95% confidence interval for the difference between mean Anger-In score for the
athletes and non-athletes. What can you conclude?

Ans: checking equality of variance by F-test

Ho: Variances are equal

Ha: Variances are not equal

α = 0.05, CI=95%

Group Statistics
Std. Error
Sports N Mean Std. Deviation Mean
Anger-In athletes 25 16.68 3.682 .736
Non-athletes 53 19.47 4.886 .671

Independent Samples Test

Levene's Test for
t-test for Equality of Means
Equality of Variances
95% Confidence
Interval of the
Difference
Sig. (2- Mean Std. Error
F Sig. t df Lower Upper
tailed) Difference Difference
Equal
variances 2.355 0.129 -2.534 76 0.013 -2.792 1.102 -4.986 -0.598
assumed
Anger-In Equal
variances
-2.802 60.994 0.007 -2.792 0.996 -4.784 -0.799
not
assumed

Sig Value: 0.129 > α = 0.05

Therefore, do not Reject Ho. Variances are equal.

We will perform Pooled T-test

µ1 = mean anger in athlete

µ2 = mean anger in non-athlete

Ho: µ1= µ2

Ha: µ1≠ µ2

α = 0.05
 P-value= 0.013 < 0.05 therefore, reject Ho

Therefore, at 95% confidence interval, data do provide evidence to conclude that there is difference
between mean anger out score between athlete and non-athlete

14. What is the overall mean Control-Out score? What is the mean Control-Out score for the athletes?
What is the mean Control-Out score for the non-athletes?
 Ans

 Overall mean is 23.69

Statistics
Control out
Valid 78
N
Missing 0
Mean 23.69

 Mean Control-Out score for the athletes is 24.68

 Mean Control-Out score for the non-athletes is 23.23

Statisticsa
Control out Statisticsa
Valid 25 Control out
N
Missing 0 Valid 53
Mean 24.68 N
Missing 0
a. Athlete/Non Athlete = Mean 23.23
Athletes
a. Athlete/Non Athlete = Non Athletes

15. Is the variable Control-out score is normally distributed if so answer the following questions.

Ans. Ho: Control-out is normally distributed

Ha: Control-out is not normally distributed

α = 0.05 CI=95%

Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
Control-Out .093 78 .091 .969 78 .054
a. Lilliefors Significance Correction
P-value > α = 0.05

Therefore, do not reject Ho. Hence, Control-out score is normally distributed.

16. At the 5% level of sig. determine if the difference in the mean Control-Out score for athletes and
non-athletes is statistically significant.
 Ans:

Assumption 01: Data has independent samples (athletes/non-athletes)

Assumption 02: Data is normally distributed. (see ans 15)

Assumption 03: Equality of variance.

Check equality of variance through F-test

Ho: Two variances are equal

Ha: Two variances are not equal

α = 0.05

Performing F-test

Independent Samples Test

Levene's Test for
t-test for Equality of Means
Equality of Variances
95% Confidence
Interval of the
Difference
Sig. (2- Mean Std. Error
F Sig. t df Lower Upper
tailed) Difference Difference
Equal
variances 3.02 0.086 1.283 76 0.203 1.454 1.133 -0.803 3.71
assumed
Control-
Equal
Out
variances
1.399 58.828 0.167 1.454 1.039 -0.626 3.533
not
assumed

P-Value = 0.086 > 0.05

Therefore, do not reject Ho.

Thus, Variances are equal and now we will for pooled T-Test

Let:

µ1= mean Control-Out score for athletes

µ2= mean Control-Out score for non-athletes

Ho: µ1= µ2

Ha: µ1 ≠ µ2

P – Value (pooled t-test) = 0.203 > 0.05

Therefore, do not reject Ho.

Conclusion: Therefore, at 5% level of significance data do not provide evidence to conclude that the
difference in the mean Control-Out score for athletes and non-athletes is statistically significant.

17. What is the variance of the Control-In scores for the athletes? What is the variance of the Control-In
scores for the non-athletes?

Ans

 Variance of the Control-In scores for the athletes is 20.477

 Variance of the Control-In scores for the non-athletes is 22.823

Statisticsa
Control in
Statisticsa
Valid 25
N Control in
Missing 0
Valid 53
Variance 20.477 N
Missing 0
a. Athlete/Non Athlete =
Variance 22.823
Athletes
a. Athlete/Non Athlete = Non Athletes

18. What is the standard error of the mean for the

Control-In scores for the athletes? What is the standard error of the mean of the Control-In scores
for the non-athletes? Why is it smaller for non-athletes?

Ans:

 Standard error of the mean for the Control-In scores for the athletes = 0.905
 Standard error of the mean for the Control-In scores for the non-athletes = 0.656.
 It is smaller because its sample mean less deviates from the actual mean of population.
 Statisticsa
a
Statistics Control in
Control in Valid 25
Valid 53 N
N Missing 0
Missing 0 Mean 24.32
Mean 20.85 Std. Error of Mean .905
Std. Error of Mean .656 a. Athlete/Non Athlete =
a. Athlete/Non Athlete = Non Athletes Athletes

19. Do athletes or non-athletes calm down more when angry? Conduct an independent sample t test to
see if the mean difference between groups in Control-In scores is statistically significant.

If we look at the chart below we can conclude that the athletes control angry feelings by calming down
or cooling off is greater than Non-athletes.
Group Statistics
Std. Error
Sports N Mean Std. Deviation Mean
Control-In Athletes 25 24.32 4.525 .905
Non-athletes 53 20.85 4.777 .656
Performing independent t-test.
Let us assume data to be normally distributed.
Let µ1= mean Control-In score for athletes

µ2= mean Control-In score for non-athletes

Ho: µ1= µ2

Ha: µ1 ≠ µ2

Checking equality of variance:

Ho: variance of Control-In score for athletes and non-athletes are same

Ha: variance of Control-In score for athletes and non-athletes are not same

Considering α = 0.05
P-value = 0.7 > 0.05 therefore, do not reject Ho, Thus variance of Control-In score for athletes and non-
athletes are same.

Now go for pool-t-test.

P-value = 0.003 < 0.05

Therefore, at 5% level of significance data do provide evidence to conclude that the mean difference
between groups in Control-In scores is statistically significant.

20. What is the correlation between the Control-In and Control-Out scores? Is this correlation
statistically significant at the 0.01 level?

Ans: Let ρ = Correlation coefficient of linear association

Ho: ρ = 0 (There is no linear correlation between Control-In and Control-Out scores)

Ha: ρ > 0 (There is positive linear correlation between Control-In and Control-Out scores)

Applying correlation test association:

Correlations
Control out Control in
Pearson Correlation 1 .719**
Control out Sig. (1-tailed) .000
N 78 78
Pearson Correlation .719** 1
Control in Sig. (1-tailed) .000
N 78 78
**. Correlation is significant at the 0.01 level (1-tailed).

Since ρ = 0.719 therefore there is a significant positive linear correlation between Control-In and
Control-out scores.

Also P value: 0.000 < 0.01

Therefore reject Ho; hence we can say that there is positive linear correlation between Control-In and
Control-Out scores.

21. Are the variable gender and sports associated?

At first we will calculate Frequency variable.

Gender * Sports Crosstabulation

Count
Sports
athletes Non-athletes Total
Gender Male 11 19 30
Female 14 34 48
Total 25 53 78
Let Ho: Gender and sports are not associated
Ha: Gender and sports are associated

Calculating percentage distribution of Sports with respect to gender.

Gender * Sports Crosstabulation

Sports
Athletes Non-athletes Total
Gender Male Count 11 19 30
% within Sports 44.0% 35.8% 38.5%
Female Count 14 34 48
% within Sports 56.0% 64.2% 61.5%
Total Count 25 53 78
% within Sports 100.0% 100.0% 100.0%
As the generated % distribution among athletes and non-athletes are different so we reject null
hypothesis. Hence, variable gender and sports are sassociated.

22. Would you expect the correlation between the Anger-Out and Control-Out scores to be positive or
negative? Compute this correlation.

Let ρ = Correlation coefficient of linear association

Ho: ρ = 0 (There is no linear correlation between Anger-Out and Control-Out scores)

Ha: ρ ≠ 0 (There is linear correlation between Anger-Out and Control-Out scores)

 Applying correlation test association:

Correlations
Anger- Control-
Out Out
Anger- Pearson
1 -.583**
Out Correlation
Sig. (1-tailed) .000
N 78 78
Control- Pearson
-.583** 1
Out Correlation
Sig. (1-tailed) .000
N 78 78
**. Correlation is significant at the 0.01 level (1-tailed).

Since ρ = -0.583 therefore there is a negative linear correlation between Anger-Out and Control-Out
scores

Also P value < 0.01

Therefore reject Ho; hence we can say that there is negative linear correlation between Anger-Out
and Control-Out scores

23. Find the regression line for predicting Anger-Out from Control-Out.
a. What is the slope?
b. What is the intercept?
c. Is the relationship at least approximately linear?
d. Test to see if the slope is significantly different from 0.
Let Ho: β = 0 (Control-out score is not a useful predictor of Anger out score)

Ha: β ≠ 0 (Control-out score is a useful predictor of Anger out score)

Analyzing linear regression:

Model Summary
Adjusted R Std. Error of
Model R R Square Square the Estimate
1 .583a .340 .331 3.450
a. Predictors: (Constant), Control-Out

This table provides the R and R2 values. The R value represents the simple correlation and is 0.583 (the
"R" Column), which indicates a high degree of correlation. The R2 value (the "R Square" column)
indicates how much of the total variation in the dependent variable, Anger-out, can be explained by the
independent variable, Control-out. In this case, 34% can be explained, which is very large.

Y(anger-out) = 28.495 – 0.524X (Control-out)

Coefficientsa
Unstandardized Standardized
Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 28.495 2.025 14.073 .000
Control-Out -.524 .084 -.583 -6.250 .000
a. Dependent Variable: Anger-Out

a) Slope = -0.583
b) Intercept = 28.495

Since P-value = 0 < 0.05; reject Ho

Therefore at 5% level of significance data do provide evidence to conclude that control-out score is a
useful predictor of anger-out score.

c) Yes relationship is not approximately linear

d) Slope is -0.583 < 0

24. What is the standard error of the estimate?

The standard error of the estimate is a measure of the accuracy of predictions. Recall that the
regression line is the line that minimizes the sum of squared deviations of prediction (also called the
sum of squares error).

In above case Standard error of estimate is 3.45