Você está na página 1de 12

Running head: DATA PRESENTATION FOR ASSIGNMENT 2, PART 3

Data Presentation for Assignment 2, Part 3


Bryan Blattert
Lance Hicks
Michelle Loding
Shellisa Mitchell
Wayne State University
SW 4810
December 22, 2015
Data Presentation for Assignment 2, Part 3

SAMPLE
The data in this study was collected from a team to look into a new treatment strategy
that is calculated to decrease major trauma symptoms in the subjects. There were 50 subjects, or
clients, that took part in the study. After the clients arrived at the treatment center an assessment
was done on each individual to gather information such as age, ethnicity, gender, and number of
traumas experienced. Before being admitted into a cottage for the youths, a pretest was
completed. This Trauma Symptoms Scale pretest consisted of 15 self-reported answers to

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3


2
indicate how frequently the youth experience or participate in detrimental behaviors as a result of
the trauma they have experienced. Some of the answers may include nightmares, sadness,
flashbacks, angry outbursts, undesirable thoughts, fear, withdrawal from society, etc.
Following the pretest, the clients are then randomly assigned to one of three cottages.
They are also randomly assigned to begin a new treatment strategy or a routine one. While
staying at the facility, data is gathered and analyzed about various things like earned off campus
privileges before treatment and after, serious behavioral incidents, and number of times a client
cancelled a treatment session. The client were also given the chance to rate how well the
relationship was with their therapist.
Analysis of the variables within the data set show that exactly half (25 individuals) of the
participants were female. Additionally, 40% labeled themselves as Caucasian, 30% as African
American, 26% as Hispanic, and 4% as other. The age range for the group was 7 to 15 years old,
with the median age being 11 years old. (See Table 1)

Table 1

Gender
Fre
quency
V
alid

M
ale
F
emale
T
otal

P
ercent

25

25

50

Valid
Percent

5
0.0
5
0.0
1
00.0

Cumula
tive Percent

50.0

50.0

50.0

100.0

100.0

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3

Ethnic
Fre

quency
V
alid

Caucasia

ercent
20

n
African

15

American
Hispanic

13

Other

Total

Valid

50

Percent
4

0.0
3
0.0
2
6.0
4.
0
1

Cumula
tive Percent

40.0

40.0

30.0

70.0

26.0

96.0

4.0

100.0

100.0

00.0

Age
Fre
quency
V
alid

7
.00
8
.00
9
.00
1
0.00
1
1.00
1
2.00
1
3.00
1
4.00
1
5.00

P
ercent

11

3
2

Valid
Percent

4.
0
6.
0
1
6.0
2
2.0
1
8.0
1
0.0
1
4.0
6.
0
4.
0

Cumula
tive Percent

4.0

4.0

6.0

10.0

16.0

26.0

22.0

48.0

18.0

66.0

10.0

76.0

14.0

90.0

6.0

96.0

4.0

100.0

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3


T
otal

50

1
00.0

100.0

The treatment groups were divided evenly (25 participants in the routine treatment and 25
participants in the new treatment). This study cannot be representative of the general youth
population due to the presumed mental health issues of this particular population and the
participants age. Not enough of a sample was obtained to presume generalizability of this study.
The following are three questions that were raised and analyzed to determine how
treatments helped to decrease trauma symptoms in youth. Data was collected and analyzed to
make the determinations. The following is the findings to the questions. The tests were all
conducted at an alpha level of 0.05.

Bivariate Analysis
Chi Square
Research Question: Is there a correlation between the cottage assignments and earned off
campus privileges?
Null hypothesis: There is no correlation between cottage assignments and earned off campus
privileges.
Alternative hypothesis: There is a correlation between the cottage assignments and earned off
campus privileges.
Independent Variable: Cottage assignment
Dependent Variable: Earned Off-Campus Privileges

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3


A cross tabulation with chi-square (table 2) was the method used to determine whether a

relationship existed between cottage assignment and off campus privileges earned. A total
number of 50 patients were used in the test, 25 who earned off-campus privileges and 25 who did
not earn off-campus privileges. The results of the cross tabulation with chi square show 64%
earned off-campus privileges, or 16 out of the 25 youth who received new treatment, while 36%
earned off-campus privileges or 9 out of the 25 youth who participated in the routine treatment.
The cross tabulation with chi-square is the appropriate statistic for determining if this difference
in the results is due to sampling error, and shows whether and how the variables are related to the
other. Table 2 below shows the Pearson chi-square value at 3.920, indicating that the new and
routine treatment group are significantly different on whether they earned off-campus privileges
or not, and the difference is not contributed to sampling error.
Table 2

Case Processing Summary


Cases
Valid

Missing

Total

P
N
Cottage * Earned
Off Campus Privileges

ercent

P
N

00.0%

ercent

ercent

0.

0%

00.0%

Chi-Square Tests
V
alue
Pearson Chi-

df
3.

920a

Square
Continuity
Correction

Asymp.

2.
880

Sig. (2-sided)

.048

.090

Exact

Exact

Sig. (2-sided)

Sig. (1-sided)

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3


Likelihood Ratio

3.

973

.046

Fisher's Exact Test

.089

Linear-by-Linear

3.

Association

842

N of Valid Cases

.044

.050

5
0

a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 12.50.
b. Computed only for a 2x2 table

Earned off campus privileges * Treatment Crosstabulation


Treatment
New

Routin

Treatment
Earned off campus
privileges

Count

es
Expected
Count
% within
Treatment
N

Count

o
Expected
Count
% within
Treatment
Total

Count

Expected
Count
% within
Treatment

Pearsons r

e Treatment

T
otal

16

12.5

12.5

64.0%

36.0%

16

12.5

12.5

36.0%

64.0%

25

25

25.0

25.0

100.0%

100.0%

2
5
2
5.0
5
0.0%
2
5
2
5.0
5
0.0%
5
0
5
0.0
1
00.0%

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3


Research Question: Does the age of the person affect the number of cancelled therapy
sessions?

Null hypothesis: Age does not affect the number of cancelled therapy sessions.
Alternative hypothesis: There is a relationship between age and number of cancelled therapy
sessions.
Independent variable: Age of person
Dependent variable: Number of cancelled sessions
The parametric formula, Pearsons r, was used to analyze whether the independent, interval level
variable (age of the person) had an effect on the dependent ratio level variable (number of
cancelled therapy sessions). The correlation coefficient calculated was -0.164 (see table 3),
which indicates an inverse relationship between age and number of cancelled sessions. This
means that older clients were less likely to cancel sessions, while younger clients were more
likely to cancel sessions. Because the value of Pearsons r is much closer to 0 than to 1, the
correlation is fairly weak.
The formula for Pearsons r assumes linear, homoscedastic data that has a fair amount of
variability for each of the two variables. A scatterplot depicting the data was generated (see table
4), in order to determine whether the data meets these assumptions. By looking at this graph, it
can be verified that the data has a sufficient amount of variability for each variable, and is
homoscedastic, with a similar amount of variance for each variable. The data also appears to be
reasonably linear on the scatterplot, therefore we can conclude that the parametric correlation,
Pearsons r, is appropriate for analyzing this data and will produce reliable results. A weak
negative correlation is visually apparent in the scatterplot, consistent with the correlation
coefficient.

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3


8
The two-tailed test of significance gives us a value of 0.254. This indicates that the correlation is
not statistically significant and could be due to sampling error. The probability that the
correlation is due to sampling error is not low enough to reject the null hypothesis.

Table 3
Pearsons r Correlations

Age

Pearson Correlation Sig. (2-tailed)


N

Age

Cancelled Therapy Session

-.164
.254
50

50
Cancelled Therapy Session

Table 4

Pearson Correlation Sig. (2-tailed)


N

-.164
.254
50

1
50

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3

T-test
The Analysis of Variance of the T-test reflected these indications:
Research Question: Is there a relationship between new or routine treatments and the number of
serious behavioral issues?
Null hypothesis: There is no relationship between new or routine treatments and the number of
serious behavioral issues.
Alternative hypothesis: There is a relationship between new or routine treatments and the
number of serious behavioral issues.
Independent variable: New or routine treatment
Dependent variable: Number of serious behavioral issues

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3


The independent t-test (table 5) was performed in analyzing the nominal level, independent

10

variable (new or routine treatments) had an effect on the ratio level, dependent variable (serious
behavioral issues). In the study the serious behavioral issues mean was 1.5000 and the new or
routine treatments was 5.5000. Moreover, this exhibits the p value in p=.525 (t=-5.095, df=49)
which is greater than p=0.05 which shows the difference in the means of treatments verses the
means of serious behavioral issues statistically have significance. The results displayed of
independent t-test in Levenes tests equality of variance results and equality of means (table 6).
The significance of the results of Levenes test displays a level of .525 which expressed a higher
p value than 0.05 rejection level. The t-Test of equality (table 7) of means can be analyze in
using this research question which shows a two-tailed p-value of .000 distinguished by equal
variance. The determining factor of the null hypotheses can be rejected because it has a lower pvalue of .000 than the rejected level of 0.05. Researchers can consider that there is no
relationship between new or routine treatments and the number of serious behavioral issues. The
variances cannot be assumed.
Table 5: Group Statistics
Group

Mean

Standard Deviation STD.

Error Mean

Students treatments

50

1.5000

.50508

.07143

SBI

50

5.5000

5.48188

.77526

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3


Table 6: Independent Sample Test- Levenes Test

11

Levenes Test for Equality of Variances

Students

Equal variances assumed

Significance

-.092

.525

Equal variances not assumed


Table 7: Independent Samples Test
t-Test for Equality of Means
t
Equal variances assumed

-5.095

Equal variances not assumed -5.095

df
49
26.00

Significance (2 tailed)
.000
.000

Mean Difference
-4.00000
-4.00000

Results
Results were as follows: For test 1 (chi square), the alternative hypothesis is true. There
was a significant difference between cottage assignment and earned off campus privileges. For
test 2 (Pearsons r) the alternative hypothesis is true with the finding of the younger clients more
apt to cancel a therapy session. Finally, the last test (t-test) found the alternative hypothesis to
hold true with there being a relationship with the type of treatment and serious behavioral issues.
There are many real world applications to this kind of research. Further studies with the new
type of treatment would warrant more research due to the positive response of the clients. The
more relationships that are examined, the more calculated the treatment for the client. For
example, examining the correlation between age and behavioral outbursts or if there is a
relationship between cottage assignment and behavioral incidents.

DATA PRESENTATION FOR ASSIGNMENT 2, PART 3

12

References
Rubin, A. (2013). Statistics for evidence-based practice and evaluation (3rd ed.). Belmont, Calif.:
Brooks/Cole.