Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
TO:
FROM:
Re:
Date:
Outcome
BMI
PreIntervention
BMI
Post
Intervention
M
25.7
1
SD
27.93 4.95
* p < .05.
95% CI for
Mean
Difference
SD
4.05
30
0.28, 4.16
df
.35
2.33*
29
Group
0
SD
SD
25.90
4.20
1
3
29.50
5.02
1
7
28.56
4.92
27.57
5.07
27.03
6.21
28.45
4.16
Marital
Gender
Alcohol
95% CI for
Mean
Difference
1
1
1
1
1
9
1
9
t
2.09
*
df
-2.89, 4.89
0.53.
28
-5.89, 3.05
0.50
8
15.3
0
-7.14, -0.064
28
* p < .05.
3. Dependent sample t-test.
A paired sample t-test was used to determine whether there was a statistically
significant difference between the BMI of the patients at post intervention and preintervention. The mean of BMI at post-intervention was found to be 2.22 units
higher than at pre-intervention, which is statistically significant at 5% level of
significance, t(29) = 2.34, p = 0.026, 95% CI for differences in mean (0.28,4.16).
To test for clinical significance, we used the cohens d effect size estimation,
d=
t
2.33
=
=0.425 , the value is, between 0.2 0.5 which is considered a small
N 30
effect size and thus the intervention strategy might not be practically significant.
The assumption for performing a paired sample t-test include normal distribution of
responses which can be assessed using a histogram; it is also assumed that the
variances for the two measurements are equal. By convention, we believe that the
observations are independent, and the observations at the two points of
measurement are from same subjects.
formulae,
d=
X m X F
s2 p
, where
2
p
Cohens d for gender was found to be 0.19 which is a small effect size and thus,
there is no clinical difference in patients BMI between Males and Females. Based on
alcohol consumption, those who had not consumed alcohol in the last 12 months
had (M = 27.03, SD = 6.21) and those who had consumed had (M = 28.45, SD =
4.16), the difference in the two means is not statistically significant at = 0.05,
t(15.30) = -0.68, p = 0.51 and 95% confidence interval for the difference (5.90,3.05). The difference between the two alcohol conditions was also not clinically
significant, cohens d = 0.3. Analysis was also done on marital status and Men had
(M = 25.90, SD = 4.20) and Females (M = 29.50, SD = 5.02), the difference in the
means was significant at 5% level of significance, t(28) = -2.09, p = 0.046, 95%
confidence interval for the difference in means (-7.14,-0.07). Although the difference
will not be considered significant if the 95% confidence interval is used since 0 is