You are on page 1of 7

Bo Guan

Undieh Midterm

1. Question 1
a. Scientific/Research Hypothesis-The purpose of this study is to look at whether wheelchair maneuvering caused different pressures on the ischial tuberosity in individuals with lower spinal
cord injury versus healthy controls.
b. Test Hypotheses
i. H0: pressureUnderLIT (SCI) = pressureUnderLIT (control)
ii. H1 : pressureUnderLIT (SCI) ≠ pressureUnderLIT (control)
c. Test Data and Assumption
i. The data consists of two group of subjects who had their pressure under LIT(left ischial tuberosity) taken. Therefore, the independent variable is “group” (spinal cord injury vs
control) and the dependent variable is the pressure. Pressure is a continuous scalar variable, hence it can be assumed to be normally distributed in the population from which the
sampled data was generated.
d. Test statistics and decision rule
i. Given that normality of pressure can be assumed, there are two groups to be compared, an independent samples t-test is selected for the data, and will be evaluated at an alpha
value=0.05.
e. Calculations

Figure 1a.
Independent Samples Test
Levene's Test for Equality of
Variances t-test for Equality of Means
95% Confidence Interval of the
Mean Std. Error Difference
F Sig. t df Sig. (2-tailed) Difference Difference Lower Upper
pressureUnderLeftIschialT Equal variances assumed .564 .462 -.569 18 .576 -7.00000 12.29444 -32.82967 18.82967
uberosity Equal variances not -.569 15.832 .577 -7.00000 12.29444 -33.08554 19.08554
assumed

Equality of variances, homogenous of variance assumption. This shows that they are the same samples because P values of the Levene’s Test is greater than .05.
Assume that the variances are the same. Accept the null hypothesis. Variance do not make a difference.
f. Statistical decision
i. The calculated t-value is -.569 which for 7 df equates a significance level of >.576 equates a significance level of .576. At the .05 alpha, the difference between the SCI and the
control group is not significant. (figure 1a)
ii. H0: pressureUnderLIT (SCI) = pressureUnderLIT (control) ; Cannot reject H0 (p>.05)
iii. H1 : pressureUnderLIT (SCI) ≠ pressureUnderLIT (control); Reject H1 (p>.05)

g. Scientific conclusion
i. There is no significant effect of spinal cord injury on the pressure on left ischial tuberosity during wheel chair use.
2. Question 2
a. Scientific/Research Hypothesis-The purpose of this study is to look at whether bone strength is different among three age groups, the young (19-49), the middle-aged (50-69), and the
elderly (70 and above).
b. Test Hypotheses
i. H0: there is no difference in bone strength between the young, middle aged, and elderly.
Bo Guan
Undieh Midterm

ii. H1 : there is a difference in bone strength between the young, middle aged, and elderly.

c. Test Data and Assumption


i. The data consists of three group of subjects whose femurs were collected from their cadevors and subjected to testing. The force to fracture the bone was measured in newtons.
Therefore the group is the independent variable(young, middle-aged, elderly) and the force needed to break the bone is the dependent variable. Bone strength, as measured by the
fore to fracture a bone, is a continuous scalar variable, hence it can be assumed to be normally distributed in the population from which the sampled data was generated.
d. Test statistics and decision rule
i. Given that normality of force to break a bone can be assumed, there are three groups to be compared, one independent variable, and one dependent variable, a one way ANOVA will
be used and evaluated at an alpha value=0.05.
e. Calculations

Figure 2a.

ANOVA
forcetoFractureBone
Sum of Mean
Squares df Square F Sig.
Between 11799.239 2 5899.619 9.256 .001
Groups
Within Groups 16571.983 26 637.384
Total 28371.222 28
f. Statistical decision
i. The ANOVA test showed that between the three groups, there is a F value of 9.256 at a significance value of .0001. At the .05 alpha, the difference between groups are significant.
(figure 2a)
ii. H0: there is no difference in bone strength between the young, middle aged, and elderly; Reject (p<.01)
iii. H1 : there is a difference in bone strength between the young, middle aged, and elderly. Cannot reject (p<.01)

g. Scientific conclusion
h. There is a difference in bone strength between the young (19-49), the middle-aged (50-69), and the elderly (70 and above)

3. Krugs Wallis
a. Scientific/Research Hypothesis-The purpose of this study is to look at whether there is a difference in pack years among the patients who have gone on the nicotine patch, bupropion
treatment, or both. The "pack years" is the average number of packs the subject smoked per day multiplied by the number of years the subject had smoked.
b. Test Hypotheses
i. H0: there is no difference in pack years between the three groups( nicotine patch, bupropion, both nicotine patch and bupropion)
ii. H1: there is no difference in pack years between the three groups( nicotine patch, bupropion, both nicotine patch and bupropion)
c. Test Data and Assumption
Bo Guan
Undieh Midterm

i. The data consists of three group of subjects (nicotine path, bupropion, both nicotine patch and bupropion) who reported their pack years as an estimate of their tobacco use.
Therefore, the independent variable is “group” category, and the dependent variable is the “pack years”. Since the patients self-assigned to treatment groups according to personal
preference and not random assignment, the pack years cannot be assumed to be normally distributed.
d. Test statistics and decision rule
i. The Kolmogorov Smirnov test showed the pack years to not be normally distributed. (3a) Given that there are three groups, one independent variable, one dependent variable, and
no normality, a Kruskal-Wallis test will be used.
e. Calculations

Figure 3a.

Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
packYears .068 164 .063 .978 164 .011
a. Lilliefors Significance Correction

Figure 3b.

Test Statisticsa,b
packYears
Kruskal-Wallis 11.664
H
df 2
Asymp. Sig. .003
a. Kruskal Wallis Test
b. Grouping Variable:
smokingCessationGroups

f. The calculated H value for Kruskal Wallis is 11.664 pack years with a significant figure of .003. At the .05 alpha, therefore, the difference between the three groups are significant.
Therefore: (figure 3b.)
i. H0: there is no difference in pack years between the three groups; Reject null hypothesis (p<.05)
ii. H1: there is no difference in pack years between the three groups. Cannot reject H1 since (p<.05)
g. There are differences in pack years among the different treatment groups.
Bo Guan
Undieh Midterm

4. Question 4
a. Scientific/Research Hypothesis-The purpose of this study is to look at whether a stress reduction program was effective in reducing anxiety patients’ stress as measured by the Hamilton
Anxiety Rating Scale. The hypothesis is that the scores should decrease from pre treatment to post treatment and decrease further three months after follow up.
b. Test Hypotheses t
i. H0: there is no difference between HARS scores between initial evaluation, pretreatment, post treatment, and 3 months follow up.
ii. H1: there is a difference between HARS scores from initial evaluation, pretreatment, post treatment, and 3 month follow up of the anxiety patients.
iii. H2: pretreatment>posttreatment
iv. H3: post treatment> 3 months follow up

c. Test Data and Assumption


i. The data consists of one group of patients who had their Hamilton Anxiety Rating Scale score taken at four different points in time including the initial evaluation, pretreatment,
post-treatment, and three-month follow up. . Therefore, the independent variable is the “time of measurement” (initial, pre, post, 3 months) and the dependent variable is the HARS
score. This scale is a continuous distribution within the studied population. The selection of patients was also random. Hence the dependent variable can be assumed to be normally
distributed.

d. Test statistics and decision rule


i. Given that there is one independent variable, one dependent variable, one group of subjects but multiple repeated tests on the same group, and normality of the HARS scale, a
repeated ANOVA is used at alpha of .05.
e. Calculations
Figure 4a.

Multivariate Testsa
Hypothesis Partial Eta
Effect Value F df Error df Sig. Squared
pointsinTime Pillai's Trace .857 21.996b 3.000 11.000 .000 .857
b
Wilks' Lambda .143 21.996 3.000 11.000 .000 .857
b
Hotelling's Trace 5.999 21.996 3.000 11.000 .000 .857
b
Roy's Largest 5.999 21.996 3.000 11.000 .000 .857
Root
a. Design: Intercept
Within Subjects Design: pointsinTime
Bo Guan
Undieh Midterm

Figure 4b.

Pairwise Comparisons
Measure: HARS
Mean 95% Confidence Interval for
(I) (J) Difference (I- Std. Differenceb
pointsinTime pointsinTime J) Error Sig.b Lower Bound Upper Bound
1 2 3.429 2.496 .193 -1.963 8.820
3 12.500* 2.064 .000 8.041 16.959
*
4 14.500 2.550 .000 8.991 20.009
2 1 -3.429 2.496 .193 -8.820 1.963
*
3 9.071 2.074 .001 4.591 13.552
4 11.071* 1.595 .000 7.626 14.517
*
3 1 -12.500 2.064 .000 -16.959 -8.041
*
2 -9.071 2.074 .001 -13.552 -4.591
4 2.000 1.997 .335 -2.315 6.315
4 1 -14.500* 2.550 .000 -20.009 -8.991
*
2 -11.071 1.595 .000 -14.517 -7.626
3 -2.000 1.997 .335 -6.315 2.315
Based on estimated marginal means
*. The mean difference is significant at the .05 level.
b. Adjustment for multiple comparisons: Least Significant Difference (equivalent to no adjustments).

f. Given that the Wilk’s Lambda multivariate test showed that the F value between the four
groups as 21.99 with a significance value of .000, there is a high significance in difference
between the four groups. (figure 4a)
i. H0: there is no difference between HARS scores between initial evaluation,
pretreatment, post treatment, and 3 months follow up. Reject because p<.05
ii. H1: there is a difference between HARS scores from initial evaluation, pretreatment,
post treatment, and 3 month follow up of the anxiety patients. Cannot reject because
p<.05

iii. H2: pretreatment>posttreatment; looking at the Pairwise Comparisons table, the mean
for pretreatment is greater than the mean for posttreatment by 9.071 at a P value
of .001. H2 cannot be rejected. (figure 4b.)
iv. H3: post treatment> 3 months follow up; looking at the Pairwise Comparison table, the
mean for post treatment is greater than three months follow up at a P value of .335.
Since P>.05, this hypothesis has to be rejected.
g. The stress reduction program is effective in reducing the Hamilton Anxiety Rating Scale
immediately after treatment but does not exhibit further effect after a three month follow up.
Bo Guan
Undieh Midterm

5. Quesition 5
a. The study is to investigate the effect of ovalbumin sensitization and exposure of air, benzaldehyde and acetaldehyde on guinea pigs’ alveolar cell count. The hypothesis is that there is a
difference between the sensitization groups and the treatment groups. Another hypothesis is that there is interaction between sensitization and treatment.
b. Test Hypotheses t
i. H0: there is no difference in alveolar count between the different treatment groups (acetaldehyde, air, and benzaldehyde) and the different ovalbumin sensitization groups(sensitized
and not sensitized)
ii. H1: there is a difference in alveolar count between the different treatment groups (acetaldehyde, air, and benzaldehyde) and the different ovalbumin sensitization groups(sensitized
and not sensitized)
iii. H2: No Ovalbumin(alveolar count) ≠ Yes Ovalbumin (alveolar count)
iv. H3: post treatment acetaldehyde (alveolar count) ≠ post treatment (benzaldehyde)
v. H4: There is interaction between ovalbumin sensitization and treatment.

c. Test Data and Assumption


i. There are two independent variables here, ovalbumin sensitization and treatment type. They are independent variables because guinea pigs are separated into groups according to
these parameters. Both of these influence on the dependent variable which is the alveolar count. Alveolar count is a continuous scalar within the studied population. The selection of
patients was also random. Hence the dependent variable can be assumed to be normally distributed.

d. Test statistics and decision rule


i. Given that there are two independent variables, one dependent variable, and normality of the alveolar count, a factorial Anova is used to test the hypothesis.
e.

Figure 5a.

Tests of Between-Subjects Effects


Dependent Variable: alveolarCount
Type III Sum Mean
Source of Squares df Square F Sig.
a
Corrected Model 17503.336 5 3500.667 10.193 .000
Intercept 74020.271 1 74020.271 215.530 .000
Oversensitized 7906.174 1 7906.174 23.021 .000
Treatment 7688.635 2 3844.318 11.194 .000
Oversensitized * 1908.527 2 954.263 2.779 .078
Treatment
Error 10303.013 30 343.434
Total 101826.620 36
Corrected Total 27806.349 35
a. R Squared = .629 (Adjusted R Squared = .568)
Bo Guan
Undieh Midterm

f. i. H0: there is no difference in alveolar count between the different treatment groups (acetaldehyde, air, and benzaldehyde) and the different ovalbumin sensitization groups(sensitized and
not sensitized). Reject because the factorial ANOVA showed that the F value for the overall model is 10.19 with a significant figure of .000. This is less than .05 and means that there is
significant difference across the groups.
i. H1: there is a difference in alveolar count between the different treatment groups (acetaldehyde, air, and benzaldehyde) and the different ovalbumin sensitization groups(sensitized
and not sensitized). Cannot reject because the factorial ANOVA showed that the F value for the overall model is 10.19 with a significant figure of .000. This is less than .05 and
means that there is significant difference across the groups.
ii. H2: No Ovalbumin(alveolar count) ≠ Yes Ovalbumin (alveolar count) Cannot reject because the factorial ANOVA showed that the F value for the ovalbumin sensitization is 23.02
with a significant figure of .000. This is less than .05 and means that there is significant difference across the groups.
iii. H3: post treatment acetaldehydet( alveolar count) ≠ post treatment (benzaldehyde) Cannot reject because the factorial ANOVA showed that the F value for treatment is 11.19 with a
significant figure of .000. This is less than .05 and means that there is significant difference across the groups.
iv. H4: There is interaction between ovalbumin sensitization and treatment. Reject because the factorial ANOVA showed that the F value for ovalbumin sensitization* treatment is 2.78
with a significant figure of .078. This is more than .05 and means that there is no significant difference across the groups.
g. There is difference among the ovalbumin sensitized and nonsensitized outcomes. There is difference between the three different exposures. However, there is no interaction.