Short Answers on Research Study

This document provides short answers to questions related to a research study. It covers the research question, hypothesis, independent and dependent variables, study design, statistical tests used, and limitations of the study.

1
1. (12 points) Short Answers
Q. Point
s
a. State the research
question behind this
study.
The research question behind this study was
based on the examination of the effect of health
warning levels (HWL) applied pictorially on the
waterpipe (WP) device on smoker’s
experience, behavior underlying puffing,
perception of harm and exposure to respiratory
toxicants (Maziak et al., 2019).
2
b. Clearly state the
hypothesis of the study, in
the form of a null and an
alternative hypothesis.
Null hypothesis: The addition of pictorial HWL
on the WP will result in no reductions in the
puffing parameters, exposure to carbon
monoxide and positive experiences among
smokers, implying that, there is a statistically
insignificant relationship between pictorial
HWL and smoking behaviors.
Alternative hypothesis: The addition of
pictorial HWL on the WP will result in
signficant reductions in the puffing parameters,
exposure to carbon monoxide and positive
experiences among smokers, implying that,
there is a statistically significant relationship
between pictorial HWL and smoking behaviors
(Maziak et al., 2019).
4
c. What is/are the independent
(exposure) and dependent
(outcome) variables in this
study?
The independent, exposure variables in the
study were the inclusion of pictorial HWL on
the WP of the smoking devices. The dependant
outcome variables for the study were: the
amount of carbon monoxide exhaled, puff
topography (puff durations and intervals
between puffs, total puff volume), smoking
experiences (satisfaction, taste, puff) and harm
perception (Maziak et al., 2019).
6
2.( 8 points) Classify the following variables according
to their type:
Quantitative/Qualitative/Nominal/Ordinal/interview/Ratio/binary/
polychotomous as applies

2
Perception of
Harm
Quantitative Variable: Ordinal variable (ranging from 1 to 7 point
scale – extremely harmful – not at all harmful).
Puff duration Quantitative variable - no ordering or categorization in variables
Total volume
inhaled
Quantitative Variable - no ordering or categorization in variables
Number of puffs Quantitative Variable - no ordering or categorization in variables

3
2. The design of this trial was a crossover design:
2a. (2 points) How would you repeat the same trial with a parallel design?
A parallel study is characterized by randomly assigning subjects to separate
groups where one group will receive a specific type of intervention (for
example intervention X) while the other group will receive another separate
type of intervention (for example intervention Y). Hence, considering the
study, instead of the same group of subjects receiving a multiple forms of
intervention in the form of a cross over design, a parallel design will involve
randomizing the subjects into two separate groups: one receiving one type
of HWL while the other group receiving a different type of HWL.
Additionally, control group will be required who will not be receiving the
HWL for the purpose of comparing the HWL effects on smoking behavior
(Grieve, 2016).
2b. (4 points) What types of considerations would you make to
minimize the effects of bias and confounding.
Participants with varied demographic characteristics such as
both males or females, or hailing from various educational
backgrounds must be incorporated to prevent generalizations
and prevent the effects of confounding variables such as
gender, education, employment etc. Further participants must
be blinded to the interventions of the study in order to prevent
biased responses. Researches must also be concealed and
blinded to the treatment and allocation sequences utilized
during intervention administration in order to prevent
researcher bias (Heale & Twycross, 2015).
2c. (4 points) If your study were indeed conducted as a parallel
trial, would you require the same/smaller/larger number of
participants for the same level of precision obtained in this
study? Justify your answer.
For the conductance of a the parallel design, a larger sample size
may be required during random allocation. This is due to the fact
that a smaller sample size in parallel designs such as
randomized controlled trials is more prone to bias and
generalization resulting in invalid results. Further, a larger
sample size ensures incorporation of varied participant
characteristics resulting in reduced confounding and increased
applicability in real-life scenarios (Bonett & Wright, 2015).
3. The authors state that “the study was not powered to detect
differences in outcome measures between the two HWLs”:

4
3a. (2 points) Explain in your own words what this statement means.
From the given statement it means that, due to the usage of a relatively small
sample size, the authors were unable to include a wider variety of variable
measurements and hence resulted in the reduced ability of the study to detect
variable outcomes which should have been detected, that is the level of intentions
of quitting among the participants. Hence, this implies that the authors were
unable to evaluate the presence of statistically significant relationships between
the outcomes exerted by the first HWL (‘Using charcoal to heat produces Cancer’)
and between the outcomes exerted by the second HWL (‘Chemicals in hookah
causes Cancer’) (George et al., 2016).
3b. (2 points) What type of error would most likely result from
the power limitation of this study?
As observed, the power limitation of this study, was due to the inclusion of a relatively
small sample size. A small sample size incorporated in a study, decreases its ability to
avoid statistical errors categorized as ‘Type II’. A error which is notified to be Type II
when the research is being able to prove the significance of the null hypothesis,
whereas, in reality, the alternative hypothesis holds true. This is due to the fact that a
reduced sample size results in a loss of the power of the study, which implies the ability
of the study to adequately identify significant outcomes and effects which actually
should be identified. A reduced sample size results in skewing of the data, further
resulting in incorrect outcome measurement, and proving of a wrong hypothesis
(Howard, 2019).
4. Looking at Table 2:
4a. (2 points) What statistical test was used to compare
topography parameters with and without HWL
The authors first obtained mean values of the puff topography
measures in both the groups receiving interventions of with
and without HWL. The statistical test used later to compare the
means between the two groups was ‘t-test’.
4b. (2points) Briefly explain why this particular test was
appropriate to use
A t test is used to assess the presence of statistical significance
between the means of two different groups receiving two
different intervention methods. Hence, a t test is important since
the presence of a statistically significant difference between the

5
two means obtained from it, implies that the two sample groups
are truly ‘different’ in terms of the results obtained from
differential interventions (Kim, 2015).
4c. (2points) Explain the meaning of the p-value comparing puff
duration
with and without HWL.
The p value for the t test was measured in terms of a significance level of
0.05. This means, that if the p value obtained is less than 0.05, then there is
a statistically significance difference between the two groups. As per the
results, the p value in puff duration was measured to be less than 0.05. This
implies that the group receiving HWL was statistically significantly different
from that receiving no HWL meaning that there truly lies a statistically
significant relationship between administration of HWL and smoking
behaviors (Delacre, Lakens & Leys, 2017).

6
References
Bonett, D. G., & Wright, T. A. (2015). Cronbach's alpha reliability: Interval
estimation, hypothesis testing, and sample size planning. Journal of
Organizational Behavior, 36(1), 3-15.
Delacre, M., Lakens, D., & Leys, C. (2017). Why psychologists should by
default use Welch’s t-test instead of Student’s t-test. International
Review of Social Psychology, 30(1).
George, B. J., Beasley, T. M., Brown, A. W., Dawson, J., Dimova, R., Divers,
J., ... & Kim, M. Y. (2016). Common scientific and statistical errors in
obesity research. Obesity, 24(4), 781-790.
Grieve, A. P. (2016). Crossover versus parallel designs. In Statistical
methodology in the pharmaceutical sciences (pp. 253-284). CRC Press.
Heale, R., & Twycross, A. (2015). Validity and reliability in quantitative
studies. Evidence-based nursing, 18(3), 66-67.
Howard, J. (2019). Research Errors. In Cognitive Errors and Diagnostic
Mistakes (pp. 537-576). Springer, Cham.
Kim, T. K. (2015). T test as a parametric statistic. Korean journal of
anesthesiology, 68(6), 540.
Maziak, W., Taleb, Z. B., Kalan, M. E., Eissenberg, T., Thrasher, J., Shihadeh,
A., & Asfar, T. (2019). Pictorial health warning labels on the waterpipe
device are effective in reducing smoking satisfaction, puffing
behaviour and exposure to CO: first evidence from a crossover clinical
laboratory study. Tobacco control, tobaccocontrol-2018.