www.pharmacyexam.

com © All Rights Reserved

RxExam’s Biostatistics
Questions & Answers
2019-2020 Edition

MANAN SHROFF

1
www.pharmacyexam.com © All Rights Reserved

This book is not intended as a substitute for the advice of physicians. Students or readers must consult their
physician about any existing problem. Do not use information in this book for any kind of self-treatment. Do
not administer any dose of mentioned drugs in this book without consulting your physician.

The author is not responsible for any kind of misinterpreted, incorrect, or misleading information or any
typographical errors in this book. Any doubtful or questionable answers should be checked in other
available reference sources.

All rights reserved.

No part of this book may be reproduced or transmitted in any form or by any means, electronically
Photocopying, recording, or otherwise, without prior written permission of the publisher.

RXEXAM® is a registered trademark of Pharmacy Exam of Krishna Publications Inc. Any unauthorized use of
this trade mark will be considered a violation of law.

2
www.pharmacyexam.com © All Rights Reserved

1. In biostatistics, confounding is normally defined 5. The scores of Naplex for 10 students are 75, 82,
as: 90, 92, 67, 95, 110, 80, 82, 86. Find the mean for the
above data.
a. One chosen from a carefully defined population
with the aid of a formal method to avoid bias. a. 83.6
b. 85.9
b. A formal method to assign subjects by chance to c. 836
one or the other treatment. d. 43

c. The effect of two or more variables that do not 6. The scores of Naplex for 10 students are 75, 82,
allow a conclusion about either one separately. 90, 92, 67, 95, 110, 80, 82, 86. Find the median for
the above data.
d. The systematic tendency of any factors
associated with the design, conduct, analysis, and a. 110
evaluation of the results of a trial to make the b. 84
estimate of a treatment effect deviate from its true c. 82
value. d. 67

2. The data that categories patients as males or 7. The scores of Naplex for 10 students are 75, 82,
females are known as: 90, 92, 67, 95, 110, 80, 82, 86. Find the mode for
the above data.
a. Random data
b. Nominal data a. 75
c. Ordinal data b. 90
d. Interval data c. 86
d. 82
3. Which of the following data represents interval
continuous data? 8. Classifying continuing educational experience
into categories including “strongly agree,”“agree,”
a. A number of cigarettes smoked per day by a and “disagree,” is an example of which type of
person. variable or data?
b. A number of children in a household.
c. Height of children. a. Nominal
d. Number of languages a person speaks. b. Ordinal
c. Interval
4. Data can be transformed by using the logarithm, d. Ratio
square root, or reciprocal. Which of the following is
the most common data transformation used in 9. Which of the following statements about the
medical research? prevalence is/are TRUE?

a. data converted to log I. It is a measure of the rate of occurrence of a

b. data converted to root condition.
c. data converted to reciprocal
d. data converted to ratio II. It is the number of new cases of a condition that
develop during a specific period of time.

III. It is the proportion of a population found to have

a condition at a single point in time.

3
www.pharmacyexam.com © All Rights Reserved

a. I only c. the risk of heart failure in patients taking the

b. III only study drug was more than the risk of heart failure in
c. II and III only patients taking placebo.
d. All
14. What is the value of relative risk reduction if an
10. “One out of every 50 adults in the United States RR is 0.33?
has depression.” This statement is an example of:
a. 1.56
a. Incidence b. 0.33
b. Prevalence c. 0.67
d. 1.33
11. Which of the following statements best describe
the relative risk (RR)? 15. What would be the absolute risk reduction if the
incidence of heart attack was 2% in the intervention
a. It is the risk of an event or outcome occurring in a group and 7% in the control group?
group of interest in relation to a control group.
a. 15%
b. It usually states how much the treatment b. 5%
reduced the risk of an outcome relative to the c. 2%
control group. d. 7%

c. It represents the absolute difference of the event 16. If the calculated absolute risk reduction (ARR)
rate between the treatment and control groups. for heart attack event was 5.1%, what would be the
number needed to treat?
d. It is a representation of the number of patients
who need to be treated to prevent one additional a. 11
event compared to treating the same number of b. 2.65
patients with the control therapy. c. 3.2
d. 20
12. Calculate the value of RR if the risk of heart
failure associated with an invention drug is 5% 17. Which of the following statements is TRUE
versus 9% with placebo? ABOUT the standard deviation (SD)?

a. 1.25 a. It is used to calculate confidence intervals.

b. 2.65
c. 0.55 b. It is used to determine variability of a sample
d. 1.8 around a sample mean.

13. Which of the following statements best describe c. It is used to display bimodal or skewed data.
the interpretation of value of RR obtained in a
previous question? d. A large SD shows that individual data points are
clustered closer to the mean.
a. there was no association between the study drug
and heart failure. 18. If the sample size in a study is 100 subjects and
the SD for blood glucose is 10 mg/dl, what is the
b. the risk of heart failure in patients taking the standard error of the mean?
study drug was less than the risk of heart failure in
patients taking placebo. a. 65.9

4
www.pharmacyexam.com © All Rights Reserved

d. the regression model explained 30% of the total b. Employees who exposed to asbestos for more
variance is not a good fit. than 10 years had 16% decreased odds of
developing asthma than those who did not.
146. In a study examining the relationship between
the drinking a coffee in the late evening and the c. The results are not significant because the
likelihood of insomnia, the r2 value for the confidence interval includes 1.0.
regression line is 1.0. What does this indicate?
d. The results are not significant because the
a. For each additional cup of coffee, chances of confidence interval is greater than 1.0.
occuring insomnia increase by 1.0%.
149. What does “OR 0.4 95%CI 0.4-0.6 p <0.05”
2
b. Since r equals 1.0, this indicates there is no mean?
relationship between drinking coffee and
developing insomnia. a. The odds of death in the intervention groups are
60% less than the odds of death in the control
c. Since r2 equals 1.0, the regression line perfectly groups with the true population effect between
fits the data. 60% and 40%. This result was statistically
significant.
d. For each additional cup of coffee, there is no
effect on sleep. b. The odds of death in the intervention groups are
60% more than the odds of death in the control
147. Which of the following information is/are TRUE groups with the true population effect between
ABOUT an odds ratio (OR)? 60% and 40%. This result was statistically
significant.
I. If the OR = 1 indicates there is no difference
between the two arms of the study. c. The odds of death in the intervention groups are
60% less than the odds of death in the control
II. If the OR is > 1 the control is better than the groups with the true population effect between
intervention. 40% and 60%. This result was statistically NOT
significant.
III. If the OR is < 1 the intervention is better than the
control. d. The odds of death in the intervention groups are
60% more than the odds of death in the control
a. I only groups with the true population effect between
b. I and II only 40% and 60%. This result was statistically NOT
c. II and III only significant.
d. All
150. A drug company-funded double blind
148. Suppose a study examined the relationship randomised controlled trial evaluated the efficacy
between workers exposed to an asbestos for more of an adenosine receptor antagonist Cangrelor vs
than 10 years and the risk of developing asthma. If Clopidogrel in patients undergoing urgent or
the study found an increased risk of asthma in the elective Percutaneous Coronary Intervention (PCI)
group who exposed to asbestos (OR: 1.16, 95% CI: who were followed up for specific complications for
0.9-1.5). What does this mean? 48 hrs (Bhatt et al. 2009). The results section
reported “OR 0.65 95% confidence interval [CI],
a. Employees who exposed to asbestos for more 0.55 to 0.83; P=0.005” What does this mean?
than 10 years were 1.16 times less likely to develop
asthma than those who did not.

21
www.pharmacyexam.com © All Rights Reserved

a. The odds of death, myocardial infarction,

ischemia-driven revascularization, or stent
thrombosis at 48 hours after randomization in the
Cangrelor arm were 35% less than in the
Clopidogrel arm with the true population effect
between 34% and 7%. This result was not
statistically significant.

b. The odds of death, myocardial infarction,

ischemia-driven revascularization, or stent
thrombosis at 48 hours after randomization in the
Cangrelor arm were 35% less than in the
Clopidogrel arm with the true population effect
between 45% and 17%. This result was statistically
significant.

c. The odds of death, myocardial infarction,

ischemia-driven revascularization, or stent
thrombosis at 48 hours after randomization in the
Cangrelor arm were 35% less than in the
Clopidogrel arm with the true population effect
between 45% and 17%. This result was statistically
NOT significant.

d. The odds of death, myocardial infarction,

ischemia-driven revascularization, or stent
thrombosis at 48 hours after randomization in the
Cangrelor arm were 35% less than in the
Clopidogrel arm with the true population effect
between 34% and 7%. This result was statistically
significant.

22
www.pharmacyexam.com © All Rights Reserved

Biostatistics Answers

1. (c) The effect of two or more variables that do “names” or categories based on the presence or
not allow a conclusion about either one separately absence of certain attributes/characteristics
is defined as confounding. without any ranking between the categories. For
example, patients are categorized by gender as
Random sample: one chosen from a carefully males or females; by religion as Hindu, Muslim, or
defined population with the aid of a formal method Christian. It also includes binominal data, which
to avoid bias and confounding. refers to two possible outcomes. For example,
outcome of cancer may be death or survival, drug
Randomization: in comparative trials, a formal therapy with drug ‘X’ will show improvement or no
method to assign subjects by chance to one or the improvement at all.
other treatment.
2. Ordinal data: It is also called as ordered,
Random errors: errors that follow the principle of categorical, or graded data. Generally, this type of
indifference. data is expressed as scores or ranks. There is a
natural order among categories, and they can be
Bias: The systematic tendency of any factors ranked or arranged in order. For example, pain
associated with the design, conduct, analysis, and associated with cancer may be classified as mild,
evaluation of the results of a trial to make the moderate, and severe. Since there is an order
estimate of a treatment effect deviate from its true between the three grades of pain, this type of data
value. is called as ordinal. To indicate the intensity of pain,
it may also be expressed as scores (mild = 1,
Confounding: The effect of two or more variables moderate = 2, severe = 3). Hence, data can be
that do not allow a conclusion about either one arranged in an order and rank.
separately.
3. Interval data: This type of data is characterized
Validity: (a). In logic, an argument is valid if the by an equal and definite interval between two
conclusions follow from the premises measurements. For example, weight is expressed as
100, 110, 120, 130, 140 lbs. The interval between
(b). In pharmacological sciences, a method is valid if 100 and 110 is same as that between 130 and 140.
it measures what it should, is reproducible and
responsive to change, e.g., by a treatment. 3. (c) Height of children.

2. (b) Nominal data. This type of data is characterized by an equal and

definite interval between two measurements. For
Observations recorded during research constitute example, weight is expressed as 100, 110, 120, 130,
data. There are three types of data i.e. nominal, 140 lbs. The interval between 100 and 110 is same
ordinal, and interval data. as that between 130 and 140.

Statistical methods for analysis mainly depend on Interval type of data can be either continuous or
type of data. Generally, data show picture of the discrete. A continuous variable can take any value
variability and central tendency. Therefore, it is very within a given range.
important to understand the types of data.
For example: hemoglobin (Hb) level may be taken
1. Nominal data: This is synonymous with as 11.3, 12.6, 13.4 gm % while a discrete variable is
categorical data where data is simply assigned usually assigned integer values i.e. does not have

1
www.pharmacyexam.com © All Rights Reserved

fractional values. For example, blood pressure 6. (b) 84.

values are generally discrete variables or number of
cigarettes smoked per day by a person. Median is an average, which is obtained by getting
middle values of a set of data arranged or ordered
Examples of Interval Continuous Data: from lowest to the highest (or vice versa). In this
process, 50% of the population has the value
1. Height of children smaller than median and 50% of samples have the
2. Weight of cars value larger than median.
3. Time to wake up in the morning
Median is a better indicator of central value when
Examples of Interval Discrete Data: one or more of the lowest or the highest
observations are wide apart or are not evenly
1. A number of cigarettes smoked per day by a distributed.
person.
2. A number of children in a household. Median in case of even number of observations is
3. A number of languages a person speaks. taken arbitrary as an average of two middle values
and in case of odd number, the central value forms
4. (a) data converted to log. the median.

Sometimes, certain data may be converted from To find median in above data, we need to arrange
one form to another form to reduce skewness and the data either ascending or descending. Therefore:
make it to follow the normal distribution. For
example, drug doses are converted to their log 67, 75, 80, 82, 82, 86, 90, 92, 95, 110
values and plotted in dose response curve to obtain
a straight line so that analysis becomes easy. Data In case of even numbers, the median should be the
can be transformed by taking the logarithm, square average of two middle values. Therefore, in above
root, or reciprocal. Logarithmic conversion is the case, it should be:
most common data transformation used in medical
research. (82+86)
= 2
5. (b) 85.90
= 84
Mean is the common measure of central tendency,
If we remove the last data set 110 from the
most widely used in calculations of averages. It is
calculation, the median in case of odd number (9)
the average of a data set. So, in the above example,
should be the central value of the data set:
the mean of the data can be calculated by adding
up the individual values and dividing the sum by
67, 75, 80, 82, 82, 86, 90, 92, 95
number of students (n).
= 82
The sample mean 𝑥̅ is calculated as

𝑥1+ 𝑥2+ 𝑥3+ 𝑥4+..𝑥𝑛 𝑥𝑖 7. (d) 82

𝑥̅ = = ∑𝑛
𝑛
The mode is the number that occurs most often
75+82+90+92+67+95+110+80+82+86 within a set of numbers or is the most common
𝑥̅ = 𝑀𝑒𝑎𝑛 = 10 number in a data set. In above case the mode
should be 82 since it appears twice in data set. To
Mean = 85.90
get better idea, look the following data set:

2
www.pharmacyexam.com © All Rights Reserved

75, 82, 75, 92, 67, 95, 75, 80, 82, 86. Prevalence is measured using cross-sectional study
designs, such as a survey or a census, or by using
In this data set, the mode will be 75 and not 82 administrative data such as medical, hospital, or
since it appears three times. prescription drug claims.

A set of numbers can have more than one mode 10. (b) This statement is an example of prevalence
(this is known as bimodal) if there are multiple since the statement is an expression of a proportion
numbers that occur with equal frequency, and more of a population found to have a condition at a single
times than the others in the set. point in time.

75, 82, 75, 92, 67, 95, 75, 80, 82, 86, 82. 11. (a) The relative risk (RR) is the risk of an event or
outcome occurring in a group of interest in relation
In this example, both the number 75 and the to a control group. The group of interest can be an
number 82 are modes; since they both appear three intervention (e.g., a medication) or a pre-specified
times. characteristic e.g., those with hypertension).

If no number in a set of numbers occurs more than An example would be the risk of heart attack in
once, that set has no mode: patients taking a study drug versus a control group
of patients taking a placebo.
For example: 75, 82, 85, 90, 92
The relative risk (RR) can be calculated by dividing
8. (b) Classifying continuing educational experience the event rate (i.e., heart attacks) in the
into categories is a categorical variable as the value intervention group by the event rate (i.e., heart
functions as a label rather than a numeric value. attacks) in the control group.
The type of categorical variable is ordinal since
categories can be ranked in a specific order. 12. (c) 0.55.

9. (b) III only. The relative risk (RR) can be calculated by using the
following formula:
Incidence is a measure of the rate of occurrence of
a condition. In other words, incidence is the number (𝑒𝑣𝑒𝑛𝑡 𝑟𝑎𝑡𝑒 𝑖𝑛 𝑖𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛 𝑔𝑟𝑜𝑢𝑝)
𝑅𝑅 =
of new cases of a condition that develop during a 𝑒𝑣𝑒𝑛𝑡 𝑟𝑎𝑡𝑒 𝑖𝑛 𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑔𝑟𝑜𝑢𝑝
specific period of time. For example, the incidence
5
of hypertension may be documented as the number 𝑅𝑅 = = 0.55
9
of new cases of hypertension per year within a
given sample or population. 13. (b) The relative risk (RR) is the risk of an event or
outcome occurring in a group of interest in relation
Incidence can be measured using methods such as to a control group. The group of interest can be an
randomized controlled trials (e.g., to determine the intervention (e.g., a medication) or a pre-specified
incidence of adverse effects of a medication) or characteristic (e.g., those with hypertension).
cohort studies, which follow groups of patients over
time. An example would be the risk of heart failure in
patients taking a study drug versus patients taking
Prevalence, on the other hand, is the proportion of placebo. The RR can be calculated by dividing the
a population found to have a condition at a single event rate (i.e., heart failure) in the intervention
point in time. For example, the prevalence of group by the event rate in the control group.
hypertension may be documented as the
percentage of the US population with hypertension.

3
www.pharmacyexam.com © All Rights Reserved

1. If the RR is 1, there is no difference between the The number needed to treat (NNT) is a
two groups; or, in other words, there is no representation of the number of patients who need
association between exposure to a factor and the to be treated to prevent one additional event
outcome of interest (e.g., there was no association compared to treating the same number of patients
between the study drug and heart failure). with the control therapy.

2. An RR less than 1 indicates a negative association It can be easily calculated as it is the inverse of the
(e.g., an RR of 0.55 - we found in question 11 - ARR. The NNT should always be rounded up to the
means the risk of heart failure in patients taking the nearest integer, as you cannot have a fraction of a
study drug was less than the risk of heart failure in patient (e.g., an NNT of 1.6 would be rounded up to
patients taking placebo). an NNT of 2).

3. While an RR greater than 1 indicates a positive 1

𝑁𝑁𝑇 =
association (e.g., an RR of 2 means the risk of heart 𝐴𝑅𝑅

failure in patients taking the study drug was greater 1

than the risk of heart failure in patients taking 𝑁𝑁𝑇 = 0.051
= 19.60 = 20
placebo).
The NNT in this example demonstrates that for
14. (c) 0.67 every 20 patients treated, 1 additional heart attack
can be avoided compared to treating those patients
The relative risk reduction (RRR) is another measure with the control therapy.
of effect size, which reports how much the
treatment reduced the risk of an outcome relative 17. (b) It is used to determine variability of a sample
to the control group. The RRR can be calculated by around a sample mean.
subtracting the relative risk (RR) from 1.
1. Standard deviations (SD) determine the
The Relative Risk Reduction (RRR) can be calculated approximate distribution of individual data points in
by the following formula: regard to a sample mean; therefore, the SD is a
measure of the variability in the data.
RRR = 1 - Relative Risk (RR)
2. Typically, a study will find a range of individual
RRR = 1 - 0.33 = 0.67 values or observations, where some values fall
either below the mean or above the mean. Thus,
15. (b) 5%. individual deviations from the mean can be either
positive or negative.
ARR represents the absolute difference of the event
rate between the treatment and control groups. 3. Variance is calculated as the mean of the
The ARR can be easily calculated by subtracting the deviations squared, and the SD is calculated as the
event rate in the intervention group from the event square root of the variance.
rate in the control group.
4. The SD provides value to readers when the
ARR = the event rate in the control group - the sample is either normally or nearly normally
event rate in the intervention group distributed and is used for continuous variables.

ARR = 7% - 2% = 5% 5. If the SD is reported for bimodal or skewed data,

it may be falsely larger than expected. Therefore, it
16. (d) 20 would not be an appropriate tool to use to measure
the variability for those distributions.

4
www.pharmacyexam.com © All Rights Reserved

margin been utilized, the treatment would have 116. (a) A categorical variable (sometimes called a
been found to be “non-inferior” if the event rates nominal variable) is one that has two or more
were as high as 8.75%. categories, but there is no intrinsic ordering to the
categories. For example, gender is a categorical
On the other hand, overly stringent NI margins can variable having two categories (male and female)
prevent an effective therapy from being deemed and there is no intrinsic ordering to the categories.
non-inferior. Hair color is also a categorical variable having a
number of categories (blonde, brown, brunette,
113. (b) After a non-inferiority clinical trial, a new red, etc.) and again, there is no agreed way to order
therapy may be accepted as effective, even if its these from highest to lowest. A purely categorical
treatment effect is slightly lesser than the current variable is one that simply allows you to assign
standard. It is therefore possible that, after a series categories, but you cannot clearly order the
of trials where the new therapy is slightly worse variables. If the variable has a clear ordering, then
than the preceding drugs, an ineffective or harmful that variable would be an ordinal variable, as
therapy might be incorrectly declared efficacious; described below.
this is known as 'bio-creep'.
An ordinal variable is similar to a categorical
To provide an example, suppose that a theoretical variable. The difference between the two is that
Drug A is deemed to be superior to placebo. Several there is a clear ordering of the variables. For
years later, Drug B is found to be non-inferior to example, suppose you have a variable, economic
drug A with a certain NI margin. Then, Drug C is status, with three categories (low, medium and
compared to Drug B via a NI study at some point high). In addition to being able to classify people
later. During each of these studies, the new agent is into these three categories, you can order the
found to be “not acceptably” worse than the categories as low, medium and high. Now consider
previous agent. It may then be unknown if Drug C is a variable like educational experience (with values
more effective than placebo. such as elementary school graduate, high school
graduate, some college and college graduate).
114. (d) The Chi-Squared Test is used to test for These also can be ordered as elementary school,
statistical significance of a dependent categorical high school, some college, and college graduate.
variable between two or more independent groups. Even though we can order these from lowest to
It is used to determine whether observed sample highest, the spacing between the values may not be
frequencies differ significantly from expected the same across the levels of the variables. If these
frequencies specified in the null hypothesis. categories were equally spaced, then the variable
would be an interval variable.
115. (c) The only assumption for the Chi-Squared
Test is that all cells must have an expected value of An interval variable is similar to an ordinal variable,
at least five. except that the intervals between the values of the
interval variable are equally spaced. For example,
The chi-square goodness of fit test is appropriate suppose you have a variable such as annual income
when the following conditions are met: that is measured in dollars, and we have three
people who make $50,000, $60,000 and $70,000.
1. The sampling method is simple random sampling. The second person makes $10,000 more than the
2. The variable under study is categorical. first person and $10,000 less than the third person,
3. The expected value of the number of sample and the size of these intervals is the same. If there
observations in each level of the variable is at least were two other people who make $100,000 and
5. $110,000, the size of that interval between these
two people is also the same ($10,000).

33
www.pharmacyexam.com © All Rights Reserved

117.(c) Stroke is a categorical dependent variable 55 x 5

Expected Cell = = 2.29
and therefore Chi Squared Test or Fischer’s Exact 120
Test can be used. The Chi Squared Test can only be
used when all cells must have an expected value of D. Control Group Without Stroke
at least five. In cases where the expected value is
not at least five for all cells, a Fischer’s Exact Test 55 x 115
Expected Cell = = 52.70
should be used. 120

The formula for calculating expected cell value in So, placing all values in above table, it should look
Chi Squared Test is: like:

Row Total x Column Total Stroke No Stroke Total

Expected Cell =
Total number subjects
Intervention 2 63 65
In above example, we can calculate it by the Group
following way:
Expected 2.70 62.29
Cell Value
Stroke No Stroke Total
Control 3 52 55
Intervention 2 63 65 Group
Group
Expected 2.29 52.70
Expected ? ?
Cell Value
Cell Value
Total 5 115 120
Control 3 52 55
Group

Expected ? ? There were two cells (e.g. 2.70 and 2.29) in our
Cell Value calculation have expected cell value less than 5.
Therefore, this data does not meet the chi-squared
Total 5 115 120 assumption for at least five expected in each cell.
So, the Fischer’s Exact Test shall be used instead.

As shown in the above table, we need to find values 118. (c) The degrees of freedom for a Chi-squared
for each expected cell denoted with “?”. Using the Test can be calculated by using the following
formula, we can calculate expected cell value in formula:
intervention and control groups.
Degrees of Freedom = (Rows – 1) x (Columns – 1)
A. Intervention Group With Stroke
Therefore, the degrees of freedom in 2 x 3
65 x 5 contingency chi squared test should be:
Expected Cell = = 2.70
120
DF = (R – 1)x(C – 1)
B. Intervention Group Without Stroke DF = (2-1)x(3-1)
DF = 1 x 2 = 2
65 x 115
Expected Cell = = 62.29
120 119. (a) HbA1c is a continuous dependent variable.
As we are testing two independent groups and it
C. Control Group With Stroke
meets the general rule of thumb having at least 30

34
www.pharmacyexam.com © All Rights Reserved

subjects in each group, the t-test is the best option. To calculate expected value for each cell, we need
to use the following formula:
Continuous variable: A continuous variable has
numeric values, and the relative magnitude of the Row Total x Column Total
Expected Cell = Total number subjects
values is significant. This means that data can be
ranked low to high or high to low. An example of a
continuous variable is value of HbA1c in a diabetic 49 x 32
patient. Expected Cell = 100
= 15.68

49 x 27
Discrete variable: Suppose we flip a coin and count Expected Cell = = 13.23
100
the number of tails. The number of tails could be
any integer value between 0 and plus infinity. We 49 x 41
Expected Cell = = 20.09
could not, for example, get 4.6 tails. 100

51 x 32
120. (a) Variables are classified as either Expected Cell = 100
= 16.32
independent or dependent variables. In research
studies, the independent variable is the variable 51 x 27
that is varied or manipulated by the researcher. Expected Cell = = 13.77
100

For example, patients in a clinical trial may be 51 x 41

Expected Cell = = 20.91
assigned to take a placebo or an anti-diabetic 100
medication. The type of study medication that the
patient is taking is an independent variable. The Expected Value
dependent variable is the outcome or response that Biology Physics Chemistry Total
is measured, such as blood glucose level, in this Male 15.68 13.23 20.09 49
example. Researchers are typically interested in Female 16.32 13.77 20.91 51
understanding the effect of the independent Total 32 27 41 100
variable (or treatment) on the dependent variable
(or outcome); for example, what is the effect of an Now, we can construct the null hypothesis.
anti-diabetic medication on blood glucose level?
Ho: Gender has NO influence on a subject selected.
121. (b) To do a Chi Square statistic calculation, we Ha: Gender has influence on a subject selected.
first want to find out values for each expected cell;
it must be greater than 5. Degrees of freedom = (Rows-1)x(Columns-1)
Degrees of freedom = (2-1)x(3-1) = 2
Observed Value
Biology Physics Chemistry Total The Chi Square value at 0.05 level of significance
Male 12 16 21 49 with a degree of freedom equals 2 would be 5.991.
Female 20 11 20 51 Please refer to Biostatistics Table file.
Total 32 27 41 100

Expected Value
Biology Physics Chemistry Total
Male ? ? ? ?
Female ? ? ? ?
Total ? ? ? ?

35
www.pharmacyexam.com © All Rights Reserved

The next step is to find out the Chi square test value 122. (d) The pain scores are an ordinal variable, so a
to compare it with our previously obtained critical non-parametric test must be used. If the data is part
value. The formula to find the Chi Square test value of one single group, Wilcoxon signed rank test shall
is: be used. If the data is part of two independent
groups, Mann Whitney U test shall be used. For
(𝑜𝑏𝑠𝑒𝑟𝑣𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 − 𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑣𝑎𝑙𝑢𝑒)2 three or more independent groups, Kruskal Wallace
ꭓ2 = ∑ test should be used.
𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑣𝑎𝑙𝑢𝑒

Therefore: In our example data is a pre/post comparison of

paired data of the same single group (same 100
(12−15.68)2 (20−16.32)2 (16−13.23)2 patients are used before and after administering
ꭓ2 = + + +
15.68 16.32 13.23 Oxycodone); therefore, the Wilcoxon Signed-Rank
test is the best option.
(11−13.77)2 (21−20.09)2 (20−20.91)2
+ +
13.77 20.09 20.91
123. (c) For the continuous dependent variables, t-
(−3.68)2 (3.68)2 (2.77)2 (−2.77)2 test, paired t-test or ANOVA can be used depend on
= + + + the presence of independent groups. If there is a
15.68 16.32 13.23 13.77
single independent group, paired t-test shall be
(0.91)2 (−0.91)2 used. If there are two independent groups, then,
+ + independent groups t-test shall be performed. For
20.09 20.91
three or more independent groups, ANOVA should
13.54 13.54 7.67 7.67 0.82 be used.
= + + + +
15.68 16.32 13.23 13.77 20.09
124. (d) All. Analysis of Variance or ANOVA is used
0.82 to test for statistical significance of a continuous
+
20.91 dependent variable between three or more
independent groups assuming the data are normally
= 0.8635 + 0.8296 + 0.5797 + 0.5570 + 0.0408 + distributed.
0.0392
When a statistically significant p-value is found
= 2.9098 using ANOVA, it means that there is a statistically
significant difference between two or more of the
Since Chi Squared value 2.9098 is less than our groups. A student’s t-test can then be used to tease
critical value (5.991); we are failed to reject the null out exactly where the differences are.
hypothesis. In other words; a gender does NOT
have any influence on the subject selected. ANOVA can be used for a comparison between two
groups, but the student’s t-test will give the same
answer and is much easier to calculate. In its
simplest form, ANOVA provides a statistical test of
whether the population means of several groups
are equal, and therefore generalizes the t-test to
more than two groups. ANOVA is useful for
comparing (testing) three or more group means for
statistical significance. It is conceptually similar to
multiple two-sample t-tests, but is more
conservative, resulting in fewer type I errors, and is
therefore suited to a wide range of practical
problems.

36
www.pharmacyexam.com © All Rights Reserved

Clopidogrel in patients undergoing urgent or

elective Percutaneous Coronary Intervention (PCI)
who were followed up for specific complications for
48 hrs (Bhatt et al. 2009). The results section
reported “OR 0.65 95% confidence interval [CI],
0.55 to 0.83; P=0.005” This means:

The odds of death, myocardial infarction, ischemia-

driven revascularization, or stent thrombosis at 48
hours after randomization in the Cangrelor arm
were 35% (1-0.65=0.35) less than in the Clopidogrel
arm with the true population effect between (1-
0.55=0.45) 45% and (1-0.83=0.17) 17%. This result
was statistically significant (p(0.005) < 0.05) since p
< 0.05.

42