Skip to main content

 Mariam

 0

 My courses

 Home

 HiYield Paper B

 Statistics

 EMIs Statistics

HiYield Paper B
Started on Tuesday, 31 May 2016, 7:18 AM

State Finished

Completed on Tuesday, 31 May 2016, 7:19 AM

Time taken 1 min 2 secs

Marks 0.00/94.00

Grade 0.00 out of 100.00

Question 1
Not answered
Marked out of 3.00
 Flag question

Question text

HiY AdvStats EMI001

Multivariate techniques
Choose one option from the given list for each description below:

A researcher wants to analyze the various

chains of influence existing among degree of Answer 1
poverty, social isolation, unemployment, Choose...
depression and physical morbidity.
A researcher wants to predict the value that a
Answer 2
binary dependent variable can take using an
Choose...
independent variable and a constant.
Within subjects analysis for paired Answer 3
observations of means in 5 different groups. Choose...

Check

Feedback

Your answer is incorrect.

Explanation:
When the term 'chain of causation' is used, path analysis is often a suitable
technique.
To make predictions, we need regression models. As a binary dependent variable
is employed, logistic regression is the most appropriate.
Repeated measures ANOVA can be used to compare paired observations within
subjects across multiple groups. Here one will model the groups (5 groups) as a
between subjects factor and time of observation (before and after) as the within-
subjects factor.
The correct answer is:
A researcher wants to analyze the various chains of influence existing among
degree of poverty, social isolation, unemployment, depression and physical
morbidity.
– Path analysis,
A researcher wants to predict the value that a binary dependent variable can
take using an independent variable and a constant.
 – Logistic regression,
Within subjects analysis for paired observations of means in 5 different groups.
– Repeated measures ANOVA

Question 2
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI002

Univariate statistics
Choose one option from the given list for each description below:

A researcher wants to assess the strength of

association between educational attainment Answer 1
rated as a continuous measure and degree of Choose...
compliance rated as a continuous measure.
A researcher wants to compare the proportion
of smoking psychiatrists and nonsmoking Answer 2
psychiatrist in a local hospital who has been Choose...
off sick at least one day in last 28 days.
The degree of agreement and reliability Answer 3
between two raters for nominal measures. Choose...

Check

Feedback

Your answer is incorrect.

Explanation: To measure the strength of association between two continuous
measures, correlation can be tested using Pearson's r. Chi-square test is useful
to compare proportions between two groups. Reliability of nominal measures can
be tested using kappa.
The correct answer is:
A researcher wants to assess the strength of association between educational
attainment rated as a continuous measure and degree of compliance rated as a
continuous measure.
– Pearson's test,
A researcher wants to compare the proportion of smoking psychiatrists and
nonsmoking psychiatrist in a local hospital who has been off sick at least one
day in last 28 days.
– Chi-square test,
The degree of agreement and reliability between two raters for nominal
measures.
– Kappa

Question 3
Not answered
Marked out of 5.00

Flag question

Question text

HiY AdvStats EMI003

Sampling methods I
Choose the most appropriate sampling technique used for each study described
below:

An addiction psychiatrist is keen to

investigate the effect of DVLA warning on
heroin users. He identifies four regular
Answer 1
patients in his methadone clinic as suitable
Choose...
for the study and distributes copies of
screening questionnaire through them to
other heroin users known to drive a car.
To study the amount of over-the-counter use
of analgesics in depressed patients, first 50 Answer 2
consecutive patients attending out patient Choose...
clinic are screened.
To study the amount of over-the-counter use
of analgesics in depressed patients, a
Answer 3
questionnaire is sent to every 10th patient in
Choose...
an out-patient register between two specified
periods of time.
In order to study the amount of over-the- Answer 4
counter use of analgesics in depressed Choose...
patients, patients in an outpatient register are
grouped by sex, age >65 or <65, presence or
absence of a history of analgesic prescription.
A random sample is picked from each group.
In order to study the amount of over-the-
counter use of analgesics in depressed
patients, patients in an outpatient register are Answer 5
grouped by sex, age >65 or <65, presence or Choose...
absence of a history of analgesic prescription.
First ten patients from each group are picked.
Check

Feedback

Your answer is incorrect.

Explanation:
Snowball sampling is used to reach normally difficult-to-engage sample frame.
Consecutive sampling is a form of convenience sampling.
Using every nth person, as if in roll calls, produces a systematic sample.
After stratifying a target population into different strata, random samples are
picked from each stratum in question 4 - this constitutes a stratified random
sampling.
In question 5, after the stratification, a proportion is selected - this is called
quota sampling.
The correct answer is:
An addiction psychiatrist is keen to investigate the effect of DVLA warning on
heroin users. He identifies four regular patients in his methadone clinic as
suitable for the study and distributes copies of screening questionnaire through
them to other heroin users known to drive a car.
– Snowball sampling,
To study the amount of over-the-counter use of analgesics in depressed patients,
first 50 consecutive patients attending out patient clinic are screened.
– Convenience sampling,
To study the amount of over-the-counter use of analgesics in depressed patients,
a questionnaire is sent to every 10th patient in an out-patient register between
two specified periods of time.
– Systematic sampling,
In order to study the amount of over-the-counter use of analgesics in depressed
patients, patients in an outpatient register are grouped by sex, age >65 or <65,
presence or absence of a history of analgesic prescription. A random sample is
picked from each group.
– Stratified random sampling,
In order to study the amount of over-the-counter use of analgesics in depressed
patients, patients in an outpatient register are grouped by sex, age >65 or <65,
presence or absence of a history of analgesic prescription. First ten patients
from each group are picked.
– Quota sampling

Question 4
Not answered
Marked out of 4.00

Flag question

Question text

HiY AdvStats EMI004

Statistical tests II
Hyperglycemia and metabolic syndrome are associated with the use of
antipsychotics in schizophrenia patients. A researcher wants to investigate this
phenomenon for the newly marketed antipsychotic paliperidone. He has various
options with respect to study designs. For each design below choose ONE
appropriate statistical test from the given list.

The mean increase in glucose levels after using

Answer 1
antipsychotic paliperidone for eight weeks in a
Choose...
sample of 50 inpatients.
The mean change in glucose levels after using
antipsychotics for eight weeks in a sample of
Answer 2
50 inpatients taking antipsychotic paliperidone
Choose...
compared to 33 patients attending day unit for
psychotherapy for the same period.
An increase in the prevalence of metabolic
syndrome in a sample of 50 inpatients after Answer 3
using antipsychotic paliperidone for eight Choose...
weeks.
Comparing the prevalence of metabolic Answer 4
syndrome in a sample of 50 inpatients after Choose...

using antipsychotic paliperidone for eight

weeks and a sample of 33-day unit patients
undertaking psychotherapy for eight weeks.
Check

Feedback

Your answer is incorrect.

Explanation:
For question 1, mean values before and after are compared in the same sample:
paired t-test can be used for this.
For question 2, differences in values of each group are first computed, then a
mean is obtained for each group which is then compared: this is effectively the
same as comparing two means which are obtained by other means. Hence, we
can use unpaired t test.
Question 3 refers to comparing proportions in a paired condition (before and
after) - McNemar's test is best used for this purpose. It is similar to Chi-square
test but can be used in non-parametric or paired situations.
Question 4 is comparing proportions between two independent groups - Chi-
square is the most suitable test.
The correct answer is:
The mean increase in glucose levels after using antipsychotic paliperidone for
eight weeks in a sample of 50 inpatients.
– Paired t-test,
The mean change in glucose levels after using antipsychotics for eight weeks in
a sample of 50 inpatients taking antipsychotic paliperidone compared to 33
patients attending day unit for psychotherapy for the same period.
– Students t-test,
An increase in the prevalence of metabolic syndrome in a sample of 50
inpatients after using antipsychotic paliperidone for eight weeks.
– McNemar's test,
Comparing the prevalence of metabolic syndrome in a sample of 50 inpatients
after using antipsychotic paliperidone for eight weeks and a sample of 33-day
unit patients undertaking psychotherapy for eight weeks.
– Chi-square test

Question 5
Not answered
Marked out of 4.00
 Flag question

Question text

HiY AdvStats EMI005

Statistical tests III
For each of the studies below, choose the most appropriate statistical test to
analyse the data.

Comparison of systolic blood pressures in

Answer 1
independent samples of schizophrenia patients
Choose...
taking clozapine with patients taking olanzapine.
Comparison of the prevalence of hepatitis C Answer 2
among IV heroin users and IV cocaine users. Choose...

Comparison of the level of platelet serotonin in

depressed adolescents and adults following Answer 3
administration of three different antidepressant Choose...
drugs.
Comparison of serum cholesterol before and
after attending a lifestyle modification clinic in a Answer 4
group of bipolar patients attending assertive Choose...
outreach service.
Check

Feedback

Your answer is incorrect.

Explanation:
Comparison of mean blood pressure between two groups can be achieved by
employing Student's unpaired t test. Prevalence is a proportion.
Comparison of proportions (categorical outcome) can be achieved using chi-
square test.
Comparing means of a continuous variable such as platelet serotonin level in
more than two groups can be done using ANOVA. Please note that while ANOVA
indicates whether or not differences exist among groups, it does not tell where
such differences are located.
Paired observations of mean values (serum cholesterol in the same subject
before and after an intervention) can be tested using paired t-test.
The correct answer is:
Comparison of systolic blood pressures in independent samples of schizophrenia
patients taking clozapine with patients taking olanzapine.
– Student t-test,
Comparison of the prevalence of hepatitis C among IV heroin users and IV
cocaine users.
– Chi-square test.,
Comparison of the level of platelet serotonin in depressed adolescents and
adults following administration of three different antidepressant drugs.
– Analysis of variance,
Comparison of serum cholesterol before and after attending a lifestyle
modification clinic in a group of bipolar patients attending assertive outreach
service.
– Paired t-test

Question 6
Not answered
Marked out of 4.00

Flag question

Question text

HiY AdvStats EMI006

Sampling methods II
For each of the techniques described below, choose the most appropriate
sampling technique from the given list.

Choosing volunteers from a trust induction

Answer 1
seminar to participate in a survey of the
Choose...
prescribing practice of psychiatrists.
Listing individuals in a locality by ethnic
Answer 2
group and choosing a proportion from each
Choose...
ethnic group at random.
Initially, numbering all the constituents of a
Answer 3
sampling frame and then using a random
Choose...
number table to pick cases.
Randomly selecting deaneries, and then
Answer 4
sampling every trainee attached to a
Choose...
deanery.
Check

Feedback

Your answer is incorrect.

Explanation:
A convenience sample is a sample where the patients are selected, in part or
whole, at the convenience of the researcher. The researcher makes no or limited
attempt to ensure representativeness of this sample to the source population. A
good example of a convenience sample is standing at a shopping mall and
selecting shoppers as they walk by to fill out a survey on shopping practices.
In simple random sampling each individual in the population has an equal chance
(probability) of being selected.
Quota sampling refers to stratification of source population into subgroups and
allocating a proportion to each group; then in a non-random manner, (e.g. first 10
in each group) the individuals are selected to be included in the study. Both
convenience sampling and quota sampling are non-random methods. One way of
obtaining a random sample is to give each individual a number and then to use a
computer-generated table of random numbers to decide which individuals should
be included.
Cluster sampling involves the use of groups as the sampling unit rather than
individuals (e.g. households, school classes or GP practices). The groups to be
studied should be randomly selected from all possible groups of the same type.
Most importantly, all members of the selected groups must be included in the
study to qualify as a cluster sample. If individuals are again randomly selected
within the random clusters, this will be a multi-stage sample (two random
sampling). In stratified sampling, a population is divided into subgroups based on
defined characteristics such as age, sex, or severity of illness, or any
combination of these; then random samples are selected from each subgroup. In
such sample, the probability of an individual being included varies according to a
predetermined characteristic (strata). This method ensures that small subgroups
that are of particular interest to the investigator are adequately represented.
The correct answer is:
Choosing volunteers from a trust induction seminar to participate in a survey of
the prescribing practice of psychiatrists.
– Convenience sampling,
Listing individuals in a locality by ethnic group and choosing a proportion from
each ethnic group at random.
– Stratified random sampling,
Initially, numbering all the constituents of a sampling frame and then using a
random number table to pick cases.
– Simple random sampling,
Randomly selecting deaneries, and then sampling every trainee attached to a
deanery.
– Cluster sampling

Question 7
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI007

Sampling methods III
For each of the techniques described below, choose the most appropriate
sampling technique from the given list.

Selecting all persons in a randomly selected Answer 1

GP practice. Choose...

Answer 2
Selecting every 4th outpatient in the register
Choose...

A sample is categorised according to

identified mediators and effect modifiers, and Answer 3
subjects are randomly picked within each Choose...
category to constitute study sample.
Check

Feedback

Your answer is incorrect.

Explanation:
The difference between cluster sampling and simple random sampling is in the
individual unit that is picked randomly. If large units such as GP practices,
postcode areas or classrooms are picked wherein all individual members of the
group are studied then it is called cluster sampling. In simple random samples,
the individual participants are directly picked in a random fashion. In stratified
random samples, larger subgroups are identified according to known mediators,
and in each group, members are picked randomly. Hence the groups are not
randomly chosen, unlike cluster sampling and within each group not all members
are included, again in contrast from cluster sampling.
Systematic sampling appears organised and 'systematic' where every nth
member is picked for a study. But the probability of being chosen is not random
and is influenced by unknown factors. Hence, it is a non-random method of
sampling.
The correct answer is:
Selecting all persons in a randomly selected GP practice.
– Cluster sampling,
Selecting every 4th outpatient in the register
– Systematic sampling,
A sample is categorised according to identified mediators and effect modifiers,
and subjects are randomly picked within each category to constitute study
sample.
– Stratified sampling

Question 8
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI008

Errors in research
For each of the errors described below, choose the most appropriate term from
the given list.

A religious leader quotes epidemiological data that

countries with the highest use of condoms and
other artificial methods of contraceptive practices Answer 1
have a higher risk of AIDS. He asserts that Choose...
condoms are ineffective in reducing HIV
transmission.
A researcher fails to show the effectiveness of Answer 2
clozapine in treatment-resistant schizophrenia Choose...

despite conducting a rigorous randomised

controlled trial for 18 months. This null-result is in
contrast to popular belief and prior research
evidence.
In a case-control study of autism and parental
bonding, gender matched siblings of same
households are chosen as controls. Despite having Answer 3
a high prevalence of poor bonding in autistic Choose...
children, the differences do not reach statistical
significance.
Check

Feedback

Your answer is incorrect.

Explanations:
In ecological studies risk factors are not examined in each individual; instead
prevalence of risk factors in a large (often naturally occurring) group such as
councils, counties, districts, countries, etc are associated with prevalence of an
outcome in the same unit. This may produce erroneous results as the individual
with a risk factor may not be the same individual who suffers the outcome. This
is called ecological fallacy. This question is a classical example of
the ecological fallacy.
A common reason for non-replication of widely established findings is type 2
error - this may be due to lack of sufficient sample size for the experiment.
Selection bias occurs when the subjects selected for the study are not
representative of the population they come from. Selecting controls in analytical
studies must be carefully done to avoid this bias. In this study the trouble with
selecting siblings for controls is that sibling of an autistic child is much more
likely to have poor bonding either due to genetic influences or parenting styles
than a child from the general population. Thus poor bonding will be
overrepresented among the controls too, and results might make the association
appear weaker than the reality.
The correct answer is:
A religious leader quotes epidemiological data that countries with the highest
use of condoms and other artificial methods of contraceptive practices have a
higher risk of AIDS. He asserts that condoms are ineffective in reducing HIV
transmission.
– Ecological fallacy,
A researcher fails to show the effectiveness of clozapine in treatment-resistant
schizophrenia despite conducting a rigorous randomised controlled trial for 18
months. This null-result is in contrast to popular belief and prior research
evidence.
– Type 2 error,
In a case-control study of autism and parental bonding, gender matched siblings
of same households are chosen as controls. Despite having a high prevalence of
poor bonding in autistic children, the differences do not reach statistical
significance.
– Selection bias

Question 9
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI009

Data Types
For each of the following descriptions choose an appropriate response from
given list:

A prospective study investigates relapse rates of Answer 1

depression in patients who have maintenance ECT. Choose...
A neuroimaging study measures insight using a
Answer 2
single item as good, moderate, mildly abnormal,
Choose...
poor or absent to correlate with parietal volume.
Lean body mass on admission at an eating disorder Answer 3
unit. Choose...

Check

Feedback

Your answer is incorrect.

Explanation:
Having a relapse or not is a categorical outcome (i.e. 'yes' or 'no' type). Hence,
the outcome is nominal. Likert scales, as described here in the insight
measurement, produce ordinal data. Lean body mass is a continuous variable
with a meaningful zero. So one can say that subject A is two times heavier than
subject B. As such ratios are possible, lean body mass is a ratio data.
Consider interval scales such as dates of admission to a psychiatric unit. These
are ordered and have real numerical units and can be subtracted from each
other, but there is no intrinsic zero to the scale so that dividing them does not
make sense. For example, it would be absurd to say one date of admission is
three times 'bigger' than another date.
The correct answer is:
A prospective study investigates relapse rates of depression in patients who
have maintenance ECT.
– Binary Nominal,
A neuroimaging study measures insight using a single item as good, moderate,
mildly abnormal, poor or absent to correlate with parietal volume.
– Ordinal,
Lean body mass on admission at an eating disorder unit.
– Ratio

Question 10
Not answered
Marked out of 5.00

Flag question

Question text

HiY AdvStats EMI010

Graphs describing data
Choose one appropriate option from the given list for each of the following
descriptions:

A graph where both vertical and horizontal

columns are made of numbers where all individual Answer 1
data are represented instead of displaying bars of Choose...
various colours to represent the magnitude.
A graph where a vertical line extends on either
Answer 2
side of a vertical bar displaying interquartile
Choose...
range.
A graph where two continuous variables are
Answer 3
plotted after paired observations resulting in
Choose...
numerous dots.
A graph used in reporting meta-analyses where in
Answer 4
the effect size is represented by a diamond-
Choose...
shaped lozenge.
A graph that could be used to identify potential Answer 5
Choose...
heterogeneity in which effect size is plotted
Choose...
against a measure of precision.
Check

Feedback

Your answer is incorrect.

Explanations: The answers are self-explanatory. L'Abbe plot is also used in a
meta-analysis. The proportion of subjects improved with placebo is plotted on
one x-axis; proportion improved on active medication is plotted on the y axis.
The individual trials are often indicated by differently sized symbols.
[Link]
The correct answer is:
A graph where both vertical and horizontal columns are made of numbers where
all individual data are represented instead of displaying bars of various colours
to represent the magnitude.
– Stem and leaf plot,
A graph where a vertical line extends on either side of a vertical bar displaying
interquartile range.
– Box-Whisker plot,
A graph where two continuous variables are plotted after paired observations
resulting in numerous dots.
– Scattergram,
A graph used in reporting meta-analyses where in the effect size is represented
by a diamond-shaped lozenge.
– Forest plot,
A graph that could be used to identify potential heterogeneity in which effect
size is plotted against a measure of precision.
– Galbraith plot

Question 11
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI011

Measures of Agreement
 Nurse
Psychiatrist Borderline + Borderline - Total
Borderline + 40 20 60
Borderline - 10 30 40
Total 50 50 100

Using the given table answer the following questions on measures of agreement

The overall agreement between the two clinicians Answer 1

out of the total Choose...

Answer 2
The overall agreement by chance alone
Choose...

Answer 3
The value of kappa
Choose...

Check

Feedback

Your answer is incorrect.

Explanation:
Step 1:
Number of occasions of agreement = 40 (agreed for presence of diagnosis)+30
(agreed for absence of diagnosis) = 70 out of 100 individual events = 70% . This
is the overall percent agreement between the two clinicians out of the total.
Step 2:
To determine the effect of chance, assume that the psychiatrist has no clinical
skills and, as a result, has a tendency to depend on his stars all the time. If this
is the case, he must be diagnosing 60 out of any 100 as borderline and 40 out of
any 100 as not having borderline irrespective of the population (from the above
data). If he evaluates the 50 patients diagnosed as borderline by the nurse, he
must call 30 of them to have the diagnosis; using the same principles he must
call 20 out of 50 patients rejected by the nurse as non-borderline. So if he
depends on chance purely, he must agree with the nurse on 30+20 = 50 out of
100 times. The overall percent agreement by chance alone must be 50%.
Step 3:
Kappa expresses the extent to which the observed agreement exceeds that
which would be expected by chance alone (numerator) relative to the most that
the observers could hope to improve their agreement (i.e., 100% - agreement
expected by chance alone) [denominator].
= (70-50) / (100-50) = 20/50 = 0.4
The correct answer is:
The overall agreement between the two clinicians out of the total
– 0.7,
The overall agreement by chance alone
– 0.5,
The value of kappa
– 0.4

Question 12
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI012

Numbers in critical appraisal I
For each of the questions below, select the one most appropriate number from
the list given.

The point at which the confidence interval for a Answer 1

difference between means becomes non-significant. Choose...

The likelihood ratio for a negative test when Answer 2

sensitivity is 60%, and specificity is 80%. Choose...

The correlation coefficient where there is an absolute Answer 3

inverse association between two variables. Choose...

Check

Feedback

Your answer is incorrect.

Explanation:
Q1. When the confidence interval range includes the value 0 for mean
differences, the results become insignificant. The same applies to a range of
ratios when they include the value 1.
Q2. The likelihood ratio of negative test = [(1-sensitivity)/specificity] =
[(1-60%)/80%] = [(1-0.6)/0.8] = 0.4/0.8 = 1/2 = 0.5
Q3. A correlation coefficient of -1 means there is an absolute inverse association
between the two variables
The correct answer is:
The point at which the confidence interval for a difference between means
becomes non-significant.
– 0,
The likelihood ratio for a negative test when sensitivity is 60%, and specificity is
80%.
– 0.5,
The correlation coefficient where there is an absolute inverse association
between two variables.
– -1

Question 13
Not answered
Marked out of 4.00

Flag question

Question text

HiY AdvStats EMI013

Sampling methods IV
Choose one option for each question below.

A researcher is interested in measuring

dimensional factors of PTSD in illegal Answer 1
immigrants. He could not find a reliable Choose...
sampling frame for recruitment.
A researcher is interested in studying the
lifestyle factors influencing metabolic
syndrome in patients who receive
antipsychotic prescriptions. He classifies the Answer 2
eligible sample into inpatients, outpatients Choose...
and CMHT clients before recruiting for his
study, to balance for the effect of one's
environment on metabolic syndrome.
A research team is interested in studying the Answer 3
prevalence of depression across all primary Choose...

care attendees in The UK. But due to the

scope of the task, they decide to pick
randomly representative primary care
authorities and survey all patients in the
selected regions.
A research nurse wants to study the
experience of some mentally ill parents with
Answer 4
respect to the demands of their teenage
Choose...
children. She designs a qualitative study
method to carry out this work.
Check

Feedback

Your answer is incorrect.

Explanation:
Question 1: To locate difficult to recruit individuals such as in the question, one
can use snowballing technique; here one participant identifies another
participant, who recruits one more and so on.
In question 2, as the existing sample is classified into various groups (strata) to
enable equitable distribution of a possible mediating factor, this is stratified
random sampling.
Instead of carrying out a whole country census, the researchers have randomly
picked GP clusters and included everyone in the clusters in question 3. Instead,
if they randomly picked individuals after randomly picking clusters, this would
have been multistage sampling.
Purposive samples are suitable for qualitative study in which required outliers
could be conveniently picked, as in question 4.
The correct answer is:
A researcher is interested in measuring dimensional factors of PTSD in illegal
immigrants. He could not find a reliable sampling frame for recruitment.
– Snowball sampling,
A researcher is interested in studying the lifestyle factors influencing metabolic
syndrome in patients who receive antipsychotic prescriptions. He classifies the
eligible sample into inpatients, outpatients and CMHT clients before recruiting
for his study, to balance for the effect of one's environment on metabolic
syndrome.
– Stratified random sampling,
A research team is interested in studying the prevalence of depression across all
primary care attendees in The UK. But due to the scope of the task, they decide
to pick randomly representative primary care authorities and survey all patients
in the selected regions.
– Cluster sampling,
A research nurse wants to study the experience of some mentally ill parents
with respect to the demands of their teenage children. She designs a qualitative
study method to carry out this work.
– Convenience sampling

Question 14
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI014

Glossary of statistical terms 1
For each description below choose the most appropriate answer from the given
list.

The ability to detect the true difference, if it Answer 1

exists, using the given sample. Choose...

Detecting a difference between studied groups Answer 2

when such difference does not exist. Choose...

The probability that a range of specified values Answer 3

contains the true value of the variable. Choose...

Check

Feedback

Your answer is incorrect.

Explanations:
Q1. The power of a study is its ability to detect a true difference if it exists.
Question 2 describes a type I error. This may be due to the high threshold for p-
value (greater than usual 0.05) or multiple testing.
Question 3 refers to confidence intervals. The confidence level is usually set at
95%.
The correct answer is:
The ability to detect the true difference, if it exists, using the given sample.
– Power,
Detecting a difference between studied groups when such difference does not
exist.
– Type 1 error,
The probability that a range of specified values contains the true value of the
variable.
– Confidence interval

Question 15
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI015

Statistical Tests 4
For each of the following vignettes choose the most appropriate statistical
test/s.

Comparison of final outcome measures between

two groups (placebo and antipsychotic) in a
Answer 1
psychosis treatment study, taking into account
Choose...
significant baseline difference between groups.
Results are normally distributed.
Comparison of final outcome measures between
the three groups (CBT, SSRI, placebo) in a Answer 2
depression treatment study. Results are normally Choose...
distributed.
Comparison of a measure of coping styles in a
Answer 3
depressed and a control group. Results are
Choose...
continuous outcomes with skewed distributions.
Check

Feedback

Your answer is incorrect.

Explanation:
Q1: In question 1, the baseline differences can be considered as covariates while
mean differences are sought between the two groups. ANCOVA will be
appropriate for this purpose.
Q2: For mean comparisons across more than two groups, ANOVA is a handy
method.
Q3: As the distribution is skewed, a non-parametric approach for comparing
unpaired observations (i.e. Mann-Whitney U Test) can be used.
The correct answer is:
Comparison of final outcome measures between two groups (placebo and
antipsychotic) in a psychosis treatment study, taking into account significant
baseline difference between groups. Results are normally distributed.
– ANCOVA.,
Comparison of final outcome measures between the three groups (CBT, SSRI,
placebo) in a depression treatment study. Results are normally distributed.
– ANOVA.,
Comparison of a measure of coping styles in a depressed and a control group.
Results are continuous outcomes with skewed distributions.
– Mann-Whitney test.

Question 16
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI016

Statistical Concepts 2
Find the appropriate measure from the list provided that corresponds to the
description below.

The 'uncertainty' surrounding a mean or Answer 1

other summary estimate. Choose...

A measure reflecting the probability of

Answer 2
making a false positive conclusion from the
Choose...
results.
The probability that your experiment will find
Answer 3
a pre-specified difference when one truly
Choose...
exists.
Check

Feedback

Your answer is incorrect.

Explanation:
Q1: Standard error reflects how much sampling fluctuation a measurement can
have by providing the uncertainty around the estimated value.
Q2: P value (generally 5%) is synonymous with type 1 error rate. It reflects the
probability of making a false positive conclusion from the results.
Q3: In simplistic terms, statistical power is the odds that you will observe a
treatment effect when it occurs (question 3).
The correct answer is:
The 'uncertainty' surrounding a mean or other summary estimate.
– Standard Error,
A measure reflecting the probability of making a false positive conclusion from
the results.
– P-value,
The probability that your experiment will find a pre-specified difference when
one truly exists.
– Power

Question 17
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI017

Numbers in critical appraisal II
For each of the questions below, select the one most appropriate number from
the list provided.

The point at which the confidence interval for an odds Answer 1

ratio becomes non-significant. Choose...
 Answer 2
Odds for a 90% event rate.
Choose...

Generally accepted usual risk, expressed as a

Answer 3
percentage, of type I error in a randomised controlled
Choose...
trial.
Check

Feedback

Your answer is incorrect.

Explanation:
Q1. When the confidence interval range includes the value 0 for mean
differences, the results become insignificant. The same applies to a range of
ratios when they include the value 1.
Q2. The event rate is a probability. Odds = [probability/(1-probability)] = [90%/(1-
90%)] = [0.9/(1-0.9)] = 0.9/0.1 = 9
Q3. 5% error is the normally accepted type 1 error rate. This converts to a p-
value of 0.05.
The correct answer is:
The point at which the confidence interval for an odds ratio becomes non-
significant.
– 1,
Odds for a 90% event rate.
– 9,
Generally accepted usual risk, expressed as a percentage, of type I error in a
randomised controlled trial.
–5

Question 18
Not answered
Marked out of 5.00

Flag question

Question text

HiY AdvStats EMI018

Glossary of Statistical Terms 2
Identify the correct terminology used in statistics for each of the following
definitions given below;
The middle value in a distribution with Answer 1
equal number of values on either side Choose...

The value that occurs most frequently in a Answer 2

distribution Choose...

The average of all the different values in a Answer 3

given set Choose...

A term that refers to 25% of the data in a Answer 4

distribution. Choose...

The difference between the 75th Answer 5

percentile and the 25th percentile Choose...

Check

Feedback

Your answer is incorrect.

Explanation: The median is the middle value with an equal number of values on
either side. The median is the 50th percentile. The mode is the value that
occurs most frequently. If there are 2 or more values that occur with equal
frequency, the distribution is described as multi-modal. The mean is the
average of all the different values. Quartile refers to 25% of the data. So, the
first quartile includes values up to the 25th percentile. The inter-quartile
range is the difference between the 75th percentile and the 25th percentile.
The correct answer is:
The middle value in a distribution with equal number of values on either side
– Median,
The value that occurs most frequently in a distribution
– Mode,
The average of all the different values in a given set
– Mean,
A term that refers to 25% of the data in a distribution.
– Quartile,
The difference between the 75th percentile and the 25th percentile
– Interquartile range

Question 19
Not answered
Marked out of 4.00
 Flag question

Question text

HiY AdvStats EMI019

Glossary of Statistical Terms 3
Identify the correct terminology used in statistics for each of the following
definitions given below:

The term used to describe a distribution

Answer 1
whose values shows either a negative or a
Choose...
positive skew.
Answer 2
It is the square root of the variance
Choose...

It is a measure of how close the sample Answer 3

mean is likely to be to the population mean. Choose...
The term used to describe a distribution in
Answer 4
which the values are dispersed equally on
Choose...
either side of the mean.
Check

Feedback

Your answer is incorrect.

Explanation:
Parametric distribution is a distribution in which the values are distributed
equally on either side of the mean. In a parametric distribution the mean =
median = mode. Non-parametric distribution refers to any distribution other
than a normal distribution. The distribution of values shows either a negative
skew or a positive skew.
Standard deviation and variance are measured of the spread of values around
the mean. The standard deviation (S.D.) is the square root of the variance. In a
parametric / normal distribution, 68% of values lie within 1 S.D. on either side of
the mean. About 95% of values lie within 2. S.D. (1.96 S.D., to be specific) on
either side of the mean.
Standard error of the mean refers to the S.D. of a distribution of sample means.
It is a measure of how close the sample mean is likely to be to the population
mean.
The correct answer is:
The term used to describe a distribution whose values shows either a negative
or a positive skew.
– Non-parametric,
It is the square root of the variance
– Standard deviation,
It is a measure of how close the sample mean is likely to be to the population
mean.
– Standard error of the mean,
The term used to describe a distribution in which the values are dispersed
equally on either side of the mean.
– Parametric

Question 20
Not answered
Marked out of 4.00

Flag question

Question text

HiY AdvStats EMI020

Plots & Graphs (1)
Using the definitions given below, identify the type of plots/graphs used in
statistical illustrations:

A survival curve in which the survival probability Answer 1

is plotted against the time from baseline. Choose...

A graph showing the median, range and Answer 2

interquartile range of a set of values Choose...

A diagram showing the proportional distribution Answer 3

of a categorical variable using a divided circle. Choose...

A diagram that illustrates the distribution of a

categorical variable by showing a separate
Answer 4
horizontal or vertical bar for each 'category', its
Choose...
length being proportional to the number of
observations in that 'category'
Check
Feedback

Your answer is incorrect.

Explanation:
1) Kaplan-Meier plot is a survival curve in which the survival probability is
plotted against the time from baseline. It is used when exact times to reach the
endpoint.
2) A box-and-whisker plot is a diagram illustrating the distribution of a variable; it
indicates the median, upper and lower quartiles, and, often, the maximum and
minimum values.
3) A pie chart is a diagram showing the frequency/proportional distribution of a
categorical or discrete variable. A circular 'pie' is split into sections, one for
each 'category'; the area of each section is proportional to the frequency in that
category.
4) A bar or column chart is a diagram that illustrates the distribution of a
categorical or discrete variable by showing a separate horizontal or vertical bar
for each 'category', its length being proportional to the (relative) frequency in
that 'category'.
(Ref: Medical statistics at a glance-Petrie & Sabin)
The correct answer is:
A survival curve in which the survival probability is plotted against the time from
baseline.
– Kaplan-Meier plot,
A graph showing the median, range and interquartile range of a set of values
– Box and whisker plot,
A diagram showing the proportional distribution of a categorical variable using a
divided circle.
– Pie Chart,
A diagram that illustrates the distribution of a categorical variable by showing a
separate horizontal or vertical bar for each 'category', its length being
proportional to the number of observations in that 'category'
– Bar chart

Question 21
Not answered
Marked out of 5.00

Flag question
Question text

HiY AdvStats EMI021

Plots & Graphs (2)
Using the definitions given below, identify the type of plots/graphs used in
statistical

A diagram relating the pre-test probability of a

Answer 1
diagnostic test result to the likelihood and the
Choose...
post-test probability.
A diagram used in a meta-analysis showing the
Answer 2
estimated effect of each trial and the pooled
Choose...
average with confidence intervals
A diagram that illustrates the relative frequency
Answer 3
distribution of a continuous variable by using
Choose...
connected bars.
The two-dimensional plot of one variable
Answer 4
against another, with each pair of observations
Choose...
marked by a point
A mixture of a diagram and a table used to Answer 5
illustrate the distribution of data. Choose...

Check

Feedback

Your answer is incorrect.

Explanation:
1) Fagan's Nomogram (also called Bayesian Nomogram) is a diagram relating the
pre-test probability of a diagnostic test result to the likelihood and the post-test
probability. It is usually used to convert the former into the latter.
2). A forest plot is a diagram used in a meta-analysis showing the estimated
effect in each trial and their average (with confidence intervals)
3). A Histogram is a graph of continuous data with the data categorized into
many classes. It illustrates the (relative) frequency distribution of a continuous
variable by using connected bars. The bar's area is proportional to the (relative)
frequency in the range specified by the boundaries of the bar.
4) Scatter diagram is a two-dimensional plot of one variable against another,
with each pair of observations marked by a point
5). A steam and leaf plot is a mixture of a diagram and a table used to illustrate
the distribution of data. It is similar to a histogram and is effectively the data
values displayed in increasing order of size.
(Ref: Petrie & Sabin, Medical statistics at a glance 1st edn.)
The correct answer is:
A diagram relating the pre-test probability of a diagnostic test result to the
likelihood and the post-test probability.
– Fagan's Nomogram,
A diagram used in a meta-analysis showing the estimated effect of each trial and
the pooled average with confidence intervals
– Forest plot,
A diagram that illustrates the relative frequency distribution of a continuous
variable by using connected bars.
– Histogram,
The two-dimensional plot of one variable against another, with each pair of
observations marked by a point
– Scatter plot,
A mixture of a diagram and a table used to illustrate the distribution of data.
– Steam and leaf plot

Question 22
Not answered
Marked out of 5.00

Flag question

Question text

HiY AdvStats EMI022

Sampling methods 5
Identify the type of sampling for each of the following descriptions;

Every individual in a population has an equal Answer 1

chance of being chosen. Choose...

Dividing the population into clusters, typically on

Answer 2
the basis of geography, and taking a random
Choose...
sample of the clusters
Dividing the population into groups on the basis Answer 3
of some suspected confounding characteristic Choose...
and then sampling each group to ensure fair
distribution of these variables.
Selecting the population on the basis of whoever Answer 4
is accessible Choose...

Choosing every nth item from a list, beginning at Answer 5

a random point. Choose...

Check

Feedback

Your answer is incorrect.

Explanation:
1) In random sampling every individual in a population has an equal chance of
being chosen.
2) Cluster sampling involves dividing the population into naturally occurring
clusters, typically on the basis of geography, and taking a random sample of the
clusters (and thus including all of the subjects that reside within a cluster)
3) Stratified sampling involves dividing the population into groups on the basis of
some suspected confounding characteristic and then sampling each group to
ensure fair distribution of these variables.
4) Convenience sampling is selecting on the basis of convenience - not using a
sampling frame but including whoever is easily accessible.
5) Systematic sampling involves choosing every nth item from a list, sometimes
beginning at a random point.
The correct answer is:
Every individual in a population has an equal chance of being chosen.
– Random sampling,
Dividing the population into clusters, typically on the basis of geography, and
taking a random sample of the clusters
– Cluster sampling,
Dividing the population into groups on the basis of some suspected confounding
characteristic and then sampling each group to ensure fair distribution of these
variables.
– Stratified sampling,
Selecting the population on the basis of whoever is accessible
– Convenience sampling,
Choosing every nth item from a list, beginning at a random point.
– Systematic sampling
Question 23
Not answered
Marked out of 4.00

Flag question

Question text

HiY AdvStats EMI023

Reliability
Identify the type of reliability for each of the following descriptions:

The level of agreement between two ratings

Answer 1
on the same material by same rater but at
Choose...
different times
The level of agreement between the ratings Answer 2
of the same material by two raters Choose...

The level of agreement between one-half of a Answer 3

test with the other. Choose...

The level of agreement between the Chinese

version and the English version of a Answer 4
depression scale in the same group of Choose...
bilingual subjects fluent in both languages.
Check

Feedback

Your answer is incorrect.

Explanation:
1) Test-retest reliability describes the level of agreement between two ratings on
the same material but at different times
2) Inter-rater reliability describes the level of agreement between the ratings on
the same material by two raters. (2 or more assessors make simultaneous
measurements)
3) Split half reliability describes the extent of agreement between one-half of a
test with the other (e.g. a 20 item scale is divided into two halves of 10 items
each)
4) Alternate form reliability describes the level of agreement between two
different versions of the same test. The alternate-form reliability can also be
used to devise modified versions of a test that helps to overcome the "practice
effect" by changing the wording of the test questions in a functionally equivalent
form.
The correct answer is:
The level of agreement between two ratings on the same material by same rater
but at different times
– Test-retest reliability,
The level of agreement between the ratings of the same material by two raters
– Inter-rater reliability,
The level of agreement between one-half of a test with the other.
– Split-half reliability,
The level of agreement between the Chinese version and the English version of a
depression scale in the same group of bilingual subjects fluent in both
languages.
– Alternate form reliability

Question 24
Not answered
Marked out of 5.00

Flag question

Question text

HiY AdvStats EMI024

Statistical tests 6
Choose the most appropriate test for each of the following situations;

In a depression treatment study, the comparison

of final outcome measures between the three Answer 1
groups (CBT, interpersonal therapy and placebo). Choose...
The results are normally distributed
Investigation of the association between several Answer 2
risk factors and the risk of developing depression Choose...

The comparison of the risk of developing Answer 3

depression in 2 groups of 500 people. One group is Choose...
comprised of people who regularly use alcohol
and the other of people who do not regularly
consume alcohol
Comparison of a measure of coping styles
measured using a rating scale in a psychotic and Answer 4
a control group. The results are continuous Choose...
outcomes with skewed distribution
Comparison of final outcome measure between 2
groups (placebo and antidepressants) in a
Answer 5
depression treatment study taking into account
Choose...
significant baseline difference between groups.
The results are normally distributed.
Check

Feedback

Your answer is incorrect.

Explanation:
1) One-way ANOVA (analysis of variance) is a parametric test used to compare 3
or more groups. The first analysis is to find out there is any difference across the
groups.
2) Logistic regression is used in data sets where the dependent variable Y is
binary in nature (e.g. diagnosis present/absent, dead/alive, etc.). The logistic
regression model can be described as that which examines the relationship
between one or more independent variables X1, X2, X3....Xk and the risk of
developing an event Y.
3) A chi-square test is used in frequency data and is a test of association
between 2 categorical variables. The categorical data from different samples are
compared regarding any real differences. It is used for compare proportions
between 2 or more groups
4) Mann-Whitney U test is the non-parametric equivalent of the 2 sample t test,
comparing the distributions of two independent groups of observation. Mann
Whitney U test is used to see whether there is a significant difference between 2
sets of data obtained from 2 different sets of subjects.
5) ANCOVA (analysis of covariance) is used to compare mean values of 2 groups
that differ at baseline. It is valid only if data is normally distributed and if
baseline and final measures are expected to be correlated.
The correct answer is:
In a depression treatment study, the comparison of final outcome measures
between the three groups (CBT, interpersonal therapy and placebo). The results
are normally distributed
– ANOVA,
Investigation of the association between several risk factors and the risk of
developing depression
– Logistic regression,
The comparison of the risk of developing depression in 2 groups of 500 people.
One group is comprised of people who regularly use alcohol and the other of
people who do not regularly consume alcohol
– Chi-squared test,
Comparison of a measure of coping styles measured using a rating scale in a
psychotic and a control group. The results are continuous outcomes with
skewed distribution
– Mann -Whitney test,
Comparison of final outcome measure between 2 groups (placebo and
antidepressants) in a depression treatment study taking into account significant
baseline difference between groups. The results are normally distributed.
– ANCOVA

Question 25
Not answered
Marked out of 3.00

Flag question

Question text

HiY AdvStats EMI025

Multivariate techniques
Identify one test for each of the following descriptions

It reduces a large
number of variables
that are highly inter-
Answer 1
correlated into a
Choose...
smaller group of
representative
variables
It predicts the value
of a single dependent
Answer 2
variable based on the
Choose...
values of several
independent variables
It predicts the value
of a dichotomous
Answer 3
variable based on the
Choose...
values of several
independent variables
Check

Feedback

Your answer is incorrect.

Explanation:
1) Factor analysis is essentially a method of data reduction - here groups of
variables that are highly inter-correlated in a dataset are identified. Then simple
correlation analysis of individual factors is carried out. A good example is factor
analysis of positive symptoms by Liddle to describe three groups of variables -
reality distortion, psychomotor poverty and disorganization syndromes.
2) Multiple linear regression is a linear regression model in which there is a
single dependent variable and 2 or more independent variables. In multiple linear
regression, several independent variables together predict a single dependent
variable. Consider y = a+b1x1 + b2 x2+ b3x3+ b4x4......................+ bnxn. Multiple
regression is a multivariate technique.
3) Logistic regression is used to predict the relationship between a binary
dependent variable and several independent variables. For linear regression, the
dependent variable must be continuous. If the dependent variable is binary, then
the logistic regression is used. It is commonly used to get OR, RR or hazard ratio
for independent variables that affect the dependent variable.
The correct answer is:
It reduces a large number of variables that are highly inter-correlated into a
smaller group of representative variables
– Factor analysis,
It predicts the value of a single dependent variable based on the values of
several independent variables
– Multiple regression,
It predicts the value of a dichotomous variable based on the values of several
independent variables
– Logistic regression

Finish review
Skip Quiz navigation

QUIZ NAVIGATION
 Question1This page Question2This page Question3This page Question4This
page Question5This page Question6This page Question7This
page Question8This page Question9This page Question10This
page Question11This page Question12This page Question13This
page Question14This page Question15This page Question16This
page Question17This page Question18This page Question19This
page Question20This page Question21This page Question22This
page Question23This page Question24This page Question25This page
Show one page at a timeFinish review

(c) 2015 SPMM Course Limited, UK

Terms & Conditions