Biostatistics Practice Questions

Q: In statistics, centiles and quartiles are considered as:

Measures of location/position. ***Measures of location/position***- Centiles (or **percentiles**) and **quartiles** are statistics that divide the data distribution into equal parts, indicating where a particular value stands relative to the rest of the data.- They are also known as **quantiles**, used to describe the location of specific data points within the distribution rather than summarizing the center or spread.*Measures of central tendency*- These statistics aim to describe the typical or **central value** of a dataset (e.g., **Mean**, **Median**, **Mode**).- While the median is technically the second quartile (**Q2**) and the 50th centile, the classifications of centiles and quartiles collectively are broader—measures of position.*Measures of dispersion*- These measures quantify the **spread** or **variability** of the data around the central value (e.g., **Standard Deviation**, **Variance**, Range).- Although quartiles are essential for calculating the **Interquartile Range (IQR)**, which is a measure of dispersion, the quartiles themselves define points of position.*Measures of correlation*- Correlation measures describe the **linear relationship** or association between **two variables** (e.g., Correlation Coefficient, R-value).- They are used in bivariate analysis and have no role in describing the position or central value of a single dataset.

Q: In a study done in a hospital, patients were categorized into three groups based on disease prevalence (Low, Medium, High), and individuals were then randomly selected from each group. What type of sampling is this?

Stratified random sampling. ***Correct: Stratified random sampling*** - This method involves dividing the population into non-overlapping subgroups (**strata**) based on a characteristic (here, disease prevalence: Low, Medium, High). - Subsequently, a **simple random sample** is drawn from *each* stratum independently to ensure representation from all groups. - This ensures that each subgroup is adequately represented in the final sample, making it ideal when the population has distinct subgroups. *Incorrect: Simple random sampling* - Every individual in the entire population has an equal and independent chance of being selected. - It does not involve dividing the population into specific subgroups or categories before selection. - This method may underrepresent or overrepresent certain subgroups by chance. *Incorrect: Systematic random sampling* - This involves selecting every *k*th element after a random start point, where *k* is the sampling interval (Population Size/Sample Size). - Like simple random sampling, it does not involve creating predefined strata based on characteristics like disease prevalence. - It's a simpler alternative to simple random sampling but doesn't ensure representation of specific subgroups. *Incorrect: Cluster random sampling* - The population is divided into natural groupings (**clusters**), such as geographical areas or schools. - Unlike stratification, entire clusters are randomly selected, and *all* individuals within the selected clusters (or a random sample thereof) are included in the study. - This differs from stratified sampling where we sample from ALL strata; in cluster sampling, we sample only SOME clusters.

Q: In a given population: - less than 15 years: 6000 - 15 to 64 years: 12000 - 65 years and above: 2000 Calculate the dependency ratio.

66 %. ***66 %*** - The **Dependency Ratio** measures the ratio of the economically dependent population to the economically productive population, usually expressed as a percentage. - **Dependent Population** (aged < 15 and ≥ 65): 6000 + 2000 = **8000** - **Productive Population** (aged 15-64): **12000** - **Dependency Ratio**: (8000 / 12000) × 100 = **66.67%** (rounded to 66%) *Incorrect: 33 %* - This figure is significantly lower than the true ratio and results from calculating the ratio of the dependent population to the **total population**, which is not the standard definition of the Dependency Ratio. - 33.3% represents the dependent population (8000) as a proportion of the total population (20000), not the dependency burden on the working population. *Incorrect: 50 %* - This result is obtained if only the young dependent group is used in the numerator (Young Dependency Ratio = 6000 / 12000 = **50%**). - However, the total Dependency Ratio must account for **both young and old dependents** to accurately reflect the economic burden. *Incorrect: 75 %* - Obtaining 75% would require the dependent population to be 9000 (i.e., 9000 / 12000), which is higher than the actual 8000 dependents. - This option represents an overestimation of the dependent burden on the working population.

Q: A new community intervention is initiated to reduce perinatal sepsis. Researchers allocate 20 Primary Health Centres (PHCs) to receive standard care and 20 PHCs to receive a community-based intervention. What type of study design is this?

Cluster-randomized control trial. ***Cluster-randomized control trial*** - A **Cluster-randomized control trial (C-RCT)** involves randomizing groups of individuals (clusters), such as entire PHCs or communities, rather than individual participants, to intervention or control arms. - Since whole **groups (20 PHCs)** are allocated to standard care or the intervention, it fits the definition of a C-RCT, which is often essential when an intervention cannot be delivered individually due to its nature (e.g., community-based programs). *Quasi-experimental study* - This design involves an intervention but lacks true **randomization** of participants or clusters, or it lacks a control group. - While both include control and intervention groups, the described scenario implies randomization (allocation), which makes a C-RCT a more specific and accurate fit than a general quasi-experimental design. *Cross-sectional study* - This design measures exposure and outcome simultaneously at a single point in time to determine **prevalence**. - It does not involve tracking groups over time or applying an **intervention**, which are key components of the described scenario. *Case-control study* - This is an observational study where individuals with a disease (cases) are compared to individuals without the disease (controls) to determine prior **exposure**. - This design is retrospective and does not involve the prospective application of a new **community intervention** as described in the question.

Q: Which of the following is the correct sequence of steps in a Randomized Controlled Trial (RCT)?

Randomisation → Manipulation → Follow-up → Assessment. **Randomisation → Manipulation → Follow-up → Assessment** is the correct sequence for conducting a Randomized Controlled Trial (RCT). - **Randomisation** is the essential first step that ensures baseline comparability between intervention and control groups, eliminating selection bias - **Manipulation** (intervention/treatment application) follows randomization - **Follow-up** involves monitoring participants over the study period - **Assessment** (outcome measurement) is performed at the end to evaluate the intervention's effect *Follow-up → Manipulation → Assessment → Randomisation* - Incorrect because randomization must occur before any intervention is applied - Starting with follow-up contradicts the fundamental RCT design *Manipulation → Assessment → Follow-up → Randomisation* - Fundamentally flawed as applying intervention before randomization introduces selection bias - This violates the core principle of RCTs (random allocation must precede intervention) *Assessment → Randomisation → Follow-up → Manipulation* - Incorrect sequence as assessment (outcome measurement) is the final stage, not the first - Manipulation must follow randomization, not come after follow-up

Q: A study is conducted to compare the mean hemoglobin (Hb) levels between two independent groups. Which statistical test is most appropriate?

Unpaired t-test. ***Unpaired t-test*** - This test is specifically used to compare the **means** of a continuous outcome variable (like hemoglobin level) between **two independent, unrelated groups**. - It is based on the assumption that the data is normally distributed and variances are equal (though modifications exist if variances are unequal, known as Welch's t-test). *Paired t-test* - The paired t-test is used when the data comes from **dependent** or **related groups**, such as measuring the same individuals before and after an intervention (pre-post study). - Since the question specifies two **independent** groups, this test is incorrect. *Chi-square test* - This test is used to analyze the association or difference between **two or more categorical variables** (e.g., comparing proportions or frequencies in nominal data). - It is unsuitable for comparing the **mean** of a **continuous variable** like Hb levels. *ANOVA* - Analysis of Variance (ANOVA) is used to compare the **means** of a continuous variable among **three or more independent groups**. - Since the study involves only **two groups**, the unpaired t-test is the simpler and more conventional choice, although ANOVA yields the same result when reduced to two groups.

Q: In a village of 100 children, 10 children have a past history of measles (i.e., they are not at risk now), 20 new cases of measles were reported this year. What is the incidence of measles in this population for the year?

22.22 %. ***22.22 %*** - Incidence is calculated as the ratio of **new cases** (20) to the **population at risk** (susceptible population) over the specified period. - The **population at risk** is the total population (100) minus those who are already immune (10), making the denominator 90. Incidence = (20/90) × 100 = **22.22 %**. *20 %* - This result is obtained by incorrectly using the **total population** (100) as the denominator (20/100 × 100), ignoring the already immune group. - Using the total population in the denominator leads to an underestimate of the true **attack rate** or incidence among the susceptible group. *10 %* - This figure represents the proportion of children who had suffered from measles in the past (10/100), reflecting a form of **past prevalence**, not incidence. - Incidence focuses exclusively on the **new cases** that developed within the year. *30 %* - This percentage represents the **cumulative prevalence** at the end of the year, including both old (10) and new (20) cases, divided by the total population (30/100). - Incidence requires the denominator to be the **population at risk** (those who could develop the disease), not the total population.

Q: The given image shows:

Spaghetti plot. ***Spaghetti plot*** - A **spaghetti plot** displays multiple overlapping individual trajectory lines, with each line representing one subject's data over time or across conditions, creating a visual that resembles **strands of spaghetti**. - It is commonly used in **longitudinal studies** to show individual patient responses or changes, allowing visualization of both individual patterns and overall trends. *Kaplan Meier plot* - A Kaplan-Meier plot is a **survival curve** that shows the probability of survival over time with characteristic **step-wise decreases** at each event occurrence. - It is specifically used for **time-to-event analysis** and does not show multiple individual trajectory lines like a spaghetti plot. *Funnel plot* - A funnel plot is used in **meta-analyses** to assess **publication bias** by plotting study effect sizes against their precision (standard error). - It creates a **symmetrical funnel shape** when no bias is present and does not show individual subject trajectories over time. *Forest plot* - A forest plot is used in **meta-analyses** to display individual study results and **pooled effect estimates** with confidence intervals. - It features **squares or diamonds** representing effect sizes with **horizontal lines** for confidence intervals, not overlapping trajectory lines.

Q: The given image shows:

Funnel plot. ***Funnel plot*** - This graph displays individual study results (trial size vs. effect size), forming a **funnel-like shape** when no publication bias is present. - The observed asymmetry in the plot, where many small trials with a larger effect size are on one side, suggests potential **publication bias**, indicating this is a funnel plot. *Kaplan Meier plot* - A Kaplan-Meier plot is used to estimate the **survival function** from lifetime data. - It displays the probability of an event (e.g., death, disease remission) over time, characterized by a **stepped curve**. *Spaghetti plot* - A spaghetti plot is used to visualize longitudinal data, showing **multiple individual trajectories** over time. - Each line in a spaghetti plot represents data from a **single subject** or entity across different time points. *Forest plot* - A forest plot graphically presents the results of individual studies included in a **meta-analysis** and the overall pooled estimate. - It typically shows **effect size** and **confidence intervals** for each study, often represented by squares and horizontal lines.

Q: The given image shows:

Funnel plot. **Funnel plot** - A **funnel plot** displays the **effect sizes** from individual studies against a measure of their precision (e.g., sample size or standard error), typically in the context of a meta-analysis. - The characteristic **funnel shape** (or inverted triangle) arises because smaller studies (lower precision) have wider confidence intervals and thus show more variability, while larger studies (higher precision) cluster more closely around the true effect size. *Kaplan Meier plot* - A **Kaplan-Meier plot** is used to estimate and display **survival probabilities** over time, particularly in clinical trials or observational studies. - It shows a stepwise-decreasing curve as events (e.g., death, disease recurrence) occur, which is not what is depicted in the image. *Spaghetti plot* - A **spaghetti plot** displays the **individual trajectories** of multiple subjects over time, often used to visualize longitudinal data. - Each line in a spaghetti plot represents a single subject's path, which is distinct from the scattered points representing individual studies on the given image. *Forest plot* - A **forest plot** is a graphical display used in **meta-analyses** to present the results of individual studies and their combined effect. - It typically shows the effect estimate and confidence interval for each study as horizontal lines, with a diamond representing the overall pooled estimate, which is different from the funnel shape seen here.

Question 1

In statistics, centiles and quartiles are considered as:

Accepted Answer

Measures of location/position

Answer

Measures of dispersion

Answer

Measures of central tendency

Answer

Measures of correlation

Question 2

In a study done in a hospital, patients were categorized into three groups based on disease prevalence (Low, Medium, High), and individuals were then randomly selected from each group. What type of sampling is this?

Accepted Answer

Stratified random sampling

Answer

Systematic random sampling

Answer

Cluster random sampling

Answer

Simple random sampling

Question 3

In a given population:
- less than 15 years: 6000
- 15 to 64 years: 12000
- 65 years and above: 2000

Calculate the dependency ratio.

Accepted Answer

66 %

Answer

50 %

Answer

75 %

Answer

33 %

Question 4

A new community intervention is initiated to reduce perinatal sepsis. Researchers allocate 20 Primary Health Centres (PHCs) to receive standard care and 20 PHCs to receive a community-based intervention. What type of study design is this?

Accepted Answer

Cluster-randomized control trial

Answer

Cross-sectional study

Answer

Case-control study

Answer

Quasi-experimental study

Question 5

Which of the following is the correct sequence of steps in a Randomized Controlled Trial (RCT)?

Accepted Answer

Randomisation → Manipulation → Follow-up → Assessment

Answer

Follow-up → Manipulation → Assessment → Randomisation

Answer

Manipulation → Assessment → Follow-up → Randomisation

Answer

Assessment → Randomisation → Follow-up → Manipulation

Question 6

A study is conducted to compare the mean hemoglobin (Hb) levels between two independent groups. Which statistical test is most appropriate?

Accepted Answer

Unpaired t-test

Answer

Paired t-test

Answer

Chi-square test

Answer

ANOVA

Question 7

In a village of 100 children, 10 children have a past history of measles (i.e., they are not at risk now), 20 new cases of measles were reported this year. What is the incidence of measles in this population for the year?

Accepted Answer

22.22 %

Answer

10 %

Answer

30 %

Answer

20 %

Question 8

The given image shows:

Accepted Answer

Spaghetti plot

Answer

Kaplan Meier plot

Answer

Funnel plot

Answer

Forest plot

Question 9

The given image shows:

Accepted Answer

Funnel plot

Answer

Kaplan Meier plot

Answer

Spaghetti plot

Answer

Forest plot

Question 10

The given image shows:

Accepted Answer

Funnel plot

Answer

Kaplan Meier plot

Answer

Spaghetti plot

Answer

Forest plot

Biostatistics — MCQs

Biostatistics — MCQs

On this page

Practice by Chapter

Want unlimited practice?