Biostatistics Practice Questions

Q: In a cohort study spanning 20 years, 50 out of 5000 smokers developed lung cancer, and 10 out of 10000 non-smokers developed lung cancer. What is the 'relative risk' among smokers for developing lung cancer?

10. ***Correct Answer: 10*** - The incidence of lung cancer in smokers = 50/5000 = 0.01 (1%) - The incidence of lung cancer in non-smokers = 10/10000 = 0.001 (0.1%) - **Relative Risk (RR) = Incidence in exposed / Incidence in unexposed** - RR = 0.01 / 0.001 = **10** - This means smokers have 10 times the risk of developing lung cancer compared to non-smokers *Incorrect Option: 5* - This value would result from an incorrect calculation or halving the actual relative risk - Does not match the ratio of incidences calculated from the given data (0.01/0.001 ≠ 5) - Would underestimate the true risk among smokers *Incorrect Option: 45* - This does not represent any standard epidemiological measure from this data - May result from confusion with absolute numbers or incorrect arithmetic - Neither the absolute risk difference nor any valid ratio yields this number *Incorrect Option: 50* - This represents the **absolute number of cases** in the smoking cohort, not a risk measure - Relative risk is a **ratio** comparing incidence rates between groups, not a count - Common error: confusing absolute numbers with relative measures

Q: Why is matching done in a case-control study?

To remove the effect of known confounders. ***To remove the effect of known confounders*** - **Matching** in a case-control study helps to control for the influence of specific variables (known confounders) that might otherwise distort the observed association between the exposure and outcome. - By matching cases and controls on characteristics like age, sex, or socioeconomic status, researchers ensure that these factors are similarly distributed in both groups, isolating the effect of the primary exposure. *To remove the effect of unknown confounders* - Matching is effective for **known confounders** that are identified and controlled during study design. - It does not address the impact of **unknown** or unmeasured confounding variables, which can still influence the study's results. *To eliminate selection bias* - **Selection bias** occurs when participants are not representative of the target population; matching primarily addresses confounding, not the initial selection process. - While careful selection is crucial to minimize selection bias, matching itself is more related to controlling for nuisance variables *after* selection. *To eliminate interviewer’s bias* - **Interviewer bias** (or observer bias) arises when the interviewer's expectations or knowledge influence data collection or interpretation. - This type of bias is typically addressed through blinding (e.g., blinded interviewers, participants, or outcome assessors), not through matching study participants.

Q: Retrospective cohort studies have the following features EXCEPT:

Generally more expensive than prospective studies. ***Generally more expensive than prospective studies*** - Retrospective cohort studies are typically **less expensive** than prospective studies because they utilize existing data, thus avoiding the costs associated with new data collection and long-term follow-up. - The primary expenses in retrospective studies often involve data retrieval and analysis, which are generally lower compared to the extensive resources needed for prospective data acquisition. *Investigator goes back in time to select study groups* - This is a hallmark feature of **retrospective cohort studies**, where researchers define study groups (exposed and unexposed) based on past exposures. - Data collection then proceeds by looking forward from the exposure to identify health outcomes that have already occurred. *Outcomes have occurred before the start of the study* - In a **retrospective cohort study**, both the exposure and the **outcomes of interest** have already taken place before the study officially begins. - Researchers identify existing records to link past exposures to these pre-existing outcomes. *Results are obtained more quickly* - Because all exposures and outcomes have already occurred and data is often readily available, **retrospective studies** can generate results much faster than prospective studies. - They do not require waiting for events to unfold in real-time, which significantly reduces the duration of the research process.

Q: In a normal distribution, Mean ± 2 S.D. contains

95.4 % values. ***95.4 % values*** - According to the **empirical rule** (or 68-95-99.7 rule) for normal distributions, approximately **95.4%** of data falls within two standard deviations of the mean. - This interval covers from (Mean - 2 S.D.) to (Mean + 2 S.D.) and represents the likelihood of a value falling in this range. *68.3 % values* - This percentage corresponds to the data contained within **Mean ± 1 S.D.** in a normal distribution, not Mean ± 2 S.D. - It signifies that roughly two-thirds of all observations lie within one standard deviation from the mean in a bell-shaped curve. *91.2 % values* - This value is not a standard percentage associated with common multiples of standard deviations (1, 2, or 3) from the mean in a normal distribution. - It does not correspond to any universally recognized interval like ±1 S.D., ±2 S.D., or ±3 S.D. *99.7 % values* - This percentage represents the data contained within **Mean ± 3 S.D.** in a normal distribution. - It indicates that almost all (99.7%) of the data points are expected to fall within three standard deviations from the mean.

Q: The appropriate statistical test to find out obesity as a significant risk factor for breast cancer is:

Chi-square test. ***Chi-square test*** - The **chi-square test** is used to determine if there is a **significant association** between two **categorical variables**. - In this scenario, both obesity (yes/no) and breast cancer (yes/no) are categorical, making chi-square appropriate to assess if obesity is a risk factor. *Wilcoxon’s signed rank test* - This is a **non-parametric test** used for comparing two related samples or repeated measurements on a single sample, especially when data are not normally distributed. - It is not suitable for assessing the association between two independent categorical variables like obesity and breast cancer. *Student’s paired ‘t’ test* - The **paired t-test** is used to compare the means of two related groups or measurements from the same subject under different conditions (e.g., before and after an intervention). - This test is designed for **continuous data** and would not be appropriate for the categorical variables of obesity and breast cancer. *Student’s unpaired ‘t’ test* - The **unpaired t-test** (also known as independent samples t-test) is used to compare the means of two independent groups for a **continuous outcome variable**. - It is not suitable when both the exposure (obesity) and the outcome (breast cancer) are categorical variables.

Q: Which of the following is/are the measure(s) of dispersion? 1. Mode 2. Median 3. Standard Deviation Select the correct answer using the code given below:

3 only. ***Correct: 3 only*** - **Standard Deviation** is a direct measure of dispersion that quantifies the amount of variation or spread of data values around the mean - It indicates how much individual data points deviate from the average, making it a key statistic for understanding the **spread** within a dataset - Other common measures of dispersion include **range, variance, interquartile range, and coefficient of variation** *Incorrect: 1, 2 and 3* - **Mode** and **Median** are measures of **central tendency**, not dispersion - They describe the center or typical value of a dataset, not the spread or variability - While they provide insight into the data's distribution, they do not quantify how spread out the data points are *Incorrect: 2 and 3 only* - **Median** is a measure of **central tendency** representing the middle value when data is ordered, not a measure of dispersion - Only **Standard Deviation** from this option is a measure of dispersion, making this choice incorrect *Incorrect: 1 and 2 only* - Both **Mode** and **Median** are measures of **central tendency** - Mode indicates the most frequent value and Median represents the middle value - Neither provides information about how **spread out** or dispersed the data points are around the center

Q: Analytical studies include the following methods of studies except:

Cross sectional studies. ***Cross sectional studies*** - **Cross-sectional studies** are generally classified as **descriptive studies** as they assess the prevalence of a disease and/or risk factors at a single point in time, rather than analyzing cause-and-effect relationships. - While they can identify associations, they cannot establish **temporality** or causality between an exposure and an outcome, which is a hallmark of analytical studies. *Randomised controlled trials* - **Randomized controlled trials (RCTs)** are the gold standard for analytical studies, specifically for evaluating interventions, due to their ability to establish **causality** through random assignment and control groups. - They are prospective in nature and aim to assess the effect of an intervention on an outcome, controlling for confounding. *Cohort studies* - **Cohort studies** are a type of **observational analytical study** where a group of individuals exposed to a risk factor and a group not exposed are followed over time to compare disease incidence. - They are crucial for studying rare exposures and the natural history of diseases, allowing for the calculation of incidence rates and relative risks. *Case control studies* - **Case-control studies** are **retrospective analytical studies** that compare individuals with a disease (cases) to individuals without the disease (controls) and look back in time to identify past exposures. - These studies are efficient for investigating rare diseases and multiple exposures for a single outcome, allowing for the calculation of odds ratios.

Q: Which one of the following is not a measure of dispersion?

Mean. ***Mean*** - The **mean** is a measure of **central tendency**, representing the average value of a dataset. - It describes the typical value around which data points cluster, rather than how spread out they are. *Mean deviation* - **Mean deviation** is a measure of **dispersion** that calculates the average of the absolute differences between each data point and the mean of the dataset. - It quantifies the average deviation of data points from the center. *Standard deviation* - **Standard deviation** is a widely used measure of **dispersion** that indicates the average amount of variability or spread around the mean. - A higher standard deviation suggests data points are more spread out from the mean. *Range* - The **range** is a simple measure of **dispersion** calculated as the difference between the highest and lowest values in a dataset. - It provides a basic idea of the total spread of data from its extremes.

Q: The agreement (yes/no) between two observers is statistically measured by:

Kappa coefficient. **Kappa coefficient** - The **kappa coefficient** measures the **inter-rater agreement** for qualitative items, such as a "yes/no" decision, beyond what would be expected by chance. - It takes into account the observed agreement and the agreement expected by chance, providing a more robust measure of agreement than simple percentage agreement. *Correlation coefficient* - The **correlation coefficient** measures the **strength and direction of a linear relationship between two quantitative variables**, not the agreement between two observers on a categorical outcome. - It is used for continuous data and indicates how closely data points fit a linear regression line. *Sensitivity* - **Sensitivity** is a measure of a test's ability to correctly identify individuals who **have a disease (true positive rate)**. - It is not used to assess the agreement between two observers but rather the performance of a diagnostic test against a gold standard. *Specificity* - **Specificity** is a measure of a test's ability to correctly identify individuals who **do not have a disease (true negative rate)**. - Like sensitivity, it evaluates the performance of a diagnostic test and not the consistency of observations between two different raters.

Question 1

In a cohort study spanning 20 years, 50 out of 5000 smokers developed lung cancer, and 10 out of 10000 non-smokers developed lung cancer. What is the 'relative risk' among smokers for developing lung cancer?

Accepted Answer

10

Answer

5

Answer

45

Answer

50

Question 2

Why is matching done in a case-control study?

Accepted Answer

To remove the effect of known confounders

Answer

To remove the effect of unknown confounders

Answer

To eliminate selection bias

Answer

To eliminate interviewer’s bias

Question 3

Retrospective cohort studies have the following features EXCEPT:

Accepted Answer

Generally more expensive than prospective studies

Answer

Investigator goes back in time to select study groups

Answer

Outcomes have occurred before the start of the study

Answer

Results are obtained more quickly

Question 4

In a normal distribution, Mean ± 2 S.D. contains

Accepted Answer

95.4 % values

Answer

68.3 % values

Answer

91.2 % values

Answer

99.7 % values

Question 5

In a case control study, confounding factors can be minimized by the following EXCEPT:

Accepted Answer

Increasing sample size for cases and controls

Answer

Matching of variables such as age and sex

Answer

Stratification during analysis

Answer

Randomization during selection

Question 6

The appropriate statistical test to find out obesity as a significant risk factor for breast cancer is:

Accepted Answer

Chi-square test

Answer

Wilcoxon’s signed rank test

Answer

Student’s paired ‘t’ test

Answer

Student’s unpaired ‘t’ test

Question 7

Which of the following is/are the measure(s) of dispersion?

1. Mode

2. Median

3. Standard Deviation
Select the correct answer using the code given below:

Accepted Answer

3 only

Answer

1, 2 and 3

Answer

2 and 3 only

Answer

1 and 2 only

Question 8

Analytical studies include the following methods of studies except:

Accepted Answer

Cross sectional studies

Answer

Randomised controlled trials

Answer

Cohort studies

Answer

Case control studies

Question 9

Which one of the following is not a measure of dispersion?

Accepted Answer

Mean

Answer

Mean deviation

Answer

Standard deviation

Answer

Range

Question 10

The agreement (yes/no) between two observers is statistically measured by:

Accepted Answer

Kappa coefficient

Answer

Correlation coefficient

Answer

Sensitivity

Answer

Specificity

Biostatistics — MCQs

Biostatistics — MCQs

On this page

Practice by Chapter

Want unlimited practice?