Which statistical test compares the means of two independent groups?
Answer: Unpaired t-test
The independent t-test compares the means of two separate groups (for example, blood pressure in men versus women). The paired t-test compares two measurements in the same subjects. ANOVA compares three or more means. Chi-square compares proportions.
Chi-square test Compares categorical data (proportions).
One-way ANOVA Three or more groups.
Pearson correlation Measures association between two continuous variables.
Paired t-test Same subjects measured twice.
Question 2Medium
Recording whether vomiting is present or absent gives which type of data?
Answer: Nominal (binary)
Data with two categories and no order (yes/no, present/absent, male/female) are nominal, also called binary or dichotomous. Ordinal data are ranked categories (mild, moderate, severe). Interval and ratio data are numerical; ratio data have a true zero (weight).
Types of data
Nominal: categories, no order (yes/no, blood group)
Ordinal: ranked (pain scale, cancer stage)
Interval/ratio: numbers (ratio has a true zero)
Remember NOIR: Nominal, Ordinal, Interval, Ratio.
Why the other options are wrong
Ordinal (ranked) Needs an order between categories.
Interval scale Numerical with no true zero (temperature °C).
Ratio scale Numerical with a true zero.
Continuous Numerical measurements.
Question 3Medium
In a case–control study, mothers of babies with birth defects remember medicines taken in pregnancy more thoroughly than mothers of healthy babies. What type of bias is this?
Answer: Recall bias
Recall bias occurs when cases remember past exposures differently from controls, common in case–control studies. Selection bias comes from how participants are chosen. Lead-time bias makes survival look longer with earlier diagnosis. Observer bias comes from assessors knowing groups. Attrition bias comes from loss to follow-up.
Common biases
Recall: cases remember differently
Lead-time: early diagnosis looks like longer survival
Selection, observer, attrition, confounding (not a bias but a mix-up)
Remember Recall bias lives in case–control studies.
Why the other options are wrong
Selection bias From how participants are recruited.
Lead-time bias Earlier diagnosis seems to prolong survival.
The maternal mortality ratio is the number of women dying from pregnancy-related causes (during pregnancy or within 42 days of its end) per 100,000 live births. The infant mortality rate is infant deaths under 1 year per 1,000 live births; the neonatal mortality rate is deaths in the first 28 days per 1,000 live births.
Vital statistics
MMR: per 100,000 live births
IMR: < 1 year per 1,000 live births
Neonatal: < 28 days per 1,000 live births
Remember Mothers per 100,000; infants per 1,000.
Why the other options are wrong
Maternal deaths per 1,000 women Not the standard denominator.
Maternal deaths per 1,000 live births Wrong multiplier.
Maternal deaths per 100 pregnancies Not the standard measure.
Infant deaths per 1,000 live births That is the infant mortality rate.
Total fertility rate is the sum of age-specific fertility rates: the average number of children born to a woman over her lifetime. Replacement level is about 2.1. Crude birth rate is births per 1,000 mid-year population. General fertility rate is births per 1,000 women aged 15–49. Net reproduction rate counts daughters surviving.
Fertility measures
TFR: children per woman (replacement ~2.1)
CBR: births per 1,000 population
GFR: births per 1,000 women 15–49
Remember TFR: how many children a woman has on average.
Why the other options are wrong
Births per 1,000 population Crude birth rate.
Births per 1,000 women 15–49 General fertility rate.
Daughters per woman Gross/net reproduction rate.
Proportion of married women Not a fertility measure.
Question 6Medium
Fasting glucose in a population is normally distributed with mean 5.0 mmol/L and SD 0.5. What percentage of people have values above 6.0 mmol/L?
Answer: About 2.5%
6.0 is 2 SD above the mean. About 95% of values lie within ±2 SD (more precisely 1.96), leaving 5% outside, split equally: 2.5% above and 2.5% below. 1 SD covers 68% (16% above), 3 SD covers 99.7% (0.15% above). Reference ranges are usually the central 95%.
Normal curve tails
±1 SD: 68% (16% each tail)
±2 SD: 95% (2.5% each tail)
±3 SD: 99.7%
Remember Two SDs out leaves 2.5% on each side.
Why the other options are wrong
About 5% Both tails together.
About 16% Above 1 SD.
About 32% Outside 1 SD both sides.
About 0.15% Above 3 SD.
Question 7Medium
The probability that a baby is a boy is 0.5, and the probability of blood group O is 0.4, independently. What is the probability that a baby is a boy with blood group O?
Answer: 0.2
For independent events, the probability that both occur is the product of their probabilities (multiplication rule): 0.5 × 0.4 = 0.2. The addition rule is used for 'either/or': for mutually exclusive events, P(A or B) = P(A) + P(B). For events that can occur together, P(A or B) = P(A) + P(B) − P(A and B). Adding 0.5 and 0.4 would answer a different question.
Probability rules
Both (independent): multiply
Either (mutually exclusive): add
Either (overlapping): add − both
Probabilities lie between 0 and 1
Remember And means multiply; or means add.
Exam trap Multiply only when events are independent.
Why the other options are wrong
0.9 This adds the probabilities.
0.1 This subtracts them.
0.45 This averages them.
0.5 This ignores blood group.
Question 8Easy
A 50-g glucose challenge test is evaluated in an antenatal clinic. Of 120 women later confirmed by OGTT to have gestational diabetes, 102 had a positive challenge test. What is the sensitivity of the challenge test?
Answer: 85%
Sensitivity is the proportion of people with the disease who test positive: true positives divided by all who have the disease. Here it is 102/120 = 85%. The remaining 15% are false negatives. Sensitivity is calculated only from the people who have the disease, so the number of women without diabetes is not needed.
Sensitivity
TP ÷ (TP + FN)
Uses only the diseased
Here 102 ÷ 120 = 85%
High sensitivity rules out (SnOut)
Remember Sensitivity looks only at those who have the disease.
Exam trap Sensitivity and false-negative rate always add up to 100%.
Why the other options are wrong
15% This is the false-negative rate.
102% A proportion cannot exceed 100%.
54% Not derived from these numbers.
98% Not derived from these numbers.
Question 9Medium
In a trial in women at risk of preterm birth, respiratory distress syndrome occurred in 20% of babies whose mothers received placebo and in 12% of those whose mothers received antenatal corticosteroids. How many women need to be treated to prevent one case?
Answer: 13
The absolute risk reduction is 20% − 12% = 8% (0.08). The number needed to treat is 1 ÷ ARR = 1 ÷ 0.08 = 12.5, which is rounded up to 13, because only whole patients can be treated and rounding down would overstate the benefit. So about 13 women must receive steroids to prevent one baby from developing respiratory distress syndrome.
Number needed to treat
ARR = 20% − 12% = 8%
NNT = 1 ÷ ARR
1 ÷ 0.08 = 12.5, round up to 13
Smaller NNT = more effective
Remember NNT is one over the absolute risk difference, rounded up.
Exam trap Always round the NNT up, never down.
Why the other options are wrong
8 This is the absolute risk reduction in percent.
40 This is the relative risk reduction in percent.
5 Uses 1 ÷ 0.20, the control risk.
60 Uses the ratio of risks, not their difference.
Question 10Medium
Cytology-based screening has reduced deaths from cervical cancer, while screening with CA-125 and ultrasound has not clearly reduced deaths from ovarian cancer. Which feature of cervical cancer mainly makes screening for it effective?
Answer: A long detectable pre-invasive phase
A key screening criterion is a recognisable latent or early stage. Cervical cancer develops over many years through cervical intraepithelial neoplasia, which can be detected by cytology or HPV testing and treated easily before invasion. Most high-grade serous ovarian cancers have no long detectable precursor stage in an accessible site, so screening tends to find them already advanced. Incidence alone does not make screening effective.
Why cervical screening works
Long pre-invasive phase (CIN)
Simple, acceptable test
Effective treatment of precursors
Fits Wilson–Jungner criteria
Remember Screening works best when there is a long, treatable early stage.
Exam trap Screening a disease without a detectable early phase mostly finds late disease and adds harm.
Why the other options are wrong
A higher incidence than ovarian cancer Incidence alone does not make screening work.
A worse prognosis once invasive Prognosis of invasive disease is not the reason.
A lower cost of treatment Cost is not the main reason.
A stronger genetic tendency Most cervical cancers are HPV-related, not hereditary.
Question 11Hard
A heat-stable uterotonic that needs no refrigeration is compared with oxytocin for preventing postpartum haemorrhage. The investigators only aim to show that it is not worse than oxytocin by more than a pre-specified margin, because it would be valuable in hot regions without cold chains. What type of trial is this?
Answer: Non-inferiority trial
A non-inferiority trial tests whether a new treatment is not unacceptably worse than an established one, by more than a pre-defined margin. It is used when the new treatment has other advantages, such as stability without refrigeration, lower cost or easier use. A superiority trial aims to show the new treatment is better. Crossover, factorial and cluster designs describe how participants receive treatments, not the hypothesis.
Non-inferiority trial
Aim: not worse by more than a margin
Margin set in advance
New drug has other advantages
Needs per-protocol and ITT analyses
Remember Not better, just not worse, but easier.
Exam trap Failing to show superiority is not the same as showing non-inferiority.
Why the other options are wrong
Superiority trial Aims to show the new drug is better.
Crossover trial Each participant gets both treatments in turn.
Factorial trial Tests two or more interventions at once.
Cluster randomised trial Randomises groups rather than individuals.
Question 12Medium
After a home delivery in a village near Rajanpur, a woman bleeds heavily. The family waits four hours for her husband to return from the fields and give permission before calling for transport. Which delay does this represent?
Answer: Delay in deciding to seek care
The three delays model explains how maternal deaths occur: delay in deciding to seek care (first delay), delay in reaching a facility (second) and delay in receiving adequate care once there (third). The first delay arises from failure to recognise danger signs, low status of women in decision-making, cost or cultural barriers. Here the family recognised the bleeding but waited for the husband's permission, a classic first delay that is common in rural Pakistan.
Three delays
1st: deciding to seek care
2nd: reaching the facility
3rd: receiving adequate care
Household decision-making drives the first delay
Remember Decide, reach, receive.
Exam trap Waiting at home for permission is the first delay even if transport is also poor.
Why the other options are wrong
Delay in reaching a health facility Transport and distance problems are the second delay.
Delay in receiving care at the facility Shortage of staff, blood or drugs at the facility is the third delay.
Delay in referral between facilities Not one of the three delays as originally described.
Delay in recognising labour onset Labour had ended; the problem was the decision to seek help.
Topics covered
Epidemiology, Descriptive statistics, Screening, Study design, Sampling, Hypothesis testing, Bias, Data presentation, Statistical tests, Data types, Risk measures, Immunisation, Public health, Screening test validity, Trials, Evidence, Vital statistics, Environmental health, Probability, Tests, Correlation, Confounding, Demography, Nutrition, Community health, Screening tests, Predictive values, Bias in screening, Phases of drug trials, Three delays model, Research methods, Clinical trials, Inference, Measurement, Survival, Reliability, Sample size, Surveillance, Normal distribution, Errors
Frequently asked questions
How many FCPS Part 1 Biostatistics questions are in the bank?
There are 249 Biostatistics questions in FCPS-1 Prep, each with an explanation for every option.
Are these past paper questions?
No. Every question is original, written by us and mapped to the published CPSP FCPS Part 1 syllabus.
