I85PREP
⌘ K
HS 326 study materials
HS 326 study resources

HS 326 practice questions

One question at a time. Build your HS 326 understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1Scientific Inquiry and Research Design

A researcher wants to determine whether a new tutoring program improves exam scores. Which design provides the strongest evidence of a causal effect?

  1. A correlational study relating hours of tutoring to scores across one school
  2. A randomized controlled trial assigning students to tutoring or a waitlist control Correct answer.
  3. A case series describing score changes in ten tutored students
  4. A cross-sectional survey comparing students who chose tutoring with those who did not

Explanation

Random assignment balances known and unknown confounders across groups, so post-intervention differences can be attributed to the program. Observational designs such as surveys, case series, and correlations cannot rule out self-selection and other confounders.

Question 2Scientific Inquiry and Research Design

In a clinical trial, participants are assigned to treatment or placebo by a random process, but neither participants nor assessors know group membership. Which two design features are described?

  1. Sampling and weighting
  2. Counterbalancing and blocking
  3. Matching and stratification
  4. Randomization and blinding Correct answer.

Explanation

Random allocation to conditions is randomization, and concealing group membership from participants and assessors is blinding. These features address confounding and expectancy bias respectively.

Question 3Scientific Inquiry and Research Design

A study reports a p-value of 0.03 for the difference between two groups. What does this value most directly indicate?

  1. The proportion of the outcome variance explained by group membership
  2. The probability that the result will replicate in a future study
  3. The probability of observing a difference at least this extreme if the null hypothesis were true Correct answer.
  4. The probability that the null hypothesis is true given the observed data

Explanation

A p-value is the probability of data at least as extreme as observed, computed under the assumption that the null hypothesis holds. It is not the probability the null is true, nor a replication probability or effect-size measure.

Question 4Scientific Inquiry and Research Design

A researcher measures depression before treatment, after treatment, and at six-month follow-up in the same participants. Which analysis approach is most appropriate for this repeated-measures structure?

  1. An independent-samples t-test comparing first and last measurements
  2. A mixed or repeated-measures model accounting for within-person correlation Correct answer.
  3. A chi-square test of independence on the three time points
  4. A simple linear regression ignoring participant identity

Explanation

Repeated observations on the same person are correlated, so models with a within-subject component or random effects are needed for valid standard errors. Treating time points as independent observations violates that assumption.

Question 5Scientific Inquiry and Research Design

A survey uses a seven-point Likert scale to measure job satisfaction. Which level of measurement does a single item represent?

  1. Interval
  2. Nominal
  3. Ordinal Correct answer.
  4. Ratio

Explanation

Likert response categories are ordered but the distances between adjacent categories are not guaranteed equal, so a single item is ordinal. Interval and ratio properties require equal units, which ordinal scales do not establish.

Question 6Scientific Inquiry and Research Design

A screening test is evaluated against a gold-standard diagnosis. Which quantity describes the proportion of people who truly have the condition and test positive?

  1. Prevalence
  2. Sensitivity Correct answer.
  3. Specificity
  4. Positive predictive value

Explanation

Sensitivity is the true-positive rate: among those with the condition, the fraction testing positive. Specificity refers to those without the condition testing negative, and predictive values condition on the test result rather than disease status.

Question 7Scientific Inquiry and Research Design

A researcher wants to estimate a population mean with a 95% confidence interval and a margin of error of 2 units, assuming a population standard deviation of 10. Which sample size is closest to the requirement?

  1. 400
  2. 96 Correct answer.
  3. 50
  4. 25

Explanation

Using n = (z·σ/E)² with z ≈ 1.96, σ = 10, and E = 2 gives (1.96×10/2)² ≈ 96. This is the standard sample-size formula for a mean with known variance.

Question 8Scientific Inquiry and Research Design

A trial reports a relative risk reduction of 50% for an event that occurs in 2% of controls. Which statement correctly interprets the absolute effect?

  1. The absolute risk reduction is 1 percentage point Correct answer.
  2. The absolute risk reduction is 50 percentage points
  3. The number needed to treat is 2
  4. The control event rate is 50%

Explanation

A 50% relative reduction applied to a 2% control risk lowers the risk to 1%, an absolute reduction of 1 percentage point. The number needed to treat is the reciprocal, about 100.

Question 9Scientific Inquiry and Research Design

A researcher recruits participants by advertising in a fitness magazine to study exercise habits among the general adult population. Which threat to inference is most prominent?

  1. Attrition bias from longitudinal dropout
  2. Selection bias from a non-representative sampling frame Correct answer.
  3. Measurement error from the survey instrument
  4. Recall bias from asking about childhood events

Explanation

Readers of a fitness magazine differ systematically from the general population in exercise behavior, so the sample is not representative. This is a sampling-frame problem rather than an issue of instrument accuracy or follow-up loss.

Question 10Scientific Inquiry and Research Design

Which statement best distinguishes internal validity from external validity?

  1. Internal validity concerns whether the observed effect is causal within the study; external validity concerns generalization beyond it Correct answer.
  2. Internal validity concerns generalization to other populations; external validity concerns causal inference within the sample
  3. Internal validity concerns ethical review; external validity concerns publication status
  4. Internal validity concerns statistical power; external validity concerns measurement reliability

Explanation

Internal validity asks whether the design supports a causal claim for the studied sample, while external validity asks whether findings extend to other settings, times, and people. They are distinct and can trade off against each other.

Question 11Scientific Inquiry and Research Design

A researcher hypothesizes that sleep duration mediates the relationship between caffeine intake and reaction time. Which analysis directly tests the mediating pathway?

  1. Comparing mean reaction time across caffeine dose groups
  2. Computing the correlation between caffeine intake and reaction time
  3. Regressing reaction time on caffeine intake and age
  4. Estimating the indirect effect of caffeine on reaction time through sleep duration Correct answer.

Explanation

Mediation is tested by estimating the indirect path from predictor through mediator to outcome, typically with bootstrapped confidence intervals. Simple correlations or group comparisons do not decompose the pathway.

Question 12Scientific Inquiry and Research Design

In a qualitative study, a researcher interviews participants until additional interviews yield no new themes. What concept does this describe?

  1. Triangulation
  2. Member checking
  3. Data saturation Correct answer.
  4. Purposive sampling

Explanation

Saturation is the point at which further data collection produces no new insights or themes. Triangulation and member checking are verification strategies, and purposive sampling is a selection approach.

Question 13Scientific Inquiry and Research Design

A researcher wants to compare three teaching methods on exam scores while controlling for prior GPA. Which technique is most suitable?

  1. Pearson correlation between method and score
  2. One-way analysis of variance without GPA
  3. Independent-samples t-tests for each pair of methods
  4. Analysis of covariance with method as a factor and GPA as a covariate Correct answer.

Explanation

ANCOVA compares group means while adjusting for a continuous covariate, increasing precision and reducing confounding by prior achievement. Plain ANOVA leaves GPA uncontrolled.

Question 14Scientific Inquiry and Research Design

A study finds an effect with p = 0.04 but a very small standardized effect size and a wide confidence interval. Which conclusion is most defensible?

  1. Statistical significance is present, but the practical importance and precision of the estimate are limited Correct answer.
  2. The result should be discarded because p exceeds 0.01
  3. The effect is large and clinically meaningful because p is below 0.05
  4. The wide interval confirms the effect is robust

Explanation

Significance indicates the result is unlikely under the null, but a small effect size and wide interval suggest limited practical importance and imprecise estimation. Significance and magnitude answer different questions.

Question 15Scientific Inquiry and Research Design

A researcher plans to test the same hypothesis in four independent subgroups and will declare success if any subgroup reaches p < 0.05. Which issue does this raise?

  1. Loss of external validity from subgroup analysis
  2. Reduced statistical power from splitting the sample
  3. Increased risk of measurement error
  4. Inflated familywise Type I error from multiple comparisons Correct answer.

Explanation

Testing multiple hypotheses at the 0.05 level raises the chance of at least one false positive, roughly to 1 − 0.95⁴. Corrections such as Bonferroni or preregistered subgroup plans address this.

Question 16Scientific Inquiry and Research Design

Which practice most directly reduces questionable research practices such as selective reporting of outcomes?

  1. Reporting only results that reach statistical significance
  2. Removing outliers until the effect becomes significant
  3. Increasing the sample size after seeing preliminary results
  4. Preregistering hypotheses, outcomes, and analysis plans before data collection Correct answer.

Explanation

Preregistration fixes the hypotheses and analyses in advance, making selective reporting and post hoc analytic flexibility visible. The other options are themselves questionable practices.

Question 17Scientific Inquiry and Research Design

A researcher measures blood pressure twice in each participant on the same day and finds the two readings agree closely. Which property does this primarily support?

  1. Construct validity
  2. Internal consistency
  3. Test-retest reliability Correct answer.
  4. Criterion validity

Explanation

Consistency of repeated measurements under stable conditions reflects test-retest reliability. Validity concerns whether the measure captures the intended construct or predicts a criterion, and internal consistency applies to multi-item scales.

Question 18Scientific Inquiry and Research Design

A case-control study finds an odds ratio of 2.5 for an exposure and a disease. Which interpretation is correct?

  1. The risk of disease is 2.5 times higher among the exposed
  2. The odds of exposure are 2.5 times higher among cases than controls Correct answer.
  3. The disease prevalence is 2.5% in the population
  4. The exposure causes disease in 2.5% of cases

Explanation

An odds ratio in a case-control design compares the odds of exposure among cases with the odds among controls. It does not directly estimate disease risk or prevalence, and it is not by itself proof of causation.

Question 19Scientific Inquiry and Research Design

A researcher wants to ensure that a new 20-item anxiety scale measures a single underlying dimension. Which analysis is most appropriate?

  1. Exploratory factor analysis of the item responses Correct answer.
  2. A paired t-test of item means
  3. Cohen's kappa between two raters
  4. A chi-square test of item frequencies

Explanation

Factor analysis examines the covariance structure among items to identify underlying dimensions, making it the standard tool for assessing dimensionality. The other methods address different questions.

Question 20Scientific Inquiry and Research Design

In a randomized trial, 30% of the treatment group and 5% of the control group drop out before the final assessment. Which concern is most serious?

  1. Ceiling effects on the outcome measure
  2. Random measurement error that inflates the variance
  3. Differential attrition that may bias the estimated treatment effect Correct answer.
  4. Insufficient randomization of covariates

Explanation

When dropout differs by group and is related to outcomes, the remaining groups are no longer comparable, biasing the effect estimate. Differential attrition is more threatening than nondifferential error.

Question 21Scientific Inquiry and Research Design

Which statement correctly describes the purpose of a control group in an experiment?

  1. It provides a comparison condition that isolates the effect of the manipulation Correct answer.
  2. It guarantees that the sample is representative of the population
  3. It ensures that measurement is perfectly reliable
  4. It eliminates the need for random assignment

Explanation

A control condition supplies the counterfactual against which the treatment effect is judged, absorbing maturation, history, and other threats. It does not by itself ensure representativeness or reliability.

Question 22Scientific Inquiry and Research Design

A researcher reports an odds ratio with a 95% confidence interval of 1.02 to 3.40. Which conclusion follows?

  1. The result cannot be interpreted without a p-value
  2. The association is not statistically significant because the interval is wide
  3. The association is statistically significant at the 0.05 level but imprecisely estimated Correct answer.
  4. The effect is necessarily large because the upper bound exceeds 3

Explanation

Because the interval excludes 1.0, the association is significant at the 0.05 level, but the wide range indicates substantial uncertainty about magnitude. Confidence intervals convey both significance and precision.

Question 23Scientific Inquiry and Research Design

A researcher uses a random number generator to select 500 adults from a complete municipal registry. Which sampling method is this?

  1. Simple random sampling Correct answer.
  2. Quota sampling
  3. Snowball sampling
  4. Convenience sampling

Explanation

Selecting at random from a complete list of the population is simple random sampling, giving every member a known chance of selection. The other methods involve nonrandom selection mechanisms.

Question 24Scientific Inquiry and Research Design

A crossover trial gives each participant both treatments in random order with a washout period. What is the primary purpose of the washout period?

  1. To increase the sample size available for analysis
  2. To prevent carryover effects from the first treatment into the second period Correct answer.
  3. To blind participants to their allocation
  4. To randomize the order of treatment assignment

Explanation

A washout interval allows the first treatment's effects to dissipate so the second period starts from a comparable baseline. Order randomization and blinding are separate design features.

Question 25Scientific Inquiry and Research Design

A meta-analysis pools studies and finds substantial heterogeneity with I² = 80%. Which interpretation is most appropriate?

  1. All included studies share the same true effect
  2. Publication bias is ruled out by the high I²
  3. The pooled estimate is invalid and should not be reported
  4. Most of the variability in effect estimates is due to between-study differences rather than chance Correct answer.

Explanation

I² estimates the proportion of total variability attributable to heterogeneity rather than sampling error, so 80% indicates considerable inconsistency. It does not by itself invalidate pooling or address publication bias.

Question 26Scientific Inquiry and Research Design

A research team defines the primary outcome, the analysis plan, and the stopping rules before enrolling any participants, then registers these details publicly. Which principle of rigorous research design does this practice most directly support?

  1. Prospective specification of the analysis to limit flexibility after seeing data Correct answer.
  2. Blinding of analysts only after the data have been cleaned
  3. Purposive selection of participants to enrich the sample
  4. Retrospective adjustment of hypotheses to match observed results

Explanation

Pre-specifying outcomes, analyses, and stopping rules before data collection constrains researcher degrees of freedom, which reduces the chance that results are shaped by the observed data. Public registration makes the specification verifiable. The core principle is that confirmatory analyses should be fixed in advance.

Question 27Scientific Inquiry and Research Design

A researcher wants to detect a difference in mean recovery time between two treatments. Which combination of design decisions most directly increases statistical power without changing the alpha level?

  1. Increasing the sample size and reducing measurement error Correct answer.
  2. Adding more secondary outcomes to the analysis plan
  3. Using a one-tailed test and widening the eligibility criteria
  4. Reducing the sample size and relaxing the inclusion criteria

Explanation

Power rises when the effect is easier to detect relative to noise, so larger samples and more precise measurement both help. Measurement error inflates variability and obscures true differences. Alpha remains fixed in this scenario, so the gain must come from sample size or precision.

Question 28Scientific Inquiry and Research Design

A cohort study follows exposed and unexposed workers for ten years. Investigators are concerned that workers who leave the job because of early symptoms are lost to follow-up and never counted as cases. What type of bias does this scenario describe?

  1. Recall bias
  2. Interviewer bias
  3. Ecological fallacy
  4. Attrition bias Correct answer.

Explanation

Differential loss of participants related to the outcome is attrition bias, because those who leave are systematically different from those who remain. If symptomatic workers drop out before being diagnosed, the observed association is distorted. The key feature is unequal and outcome-related dropout over time.

Question 29Scientific Inquiry and Research Design

In a randomized trial, participants are assigned to treatment or placebo, but neither the participants nor the clinicians assessing outcomes know the assignment. Which validity threat is most directly reduced by this design?

  1. Limited generalizability to non-trial populations
  2. Random sampling error in the control group
  3. Confounding by measured baseline covariates
  4. Detection bias from differential outcome ascertainment Correct answer.

Explanation

When assessors know the assignment, outcome ascertainment can differ between groups, producing detection bias. Blinding participants and outcome assessors prevents this differential measurement. Randomization already addresses measured and unmeasured confounding, so blinding targets a different threat.

Question 30Scientific Inquiry and Research Design

A researcher collects data on 200 patients and reports a correlation of r = 0.30 between a biomarker and disease severity, with a p-value of 0.001. Which interpretation is most appropriate?

  1. The p-value proves the biomarker causes worse severity
  2. The biomarker explains most of the variance in severity
  3. The association is statistically significant but modest in magnitude Correct answer.
  4. The result is not meaningful because the sample is small

Explanation

A small p-value indicates that an association of this size is unlikely under the null, but r = 0.30 corresponds to only about 9% shared variance. Statistical significance and effect magnitude are separate considerations. Reporting both is essential for sound interpretation.

Question 31Scientific Inquiry and Research Design

A researcher plans to compare mean pain scores across four drug regimens and also examine whether the effect differs by sex. Which analysis strategy is most appropriate for testing the interaction?

  1. A paired t-test of pre-treatment and post-treatment scores
  2. A chi-square test of independence on pain categories
  3. Four separate t-tests comparing each regimen with control
  4. A two-way analysis of variance with regimen and sex as factors Correct answer.

Explanation

An interaction asks whether the effect of one factor depends on the level of another, which a two-way analysis of variance tests directly through the interaction term. This also preserves the overall error rate better than multiple separate tests. A factorial structure matches the two factors in the design.

Question 32Scientific Inquiry and Research Design

A study measures self-reported physical activity and simultaneously measures cardiorespiratory fitness at one time point. The authors conclude that activity improves fitness. Which design limitation most undermines this conclusion?

  1. The study lacks a control group and therefore cannot compute associations
  2. Self-report always produces random rather than systematic error
  3. Fitness is a continuous variable and cannot be analyzed causally
  4. Cross-sectional data cannot establish temporal order between exposure and outcome Correct answer.

Explanation

When exposure and outcome are measured at the same time, the data cannot show which came first, which is a prerequisite for causal inference. Temporality is one of the core criteria for causal reasoning. A longitudinal design would be needed to support the claim.

Question 33Scientific Inquiry and Research Design

A researcher wants to study rare adverse events following a new vaccine. Which design is best suited for this purpose?

  1. A small randomized trial with 40 volunteers
  2. A qualitative interview study of affected patients
  3. A case-control study nested in a large vaccinated cohort Correct answer.
  4. A cross-sectional survey of current symptoms

Explanation

Rare outcomes require designs that efficiently sample many exposed persons and compare those with and without the event. A case-control study nested in a cohort does exactly this and can link vaccination records to outcomes. This design also allows estimation of odds ratios for the rare event.

Question 34Scientific Inquiry and Research Design

A survey is mailed to a random sample of 1,000 households, but only 25% respond, and responders tend to be older and more health-conscious than non-responders. Which threat to inference is most prominent?

  1. Nonresponse bias affecting representativeness Correct answer.
  2. Ceiling effects in the outcome scale
  3. Measurement error from the survey instrument
  4. Confounding by the randomization procedure

Explanation

When responders differ systematically from non-responders on characteristics related to the outcome, the sample no longer represents the target population. This is nonresponse bias, a form of selection problem. A high response rate alone does not guarantee representativeness, but differential response is the key concern here.

Question 35Scientific Inquiry and Research Design

A researcher plans a study and calculates that 120 participants per group are needed to detect a clinically meaningful difference with 80% power at alpha 0.05. After enrolling 90 per group, the observed difference is not significant. Which conclusion is most defensible?

  1. The treatment is proven ineffective because the result was not significant
  2. The study is underpowered, so a meaningful effect cannot be ruled out Correct answer.
  3. The alpha level should be raised to 0.20 to obtain significance
  4. The effect size must be smaller than any clinically relevant threshold

Explanation

Failing to reach the planned sample size reduces power, so a non-significant result may reflect insufficient precision rather than absence of effect. Absence of evidence is not evidence of absence. The correct interpretation acknowledges the study's limited ability to detect the target difference.

Question 36Scientific Inquiry and Research Design

A researcher wants to estimate the prevalence of a chronic condition in a city and needs results that generalize to the whole adult population. Which sampling approach best supports this goal?

  1. Snowball sampling starting from patient support groups
  2. Quota sampling by age and sex at a shopping center
  3. Probability sampling from a complete population register Correct answer.
  4. Convenience sampling at a single clinic

Explanation

Probability sampling gives every member of the target population a known chance of selection, which supports unbiased prevalence estimation and generalizability. A complete register provides the sampling frame needed for such selection. Non-probability methods cannot guarantee representativeness.

Question 37Scientific Inquiry and Research Design

A trial reports that a drug reduces the risk of a composite endpoint from 10% to 7%. Which statement correctly expresses the absolute effect?

  1. The absolute risk reduction is 3 percentage points Correct answer.
  2. The number needed to treat is 3
  3. The odds ratio equals 0.03
  4. The relative risk reduction is 3%

Explanation

Absolute risk reduction is the arithmetic difference between control and treatment event rates, here 10% minus 7%, or 3 percentage points. This differs from the relative reduction, which is 30%. Absolute measures convey the clinical magnitude of benefit more directly.

Question 38Scientific Inquiry and Research Design

A team conducts a study in which neither participants nor researchers know group assignment, and the analysis plan is registered before data collection. A reviewer nonetheless asks whether the finding applies to routine care. Which concept is the reviewer questioning?

  1. External validity Correct answer.
  2. Statistical conclusion validity
  3. Internal validity
  4. Construct validity

Explanation

External validity concerns whether results generalize to other populations, settings, and conditions such as routine care. Internal validity concerns whether the observed effect is causal within the study. The reviewer is asking about applicability beyond the trial context.

Question 39Scientific Inquiry and Research Design

A systematic review includes 12 trials, and a funnel plot shows that small studies with null results are conspicuously absent while small studies with large effects are overrepresented. Which interpretation is most appropriate?

  1. The large studies are necessarily more biased than small ones
  2. The pooled estimate may be inflated by publication bias Correct answer.
  3. The heterogeneity is too low to permit pooling
  4. The funnel plot asymmetry confirms a causal mechanism

Explanation

Asymmetric funnel plots often signal that small null or negative studies were not published, which can inflate pooled effects. This is a recognized concern in meta-analysis. The pattern suggests selective availability of results rather than a biological mechanism.

Question 40Scientific Inquiry and Research Design

A researcher wants to reduce the risk that a single analyst's expectations influence how ambiguous qualitative data are coded. Which strategy is most appropriate?

  1. Reporting only the codes that support the main theme
  2. Converting qualitative data into numerical scores early
  3. Having a second coder independently code a subset and comparing results Correct answer.
  4. Increasing the sample size to several hundred interviews

Explanation

Independent double coding lets researchers check whether interpretations are consistent across analysts, which reduces the influence of a single person's expectations. Discrepancies are then discussed and resolved. This is a standard credibility strategy in qualitative research.

Question 41Evidence Appraisal and Critical Reading of the Literature

A trial reports a hazard ratio of 0.80 with a 95% confidence interval of 0.65 to 0.98 for the primary outcome. Which interpretation of the result is most defensible?

  1. The result proves the intervention has no clinically meaningful effect.
  2. The result is statistically significant at the 0.05 level and favors the intervention. Correct answer.
  3. The result is not statistically significant because the interval is wide.
  4. The point estimate indicates a 20% absolute risk reduction.

Explanation

Because the 95% confidence interval excludes 1.0, the effect is statistically significant at the two-sided 0.05 level, and the point estimate below 1.0 favors the intervention. The width of the interval reflects precision, not whether significance is achieved. A hazard ratio is a relative measure, so it cannot be read as an absolute risk reduction.

Question 42Evidence Appraisal and Critical Reading of the Literature

In a randomized controlled trial, what does allocation concealment primarily protect against?

  1. Reporting bias from selective publication of outcomes
  2. Selection bias arising from foreknowledge of upcoming treatment assignments Correct answer.
  3. Attrition bias caused by differential loss to follow-up
  4. Detection bias from unblinded outcome assessors

Explanation

Allocation concealment prevents investigators from anticipating or influencing which group the next participant is assigned to, thereby preventing selection bias at enrollment. It is distinct from blinding, which addresses bias after randomization. Loss to follow-up, assessor blinding, and publication practices concern other bias domains.

Question 43Evidence Appraisal and Critical Reading of the Literature

A systematic review pools studies of varying designs and reports substantial statistical heterogeneity (I-squared = 78%). Which action is most appropriate when interpreting the pooled estimate?

  1. Ignore heterogeneity because the pooled estimate remains statistically significant
  2. Present the pooled estimate cautiously and explore sources of heterogeneity Correct answer.
  3. Report only the single largest study as the definitive result
  4. Discard all included studies and conclude no evidence exists

Explanation

High I-squared indicates that a large share of variability across studies is due to true differences rather than chance, so the pooled estimate should be interpreted cautiously with exploration of clinical and methodological sources. Statistical significance of the pooled effect does not resolve inconsistency. Discarding all evidence or relying on one study is unjustified.

Question 44Evidence Appraisal and Critical Reading of the Literature

Which feature most strongly supports a causal interpretation of an observed association in an observational study?

  1. Publication of the findings in a high-impact journal
  2. Use of a validated measurement instrument
  3. A large sample size with a narrow confidence interval
  4. A dose-response relationship between exposure and outcome Correct answer.

Explanation

A dose-response gradient is one of the classic criteria supporting causal inference because increasing exposure producing increasing effect is unlikely to arise from confounding alone. Large samples improve precision but not validity against confounding. Validated instruments and journal prestige do not establish causality.

Question 45Evidence Appraisal and Critical Reading of the Literature

A study reports a number needed to treat of 50. Which statement correctly describes this value?

  1. One in fifty treated patients will experience an adverse event
  2. Fifty patients must be treated for one additional patient to benefit Correct answer.
  3. The treatment reduces relative risk by 50%
  4. Fifty percent of treated patients will experience the benefit

Explanation

The number needed to treat is the reciprocal of the absolute risk reduction, representing how many patients must receive the intervention for one additional favorable outcome. It is an absolute measure and does not describe relative risk reduction or harm.

Question 46Evidence Appraisal and Critical Reading of the Literature

A diagnostic accuracy study reports a sensitivity of 90% and a specificity of 70% for a screening test in a population with 2% prevalence. What is the most important consequence for interpreting a positive result?

  1. The positive predictive value will be low despite high sensitivity Correct answer.
  2. Sensitivity will fall when the test is used in a higher-prevalence population
  3. Specificity will rise because prevalence changes the cutoff
  4. The negative predictive value will be low because specificity is imperfect

Explanation

When prevalence is very low, even a highly sensitive test yields many false positives, so the positive predictive value is low. Sensitivity and specificity are properties of the test and are generally stable across prevalence, whereas predictive values depend on prevalence. A specificity of 70% guarantees substantial false positives in a low-prevalence setting.

Question 47Evidence Appraisal and Critical Reading of the Literature

Which design is most vulnerable to recall bias when studying an exposure-outcome association?

  1. A cross-sectional survey using a validated screening tool
  2. A case-control study relying on participant-reported past exposures Correct answer.
  3. A prospective cohort with standardized exposure measurement
  4. A randomized double-blind trial with objective outcome measurement

Explanation

Recall bias arises when cases and controls remember or report past exposures differently, which is a particular threat in retrospective case-control studies using self-report. Prospective designs measure exposure before outcome and reduce this threat. The other options either measure exposure prospectively or do not rely on differential recall.

Question 48Evidence Appraisal and Critical Reading of the Literature

A trial reports an intention-to-treat analysis. What is the principal reason for this approach?

  1. It guarantees that confounding by indication is removed
  2. It maximizes the observed treatment effect by excluding nonadherent participants
  3. It preserves the benefits of randomization by analyzing groups as assigned Correct answer.
  4. It eliminates the need for a per-protocol analysis

Explanation

Intention-to-treat analysis compares participants according to randomized assignment, preserving the comparability created by randomization and avoiding bias from post-randomization selection. It typically gives a conservative estimate rather than maximizing the effect. It does not remove confounding by indication, which randomization addresses at baseline.

Question 49Evidence Appraisal and Critical Reading of the Literature

In appraising a meta-analysis, which practice most directly reduces the risk of selective reporting bias?

  1. Pooling only studies published in the last two years
  2. Registering the review protocol and searching trial registries for unpublished studies Correct answer.
  3. Using a fixed-effects model instead of a random-effects model
  4. Restricting inclusion to English-language publications

Explanation

Prospective protocol registration and searching registries or grey literature help identify unpublished or suppressed studies, directly mitigating selective reporting and publication bias. Language restriction and recency filters can worsen bias. The choice of pooling model affects weighting, not reporting bias.

Question 50Evidence Appraisal and Critical Reading of the Literature

A forest plot shows a study with a wide confidence interval crossing the line of no effect, while the overall diamond does not cross it. What is the best interpretation?

  1. The individual study contradicts the pooled estimate and should be removed
  2. The individual study is inconclusive but contributes to a significant pooled effect Correct answer.
  3. The overall pooled result is invalid because one study is imprecise
  4. The wide interval indicates the study contains a data error

Explanation

A single study whose interval crosses the null is inconclusive on its own but still contributes information to the pooled estimate, which can be significant when combined with others. Imprecision is not grounds for removal or evidence of error. Removing studies solely because they cross the null would bias the synthesis.