Statistical Questions

Philip Sedgwick, a medical statisitican at St Georges, has written many short Statistical question articles in the BMJ, These take to form of a summary of the statistical methodology in pubished paper, follow by a question about which of four statements are correct or incorrect. This is then followed by an explanation of the correct and incorrect answers. A list of these articles is given below, many of them are relevant for undergraduate students.

Absence of evidence is not evidence of absencehttp://www.bmj.com/content/342/bmj.d3126
Absolute and relative riskshttp://www.bmj.com/content/345/bmj.e5613
Allocation concealmenthttp://www.bmj.com/content/340/bmj.c537.full
Allocation concealmenthttp://www.bmj.com/content/344/bmj.e156
Analysing case-control studies: adjusting for confoundinghttp://www.bmj.com/content/346/bmj.f25
Analysis by intention to treathttp://www.bmj.com/content/342/bmj.d2212
Analysis by per protocolhttp://www.bmj.com/content/342/bmj.d2330
Analysis of longitudinal studieshttp://www.bmj.com/content/346/bmj.f363
Analysis of outcome measures within treatment groupshttp://www.bmj.com/content/345/bmj.e7201
Bias in clinical trialshttp://www.bmj.com/content/343/bmj.d4176
Block randomisationhttp://www.bmj.com/content/343/bmj.d7139
Case-control studies: measures of riskhttp://www.bmj.com/content/346/bmj.f1185
Cluster randomised controlled trialshttp://www.bmj.com/content/345/bmj.e4654
Cluster randomised controlled trials: sample size calculationshttp://www.bmj.com/content/346/bmj.f2839
Cohen's coefficient Kappahttp://www.bmj.com/content/344/bmj.e1178
Cohort studieshttp://www.bmj.com/content/340/bmj.c1002.full
Cohort studies: sources of biashttp://www.bmj.com/content/343/bmj.d7839
Confidence intervals and statistical significancehttp://www.bmj.com/content/344/bmj.e2238
Confidence intervals and statistical significance: rules of thumbhttp://www.bmj.com/content/345/bmj.e4960
Confidence intervals: predicting uncertaintyhttp://www.bmj.com/content/344/bmj.e3147
Confounding in case-control studieshttp://www.bmj.com/content/341/bmj.c5136.full
Confounding in case-control studies IIhttp://www.bmj.com/content/341/bmj.c5290.full
Confounding in clinical trialshttp://www.bmj.com/content/345/bmj.e7951
Confounding in randomised controlled trialshttp://www.bmj.com/content/341/bmj.c5403.full
Control groupshttp://www.bmj.com/content/340/bmj.c391.full
Control treatmentshttp://www.bmj.com/content/342/bmj.d2109
Correlationhttp://www.bmj.com/content/345/bmj.e5407
Correlation versus linear regressionhttp://www.bmj.com/content/346/bmj.f2686
Cox proportional hazards regressionhttp://www.bmj.com/content/347/bmj.f4919
Cross sectional studieshttp://www.bmj.com/content/340/bmj.c846.full
Crossover trialshttp://www.bmj.com/content/344/bmj.e3710
Derivation of hazard ratioshttp://www.bmj.com/content/343/bmj.d6994
Describing the spread of data Ihttp://www.bmj.com/content/340/bmj.c1116.full
Describing the spread of data IIhttp://www.bmj.com/content/340/bmj.c1282.full
Double dummy trialshttp://www.bmj.com/content/343/bmj.d7294
Double dummy trials incorporating factorial designshttp://www.bmj.com/content/346/bmj.f3539
The ecological fallacyhttp://www.bmj.com/content/343/bmj.d4670
Effect sizeshttp://www.bmj.com/content/345/bmj.e7370
Equivalence trialshttp://www.bmj.com/content/346/bmj.f184
Errors when statistical hypothesis testinghttp://www.bmj.com/content/340/bmj.c2348.full
Estimating the population at riskhttp://www.bmj.com/content/345/bmj.e6859
External and internal validity in clinical trialshttp://www.bmj.com/content/344/bmj.e1004
Factorial trialshttp://www.bmj.com/content/340/bmj.c3401.full
Generalisation and extrapolationhttp://www.bmj.com/content/340/bmj.c717.full
Generalisation and extrapolation of study resultshttp://www.bmj.com/content/346/bmj.f3022
Hazard ratioshttp://www.bmj.com/content/341/bmj.c4566.full
Hazards and hazard ratioshttp://www.bmj.com/content/345/bmj.e5980
The healthy entrant effecthttp://www.bmj.com/content/344/bmj.e2728
How to read a forest plothttp://www.bmj.com/content/345/bmj.e8335
Incidence rate ratiohttp://www.bmj.com/content/341/bmj.c4804.full
Incidence rateshttp://www.bmj.com/content/344/bmj.e1589
Independent samples t testhttp://www.bmj.com/content/340/bmj.c2673.full
Internal and external validityhttp://www.bmj.com/content/340/bmj.c1705.full
Intraclass correlation coefficienthttp://www.bmj.com/content/346/bmj.f1816
Kaplan-Meier survival analysis: types of censored observationshttp://www.bmj.com/content/347/bmj.f4663
Limits of agreement (Bland-Altman method)http://www.bmj.com/content/346/bmj.f1630
The log rank testhttp://www.bmj.com/content/341/bmj.c3773.full
Log transformation of datahttp://www.bmj.com/content/345/bmj.e6727
Logistic regressionhttp://www.bmj.com/content/347/bmj.f4488
Measurement of datahttp://www.bmj.com/content/341/bmj.c5929.full
Meta-analyses Ihttp://www.bmj.com/content/342/bmj.d45.full
Meta-analyses IIhttp://www.bmj.com/content/341/bmj.d229.full
Meta-analyses IIIhttp://www.bmj.com/content/342/bmj.d244.full
Meta-analyses IVhttp://www.bmj.com/content/342/bmj.d540.full
Meta-analyses Vhttp://www.bmj.com/content/342/bmj.d686.full
Meta-analyses VIhttp://www.bmj.com/content/342/bmj.d937.full
Meta-analyses VIIhttp://www.bmj.com/content/342/bmj.d1108.full
Meta-analyses: heterogeneity and subgroup analysishttp://www.bmj.com/content/346/bmj.f4040
Meta-analyses: how to read a funnel plothttp://www.bmj.com/content/346/bmj.f1342
Meta-analyses: tests of heterogeneityhttp://www.bmj.com/content/344/bmj.e3971
Multiple regressionhttp://www.bmj.com/content/347/bmj.f4373
Multiple significance testshttp://www.bmj.com/content/340/bmj.c2963.full
Multiple significance tests: the Bonferroni correctionhttp://www.bmj.com/content/344/bmj.e691
n of 1 trialshttp://www.bmj.com/content/344/bmj.e844
Nested case-control studieshttp://www.bmj.com/content/340/bmj.c2582.full
Non-inferiority trialshttp://www.bmj.com/content/342/bmj.d3253
Non-parametric statistical tests for independent groups: numerical datahttp://www.bmj.com/content/344/bmj.e3354
Non-parametric statistical tests for two related groups: numerical datahttp://www.bmj.com/content/344/bmj.e2537
The Normal distributionhttp://www.bmj.com/content/341/bmj.c6085.full
The normal distributionhttp://www.bmj.com/content/345/bmj.e6533
Normal rangeshttp://www.bmj.com/content/346/bmj.f1343
Number needed to harmhttp://www.bmj.com/content/342/bmj.d2811
Number needed to treathttp://www.bmj.com/content/341/bmj.c5614.full
Number needed to treat Ihttp://www.bmj.com/content/342/bmj.d2463
Number needed to treat IIhttp://www.bmj.com/content/342/bmj.d2664
Numbers needed to treat and harmhttp://www.bmj.com/content/341/bmj.c5731.full
Observational study designhttp://www.bmj.com/content/345/bmj.e5856
Observational Study Design Ihttp://www.bmj.com/content/342/bmj.d1742.full
Observational study design IIhttp://www.bmj.com/content/342/bmj.d1903
Observational study designshttp://www.bmj.com/content/344/bmj.e4209
Odds and odds ratioshttp://www.bmj.com/content/347/bmj.f5067
Odds ratioshttp://www.bmj.com/content/341/bmj.c4414.full
Odds ratios and adjusting for confoundinghttp://www.bmj.com/content/343/bmj.d6833
Odds ratios IIhttp://www.bmj.com/content/341/bmj.c4971.full
One sided and two sided hypothesis testshttp://www.bmj.com/content/340/bmj.c2458.full
One way analysis of variancehttp://www.bmj.com/content/344/bmj.e2427
Open clinical trialshttp://www.bmj.com/content/345/bmj.e4749
Open label crossover trialshttp://www.bmj.com/content/346/bmj.f4186
Open label trialshttp://www.bmj.com/content/342/bmj.d3427
P valueshttp://www.bmj.com/content/340/bmj.c2203.full
P values or confidence intervals?http://www.bmj.com/content/346/bmj.f3212
Parametric statistical tests for independent groups: numerical datahttp://www.bmj.com/content/345/bmj.e8145
Parametric v non-parametric statistical testshttp://www.bmj.com/content/344/bmj.e1753
Patient preference trialshttp://www.bmj.com/content/342/bmj.d3602
Pearson's correlation coefficienthttp://www.bmj.com/content/345/bmj.e4483
Per protocol analysishttp://www.bmj.com/content/340/bmj.c1825.full
The placebo effecthttp://www.bmj.com/content/343/bmj.d7665
Placebo run in periodshttp://www.bmj.com/content/344/bmj.e3827
Placeboshttp://www.bmj.com/content/340/bmj.c247.full
Populations and sampleshttp://www.bmj.com/content/344/bmj.e3048
Prevalence and incidencehttp://www.bmj.com/content/341/bmj.c4709.full
Primary and secondary outcome measureshttp://www.bmj.com/content/340/bmj.c1938.full
Proportional quota samplinghttp://www.bmj.com/content/345/bmj.e6336
The purpose of control groupshttp://www.bmj.com/content/346/bmj.f658
Questionnaire surveyshttp://www.bmj.com/content/343/bmj.d4449
Questionnaire surveys: sources of biashttp://www.bmj.com/content/347/bmj.f5265
Random allocation IIIhttp://www.bmj.com/content/339/bmj.b5072.full
Random sampling versus random allocationhttp://www.bmj.com/content/343/bmj.d7453
Randomised controlled trials with full factorial designshttp://www.bmj.com/content/345/bmj.e5114
Receiver operating characteristic curveshttp://www.bmj.com/content/343/bmj.d4302
Receiver operating characteristic curveshttp://www.bmj.com/content/346/bmj.f2493
Reference and normal rangeshttp://www.bmj.com/content/341/bmj.c6666.full
Relative riskshttp://www.bmj.com/content/341/bmj.c3983.full
Relative risks and confidence intervalshttp://www.bmj.com/content/341/bmj.c4166.full
Relative risks and statistical significancehttp://www.bmj.com/content/341/bmj.c4541.full
Restricted randomisationhttp://www.bmj.com/content/344/bmj.e1324
Sample size calculations Ihttp://www.bmj.com/content/340/bmj.c3104.full
Sample size calculations IIhttp://www.bmj.com/content/340/bmj.c3299.full
Sample size: how many participants are needed in a trial?http://www.bmj.com/content/346/bmj.f1041
Sampling Ihttp://www.bmj.com/content/339/bmj.b5512.full
Sampling IIhttp://www.bmj.com/content/340/bmj.b5677.full
Sampling IIIhttp://www.bmj.com/content/340/bmj.c93.full
Sampling methods Ihttp://www.bmj.com/content/342/bmj.d1249.full
Sampling methods IIhttp://www.bmj.com/content/342/bmj.d1387.full
Sampling methods IIIhttp://www.bmj.com/content/342/bmj.d1537.full
Screening tests: indices of performancehttp://www.bmj.com/content/342/bmj.d3763
Screening tests: likelihood ratioshttp://www.bmj.com/content/342/bmj.d3986
Selection bias versus allocation biashttp://www.bmj.com/content/346/bmj.f3345
Sequential trialshttp://www.bmj.com/content/344/bmj.e2102
Simple linear regressionhttp://www.bmj.com/content/346/bmj.f2340
Skewed distributionshttp://www.bmj.com/content/341/bmj.c6276.full
Skewed distributionshttp://www.bmj.com/content/345/bmj.e7534
Skewed distributions IIhttp://www.bmj.com/content/341/bmj.c6448.full
Sources of bias in randomised controlled trialshttp://www.bmj.com/content/341/bmj.c6995.full
Sources of bias in randomised controlled trials IIhttp://www.bmj.com/content/341/bmj.c7053.full
Standard deviation versus standard errorhttp://www.bmj.com/content/343/bmj.d8010
Standard error of the meanhttp://www.bmj.com/content/340/bmj.c1437.full
Standardisation of outcome measures (z scores)http://www.bmj.com/content/345/bmj.e6178
Statistical hypothesis testinghttp://www.bmj.com/content/340/bmj.c2059.full
Statistical tests for independent groups: categorical datahttp://www.bmj.com/content/344/bmj.e344
Statistical tests for independent groups: time to event datahttp://www.bmj.com/content/345/bmj.e5257
Statistical tests: matched pairs categorical datahttp://www.bmj.com/content/343/bmj.d6623
Stratified random allocationhttp://www.bmj.com/content/346/bmj.f822
Study designhttp://www.bmj.com/content/339/bmj.b5263.full
Study designhttp://www.bmj.com/content/340/bmj.c2667.full
Superiority trialshttp://www.bmj.com/content/342/bmj.d2981
Survival (time to event) data Ihttp://www.bmj.com/content/341/bmj.c3537.full
Survival (time to event) data IIhttp://www.bmj.com/content/341/bmj.c3665.full
Survival (time to event) data: censored observationshttp://www.bmj.com/content/343/bmj.d4816
Survival (time to event) data: median survival timeshttp://www.bmj.com/content/343/bmj.d4890
T scores and z scoreshttp://www.bmj.com/content/341/bmj.c7362.full
Units of analysishttp://www.bmj.com/content/346/bmj.f2128
Variables and parametershttp://www.bmj.com/content/340/bmj.c1568.full
Variables sample estimates and population parametershttp://www.bmj.com/content/346/bmj.f2019
What are odds?http://www.bmj.com/content/344/bmj.e2853
What is a P value?http://www.bmj.com/content/345/bmj.e7767
What is a superiority trial?http://www.bmj.com/content/347/bmj.f5420
What is intention to treat analysis?http://www.bmj.com/content/346/bmj.f3662
What is number needed to harm (NNH)?http://www.bmj.com/content/347/bmj.f4869
What is number needed to treat (NNT)?http://www.bmj.com/content/347/bmj.f4605
What is per protocol analysis?http://www.bmj.com/content/346/bmj.f3748
What is recall bias?http://www.bmj.com/content/344/bmj.e3519
What Is Risk?http://www.bmj.com/content/345/bmj.e7036
What is sampling error?http://www.bmj.com/content/344/bmj.e4285
What is the standard error of the mean?http://www.bmj.com/content/346/bmj.f532
Why match in case-control studies?http://www.bmj.com/content/344/bmj.e691
Why randomise in clinical trials?http://www.bmj.com/content/345/bmj.e5584
z scoreshttp://www.bmj.com/content/341/bmj.c6746.full
Zelen's designhttp://www.bmj.com/content/345/bmj.e8505

List updated on 29 October 2013

Please Report any issues to Dr Trevor Bryant