Statistics Guide for Research Grant Applicants

Checklist

This checklist can be downloaded as an Adobe Acrobat PDF file, for ease of printing.

Tick [ ] those that apply

Design

• Is the study observational? [   ]
  (see A-1.1 and A-1.2)
• If Yes
  • Is it a cohort study? [   ]
    (see A-1.3)
  • Is it a case-control study? [   ]
    (see A-1.4)
  • Is it a cross-sectional study? [   ] (see A-1.5)
  • If Yes to cross-sectional study
    • Is it a prevalence study? [   ]
      (see A-1.5a)
    • Is it estimating sensitivity and specificity? [   ]
      (see A-1.5b and A-1.5c)
    • Is it an ecological study? [   ]
      (see A-1.5d)
  • Have you addressed the issue of confounding in your proposal? [   ]
    (see A-1.6)
• Is the study experimental? [   ]
  (see A-1.1, A-1.2)
• If Yes
  • Is the study a type of trial e.g. a clinical trial? [   ]
    (see A-1.7, A-1.8)
  • If Yes
    • Is the study a controlled trial (i.e. is there a control group)? [   ]
      (see A-1.8 and B-3)
    • Is it a randomised trial (i.e. are study subjects randomly allocated to groups)? [   ]
      (see A-1.8 and B-5)
    • If Yes to randomised trial
      • Is it important to have similar numbers in each randomisation group? (you may need to use blocks) [   ]
        (see B-5.6)
      • Are any known factors strongly prognostic? [   ]
        (you may need to randomise in strata)
        (see B-5.7)
      • If Yes to strongly prognostic factors
        • Is the proposed sample size small? [   ]
          (see B-5.8)
      • Are groups of individuals to be randomised together to the same treatment? [   ]
        (see B-5.9)
      • Is this a cross over trial? [   ]
        (see B-5.10b)
      • Is the assessor blind? [   ]
        (see A-1.8 and B-4)
      • Are the study subjects blind? [   ]
        (see A-1.8 and B-4)
• Is the study prospective? [   ]
  (see A-2)
• If Yes
  • Have you specified the length of follow up? [   ]
    (see A-2)

The Study Subjects

• Have you described where they come from? [   ]
• Have you explained why they are an appropriate group? [   ]
• Have you described how the study subjects will be selected? [   ]
• Have you specified inclusion / exclusion criteria? [   ]
• Have you specified your proposed sample size taking into account refusals/drop-outs? [   ]

Types of Variables

• Have you described all outcome and explanatory variables in terms of data type and scale of measurement? [   ]
  (see A-4.1 and A-4.2)
• Have you described how the data will be collected? [   ]
  (see A-4.3)
• If using a questionnaire or a non-standard measurement, have you provided information on its reliability and validity? [   ]
  (see A-4.4, A-4.4a, A-4.4b, A-4.4c)

Sample Size

• Have you provided a sample size calculation? [   ]
  (see D-1)
• Have you defined the outcome variable(s) used in the sample size calculation? [   ]
  (see D-5)
• Have you defined the effect size which would be of clinical importance? [   ]
  (see D-4.5)
• Have you described the power and significance level of the sample size calculation? [   ]
  (see D-4.3 and D-4.4)
• Has your sample size made allowance for expected response rates and other sample attrition? [   ]
  (see D-6)
• Is your sample size consistent with the study aims? [   ]
  (see D-7)
• Is your sample size consistent with the proposed analysis of the study? [   ]
  (see D-7)
• Is your description of the sample size calculation adequate? [   ]
  (See examples in D-8)

Statistical Analysis

• Have you described the proposed statistical methods using appropriate terminology? [   ]
  (see E-1.1, E-1.2)
• Are the proposed methods appropriate for the types of data generated by your study? [   ]
  (see E-2, E-2.1, E-11)
• Will the assumptions made by the proposed methods hold? [   ]
  (see E-4, E-4.1)
• Do the proposed methods take account of the structure of the data set (structure such as hierarchy, clustering, matching, paired data)? [   ]
  (see E-3, E-6, E-6.2, E-10)
• Have important confounding factors been listed and methods of adjusting for them presented? [   ]
  (see E-5)
• Will the proposed methods take account of multiple testing where appropriate? [   ]
  (see E-7.1, E-7.2, E-7.3, E-7.4, E-7.4a, E-7.4b, E-7.4c, E-7.4d, E-7.4e, E-7.4f)
• Have biases due to measurement error been considered e.g. regression towards the mean? [   ]
  (see E-8)
• Have details on the calculation of confidence intervals been provided? [   ]
  (see E-12)

• Have you specified that your analysis will be by intention to treat? [   ]
  (see E-9)

Back to Brief Table of Contents.

Back to Martin Bland's home page.

This page is maintained by Martin Bland.

Last updated: 10 September, 2009.

Back to top