2.0.statistical methods and determination of sample sizesalummkata1
statistical methods and determination of sample size
These guidelines focus on the validation of the bioanalytical methods generating quantitative concentration data used for pharmacokinetic and toxicokinetic parameter determinations.
Statistical Significance Testing in Information Retrieval: An Empirical Analy...Julián Urbano
Statistical significance testing is widely accepted as a means to assess how well a difference in effectiveness reflects an actual difference between systems, as opposed to random noise because of the selection of topics. According to recent surveys on SIGIR, CIKM, ECIR and TOIS papers, the t-test is the most popular choice among IR researchers. However, previous work has suggested computer intensive tests like the bootstrap or the permutation test, based mainly on theoretical arguments. On empirical grounds, others have suggested non-parametric alternatives such as the Wilcoxon test. Indeed, the question of which tests we should use has accompanied IR and related fields for decades now. Previous theoretical studies on this matter were limited in that we know that test assumptions are not met in IR experiments, and empirical studies were limited in that we do not have the necessary control over the null hypotheses to compute actual Type I and Type II error rates under realistic conditions. Therefore, not only is it unclear which test to use, but also how much trust we should put in them. In contrast to past studies, in this paper we employ a recent simulation methodology from TREC data to go around these limitations. Our study comprises over 500 million p-values computed for a range of tests, systems, effectiveness measures, topic set sizes and effect sizes, and for both the 2-tail and 1-tail cases. Having such a large supply of IR evaluation data with full knowledge of the null hypotheses, we are finally in a position to evaluate how well statistical significance tests really behave with IR data, and make sound recommendations for practitioners.
2.0.statistical methods and determination of sample sizesalummkata1
statistical methods and determination of sample size
These guidelines focus on the validation of the bioanalytical methods generating quantitative concentration data used for pharmacokinetic and toxicokinetic parameter determinations.
Statistical Significance Testing in Information Retrieval: An Empirical Analy...Julián Urbano
Statistical significance testing is widely accepted as a means to assess how well a difference in effectiveness reflects an actual difference between systems, as opposed to random noise because of the selection of topics. According to recent surveys on SIGIR, CIKM, ECIR and TOIS papers, the t-test is the most popular choice among IR researchers. However, previous work has suggested computer intensive tests like the bootstrap or the permutation test, based mainly on theoretical arguments. On empirical grounds, others have suggested non-parametric alternatives such as the Wilcoxon test. Indeed, the question of which tests we should use has accompanied IR and related fields for decades now. Previous theoretical studies on this matter were limited in that we know that test assumptions are not met in IR experiments, and empirical studies were limited in that we do not have the necessary control over the null hypotheses to compute actual Type I and Type II error rates under realistic conditions. Therefore, not only is it unclear which test to use, but also how much trust we should put in them. In contrast to past studies, in this paper we employ a recent simulation methodology from TREC data to go around these limitations. Our study comprises over 500 million p-values computed for a range of tests, systems, effectiveness measures, topic set sizes and effect sizes, and for both the 2-tail and 1-tail cases. Having such a large supply of IR evaluation data with full knowledge of the null hypotheses, we are finally in a position to evaluate how well statistical significance tests really behave with IR data, and make sound recommendations for practitioners.
Marketing Research Project on T test and Sample Designing, Detail Analysis of all the aspect of T test and usage of all the tools for finding out the different variants.
This ppt includes basic concepts about data types, levels of measurements. It also explains which descriptive measure, graph and tests should be used for different types of data. A brief of Pivot tables and charts is also included.
How to Structure the “Approach” Section of a Grant Application by David Elash...UCLA CTSI
David Elashoff, PhD speaks on the topic of "How to Structure the “Approach” Section of a Grant Application" at the November 08, 2018 R Award Workshop at UCLA.
Read| The latest issue of The Challenger is here! We are thrilled to announce that our school paper has qualified for the NATIONAL SCHOOLS PRESS CONFERENCE (NSPC) 2024. Thank you for your unwavering support and trust. Dive into the stories that made us stand out!
Delivering Micro-Credentials in Technical and Vocational Education and TrainingAG2 Design
Explore how micro-credentials are transforming Technical and Vocational Education and Training (TVET) with this comprehensive slide deck. Discover what micro-credentials are, their importance in TVET, the advantages they offer, and the insights from industry experts. Additionally, learn about the top software applications available for creating and managing micro-credentials. This presentation also includes valuable resources and a discussion on the future of these specialised certifications.
For more detailed information on delivering micro-credentials in TVET, visit this https://tvettrainer.com/delivering-micro-credentials-in-tvet/
Normal Labour/ Stages of Labour/ Mechanism of LabourWasim Ak
Normal labor is also termed spontaneous labor, defined as the natural physiological process through which the fetus, placenta, and membranes are expelled from the uterus through the birth canal at term (37 to 42 weeks
it describes the bony anatomy including the femoral head , acetabulum, labrum . also discusses the capsule , ligaments . muscle that act on the hip joint and the range of motion are outlined. factors affecting hip joint stability and weight transmission through the joint are summarized.
A workshop hosted by the South African Journal of Science aimed at postgraduate students and early career researchers with little or no experience in writing and publishing journal articles.
Biological screening of herbal drugs: Introduction and Need for
Phyto-Pharmacological Screening, New Strategies for evaluating
Natural Products, In vitro evaluation techniques for Antioxidants, Antimicrobial and Anticancer drugs. In vivo evaluation techniques
for Anti-inflammatory, Antiulcer, Anticancer, Wound healing, Antidiabetic, Hepatoprotective, Cardio protective, Diuretics and
Antifertility, Toxicity studies as per OECD guidelines
June 3, 2024 Anti-Semitism Letter Sent to MIT President Kornbluth and MIT Cor...Levi Shapiro
Letter from the Congress of the United States regarding Anti-Semitism sent June 3rd to MIT President Sally Kornbluth, MIT Corp Chair, Mark Gorenberg
Dear Dr. Kornbluth and Mr. Gorenberg,
The US House of Representatives is deeply concerned by ongoing and pervasive acts of antisemitic
harassment and intimidation at the Massachusetts Institute of Technology (MIT). Failing to act decisively to ensure a safe learning environment for all students would be a grave dereliction of your responsibilities as President of MIT and Chair of the MIT Corporation.
This Congress will not stand idly by and allow an environment hostile to Jewish students to persist. The House believes that your institution is in violation of Title VI of the Civil Rights Act, and the inability or
unwillingness to rectify this violation through action requires accountability.
Postsecondary education is a unique opportunity for students to learn and have their ideas and beliefs challenged. However, universities receiving hundreds of millions of federal funds annually have denied
students that opportunity and have been hijacked to become venues for the promotion of terrorism, antisemitic harassment and intimidation, unlawful encampments, and in some cases, assaults and riots.
The House of Representatives will not countenance the use of federal funds to indoctrinate students into hateful, antisemitic, anti-American supporters of terrorism. Investigations into campus antisemitism by the Committee on Education and the Workforce and the Committee on Ways and Means have been expanded into a Congress-wide probe across all relevant jurisdictions to address this national crisis. The undersigned Committees will conduct oversight into the use of federal funds at MIT and its learning environment under authorities granted to each Committee.
• The Committee on Education and the Workforce has been investigating your institution since December 7, 2023. The Committee has broad jurisdiction over postsecondary education, including its compliance with Title VI of the Civil Rights Act, campus safety concerns over disruptions to the learning environment, and the awarding of federal student aid under the Higher Education Act.
• The Committee on Oversight and Accountability is investigating the sources of funding and other support flowing to groups espousing pro-Hamas propaganda and engaged in antisemitic harassment and intimidation of students. The Committee on Oversight and Accountability is the principal oversight committee of the US House of Representatives and has broad authority to investigate “any matter” at “any time” under House Rule X.
• The Committee on Ways and Means has been investigating several universities since November 15, 2023, when the Committee held a hearing entitled From Ivory Towers to Dark Corners: Investigating the Nexus Between Antisemitism, Tax-Exempt Universities, and Terror Financing. The Committee followed the hearing with letters to those institutions on January 10, 202
Unit 8 - Information and Communication Technology (Paper I).pdfThiyagu K
This slides describes the basic concepts of ICT, basics of Email, Emerging Technology and Digital Initiatives in Education. This presentations aligns with the UGC Paper I syllabus.
Safalta Digital marketing institute in Noida, provide complete applications that encompass a huge range of virtual advertising and marketing additives, which includes search engine optimization, virtual communication advertising, pay-per-click on marketing, content material advertising, internet analytics, and greater. These university courses are designed for students who possess a comprehensive understanding of virtual marketing strategies and attributes.Safalta Digital Marketing Institute in Noida is a first choice for young individuals or students who are looking to start their careers in the field of digital advertising. The institute gives specialized courses designed and certification.
for beginners, providing thorough training in areas such as SEO, digital communication marketing, and PPC training in Noida. After finishing the program, students receive the certifications recognised by top different universitie, setting a strong foundation for a successful career in digital marketing.
How to Build a Module in Odoo 17 Using the Scaffold MethodCeline George
Odoo provides an option for creating a module by using a single line command. By using this command the user can make a whole structure of a module. It is very easy for a beginner to make a module. There is no need to make each file manually. This slide will show how to create a module using the scaffold method.
2024.06.01 Introducing a competency framework for languag learning materials ...Sandy Millin
http://sandymillin.wordpress.com/iateflwebinar2024
Published classroom materials form the basis of syllabuses, drive teacher professional development, and have a potentially huge influence on learners, teachers and education systems. All teachers also create their own materials, whether a few sentences on a blackboard, a highly-structured fully-realised online course, or anything in between. Despite this, the knowledge and skills needed to create effective language learning materials are rarely part of teacher training, and are mostly learnt by trial and error.
Knowledge and skills frameworks, generally called competency frameworks, for ELT teachers, trainers and managers have existed for a few years now. However, until I created one for my MA dissertation, there wasn’t one drawing together what we need to know and do to be able to effectively produce language learning materials.
This webinar will introduce you to my framework, highlighting the key competencies I identified from my research. It will also show how anybody involved in language teaching (any language, not just English!), teacher training, managing schools or developing language learning materials can benefit from using the framework.
Acetabularia Information For Class 9 .docxvaibhavrinwa19
Acetabularia acetabulum is a single-celled green alga that in its vegetative state is morphologically differentiated into a basal rhizoid and an axially elongated stalk, which bears whorls of branching hairs. The single diploid nucleus resides in the rhizoid.
1. Missing data
Julian Higgins
MRC Biostatistics Unit
Cambridge, UK
with thanks to Ian White, Fred Wolf, Angela Wood, Alex Sutton
SMG advanced workshop, Cardiff, March 4-5, 2010
2. Potential sources of missing data
in a meta-analysis
Potential sources of missing data
in a meta-analysis
• Studies not found
• Outcome not reported
• Outcome partially reported (e.g. missing SD or SE)
• Extra information required for meta-analysis
(e.g. imputed correlation coefficient)
• Missing participants
• No information on study characteristic for heterogeneity
analysis
3. Concepts in missing dataConcepts in missing data
1. ‘Missing completely at random’
• As if missing observation was randomly picked
• Missingness unrelated to the true (missing) value
• e.g.
– (genuinely) accidental deletion of a file
– observations measured on a random sample
– statistics not reported due to ignorance of their
importance(?)
• OK (unbiased) to analyse just the data available, but
sample size is reduced - which may not be ideal
• Not very common!
4. Concepts in missing dataConcepts in missing data
2. ‘Missing at random’
• Missingness depends on things you know about, and
not on the missing data themselves
• e.g.
– older people more likely to drop out (irrespective of
the effect of treatment)
– recent cluster trials more likely to report intraclass
correlation
– experimental treatment more likely to cause drop-
out (independent of therapeutic effect)
• Usually surmountable
5. Concepts in missing dataConcepts in missing data
3. ‘Informatively missing’
• The fact that an observation is missing is a
consequence of the value of the missing observation
• e.g.
– publication bias (non-significant result leads to
missingness)
– drop-out due to treatment failure (bad outcome
leads to missingness)
– selective reporting
• methods not reported because they were poor
• unexpected or disliked findings suppressed
• Requires careful consideration
• Very common!
6. Strategies for dealing with missing dataStrategies for dealing with missing data
• Ignore (exclude) the missing data
– Generally unwise
• Obtain the missing data
– Contact primary investigators
– Compute from available information
• Re-interpret the analysis
– Focus on sub-population, or place results in context
• Simple imputation
– ‘Fill in’ a value for each missing datum
• Multiple imputation
– Impute several times from a random distribution and
combine over results of multiple analyses
• Analytical techniques
– e.g. EM algorithm, maximum likelihood techniques
7. Strategies for dealing with missing dataStrategies for dealing with missing data
• Do sensitivity analyses
– But what if it makes a big difference?
9. Missing studiesMissing studies
• If studies are informatively missing, this is
publication bias
• If:
1. a funnel plot is asymmetrical
2. publication bias is assumed to be the cause
3. a fixed-effect of random-effects assumption is
reasonable
• then trim and fill provides a simple imputation method
– but I think these assumptions are difficult to believe
– I favour extrapolation of a regression line
Moreno, Sutton, Turner, Abrams, Cooper, Palmer and Ades. BMJ 2009; 339: b2981
10. The trim and fill methodThe trim and fill method
• Formalises the use of the funnel plot
• Rank-based ‘data augmentation’ technique
– estimates and adjusts for the number and outcomes of
missing studies
• Relatively simple to implement
• Simulations suggest it performs well
Duval and Tweedie. Biometrics 2000; 56: 455–463
11. The trim and fill method (cont)The trim and fill method (cont)
• Funnel plot of the
effect of gangliosides
in acute ischaemic
stroke
12. The trim and fill method (cont)The trim and fill method (cont)
• Estimate number and
trim asymmetric
studies
13. The trim and fill method (cont)The trim and fill method (cont)
• Calculate ‘centre’ of
remainder
14. The trim and fill method (cont)The trim and fill method (cont)
• Replace the trimmed
studies and fill with
their missing mirror
image counterparts to
estimate effect and
its variance
15. Extrapolation of funnel plot
regression line
Extrapolation of funnel plot
regression line
Odds ratio
0.1 0.3 1 3
3
2
1
0
100.6
Standarderror
Maximum
observed precision
or
Infinite precision
18. Missing standard deviationsMissing standard deviations
• Make sure these are computed where possible (e.g. from
T statistics, F statistics, standard errors, P values)
– For non-exact P values (e.g. P < 0.05, P > 0.05)?
• SDs are not necessary: the generic inverse variance
method in RevMan can analyse estimates and SEs
• Where SDs are genuinely missing and needed, consider
imputation
– borrow from a similar study?
– is it better to impute than to leave the study out of the
meta-analysis?
19. Missing standard deviationsMissing standard deviations
“It was decided that missing standard deviations for caries
increments that were not revealed by contacting the
original researchers would be imputed through linear
regression of log(standard deviation)s on log(mean
caries) increments. This is a suitable approach for caries
prevention studies since, as they follow an approximate
Poisson distribution, caries increments are closely related
to their standard deviations (van Rijkom 1998).”
Marinho, Higgins, Logan and Sheiham, CDSR 2003, Issue 1. Art. No.: CD002278
20. Sensitivity analysis
Asthma self management & ER visits BMJ 2003, from Fred Wolf
Sensitivity analysis
Asthma self management & ER visits BMJ 2003, from Fred Wolf
N = 7 studies (n=704)
(no missing data studies)
• SMD (CI)
-0.25 (-0.40, -0.10)
• Larger effect size
(over estimation?)
• Less precision
(more error?)
N = 12 studies (n=1114)
(5 with imputed data)
• SMD (CI)
-0.21 (-0.33, -0.09)
• Smaller effect size
• Greater precision
• > statistical Power
(> # studies & subjects)
Other possible sensitivity analyses:
vary imputation methods & assumptions
22. Missing correlation coefficientsMissing correlation coefficients
• Common problem for change-from-baseline / ANCOVA,
cross-over trials, cluster-randomized trials, combining or
comparing outcomes / time points, multivariate meta-
analysis
• If analysis fails to account for pairing or clustering, can
adjust it using an imputed correlation coefficient
– Correlation can often be computed for cross-over trials
from paired and unpaired results
• then lent to another study that doesn’t report paired results
– For cluster-randomized trials, ICC resources exist:
• Campbell et al, Statistics in Medicine 2001; 20:391-9
• Ukoumunne et al, Health Technology Assessment 1999; 3 (no 5)
• Health Services Research Unit Aberdeen
www.abdn.ac.uk/hsru/epp/cluster.shtml
23. Missing correlation coefficientsMissing correlation coefficients
• Guidance on change from baseline, cross-over trials,
cluster-randomized trials available in the Handbook
• See also Follmann (JCE, 1992)
25. Missing outcomes from individual
participants
Missing outcomes from individual
participants
• Consider first a dichotomous outcome
• From each trial, a 32 table
nCmCfCrCControl
nTmTfTrTTreatment
TotalMissingFailureSuccess
27. Typical practice in Cochrane reviewsTypical practice in Cochrane reviews
• Ignore missing data completely
– Call the above an available case analysis
– often called complete case analysis
• Impute success, failure, best-case, worst-case etc
– Call this an imputed case analysis
– analysis often proceeds assuming imputed data are
known
29. Available case analysisAvailable case analysis
• (a.k.a. complete case analysis)
• Ignore missing data
• May be biased
• Over-precise
– Consider a trial of 80 patients,
with 60 followed up. We
should be more uncertain with
60/80 of data than with 60/60
of data
Treatment Control
vs
34. Using reasons for missingnessUsing reasons for missingness
• Sometimes reasons for missing data are available
• Use these to impute particular outcomes for missing
participants
• e.g. in haloperidol trials:
– Lack of efficacy, relapse: impute failure
– Positive response: impute success
– Adverse effects, non-compliance: impute control event rate
– Loss to follow-up, administrative
reasons, patient sleeping: impute group-specific event rate
Higgins , White and Wood. Clinical Trials 2008; 5: 225–239
35. Application to haloperidolApplication to haloperidol
1.8 (1.4, 2.1) Q=22
1.5 (1.2, 1.7) Q=31
1.3 (1.1, 1.5) Q=34
1.4 (1.2, 1.6) Q=31
0.9 (0.8, 1.1) Q=62
2.4 (1.9, 3.0) Q=22
1.9 (1.5, 2.4) Q=22
1.2 (1.0, 1.3) Q=40
1.6 (1.3, 1.9) Q=27 (16 df)
Fixed-effect meta-analysis
incorporating available reasons for
missing data
according to observed group-specific
event rate
according to observed treatment
event rate, pT
according to observed control event
rate, pC
missing = success (1)
missing = failure (0)
Imputation
worst case scenario for treatment
best case scenario for treatment
nothing (available case analysis)
36. Application to haloperidolApplication to haloperidol
Imputation
1.3 (28%)
1.0 (36%)
1.0 (25%)
1.0 (37%)
0.5 (33%)
2.5 (30%)
1.4 (25%)
0.9 (36%)
1.0 (31%)
RR (weight)
1.8 (22%)
1.5 (32%)
1.1 (51%)
1.3 (27%)
0.7 (26%)
4.0 (11%)
2.4 (10%)
1.1 (47%)
1.5 (19%)
RR (weight)
incorporating available reasons for
missing data
according to observed group-specific
event rate
according to observed treatment
event rate, pT
according to observed control event
rate, pC
missing = success (1)
missing = failure (0)
worst case scenario for treatment
best case scenario for treatment
nothing (available case analysis)
SelmanBeasley
37. Generalization of imputing schemesGeneralization of imputing schemes
• Consider control group only
• To reflect informative missingness,
specify odds ratio comparing
event rate among missing participants
vs
event rate among observed participants
• Call this informative missingness odds
ratio (IMORC)
• Similarly for treatment group (IMORT)
vs
38. incorporating available reasons for
missing data
11according to observed group-specific
event rate
1all according to observed treatment
event rate, pT
1all according to observed control
event rate, pC
missing = success (1)
00missing = failure (0)
IMORCIMORTImputation
C T
C T
p 1 p
1 p p
T C
T C
p 1 p
1 p p
M
T T
M
T T
p 1 p
1 p p
M
C C
M
C C
p 1 p
1 p p
0 [or ] [or 0]
Impute to create worst case scenario
for treatment
[or 0]0 [or ]
to create best case scenario for
treatment
ConnectionsConnections
39. Weighting studiesWeighting studies
• Suppose we were to treat imputed data as known
– Standard errors will tend to be too small and weights too
big
• Some simple alternative weights might be used
– use the available case weights
– use an effective sample size argument, but with revised
event rates
• We have derived more suitable weights based on IMORs: an
analytic strategy
40. metamiss in Statametamiss in Stata
• All the above strategies are implemented in our program
metamiss
• Download it within Stata using
ssc install metamiss
• or download the latest version using
net from http://www.mrc-
bsu.cam.ac.uk/BSUsite/
Software/pub/software/stata/meta
White and Higgins. The Stata Journal 2009; 9: 57-69.
41.
42. Overall (I-squared = 26.8%, p = 0.148)
Selman
Nishikawa_82
Borison
Howard
ID
Chouinard
Reschke
Marder
Simpson
Serafetinides
Durost
Garry
Spencer
Nishikawa_84
Beasley
Bechelli
Vichaiya
Arvanitis
Study
1.2 1 5
Risk ratio, intervention vs. control
Haloperidol meta-analysis: using
reasons
Haloperidol meta-analysis: using
reasons
43. Implications for practiceImplications for practice
• We desire a strategy for primary analysis and for
sensitivity analyses
• Primary analysis
– present available case analysis as a reference point
– make use of available reasons for missingness
– consider using IMORs based on external evidence or
heuristic arguments
44. Implications for practiceImplications for practice
• Sensitivity analysis
– Best-case/worst case scenarios are fine but probably
too unrealistic
– We suggest to use IMORs to move towards these
analyses, using clinically realistic values
– Should also evaluate impact of changing weights
46. Application to haloperidolApplication to haloperidol
1.8 (1.4, 2.2) Q=22
1.4 (1.1, 1.7) Q=34
1.7 (1.4, 2.1) Q=25
1.4 (1.2, 1.7) Q=31
1.6 (1.3, 1.9) Q=27 (16 df)
Fixed-effect meta-analysis
IMORT = 1/2, IMORC = 1/2
IMORT = 1/2, IMORC = 2
Imputation
Available case
IMORT = 2, IMORC = 1/2
IMORT = 2, IMORC = 2
Conclusion: no important impact of missing data
47. Missing data introduce extra uncertaintyMissing data introduce extra uncertainty
• Even if we make assumptions about the missing data, we
don’t know if they are correct
• For example, we don’t know if IMOR is 0 or 1.
• Allowing for missing data should introduce extra
uncertainty
– increase the standard error
– hence down-weight trials with more missing data in the
meta-analysis
• We do this by allowing for uncertainty in the IMORs
– sdlogimor(1) option in metamiss specifies a sensible
degree of uncertainty about the IMOR
White, Wood and Higgins. Statistics in Medicine 2008; 27: 711–727.
48. sdlogimor(): technical detailssdlogimor(): technical details
• We place a normal distribution on the log IMOR
• For example, we might give the log IMOR mean -1 and
standard deviation 1
– our “best guess” is log IMOR = -1 (IMOR = 0.37)
– we are 68% sure the log IMOR lies between -2 and
0 (IMOR lies between 0.14 and 1)
– we are 95% sure the log IMOR lies between -3 and
+1 (IMOR lies between 0.05 and 2.7)
49. Continuous dataContinuous data
• Many methods available to primary trialists, e.g.
– last observation carried forward
– regression imputation
– analytic approaches
• Imputing is much more difficult for the meta-analyst
• You can
– Impute using the mean (corresponds to increasing
sample size to include missing people)
– Impute using a specific value
– Apply an ‘IMMD’ or ‘IMMR’
(‘informative missingness mean difference’ or ‘…mean
ratio’)
• Methods to properly account for uncertainty are not yet
developed
50. RemarksRemarks
• Many imputing schemes are available, but these should
not be used to enter ‘filled-out’ data into RevMan
• However, the point estimates from such analyses may be
used with weights from an available case analysis
(requires analyses outside of RevMan)
51. Remarks (ctd)Remarks (ctd)
• Informative missingness odds ratios offer advantages of
– a generalization of the imputation schemes
– can reflect ‘realistic’ scenarios
– statistical expressions for variances
– ability to incorporate prior distributions on IMORs
• Sensitivity analyses are essential, and should ideally
address both estimates and weights
• See: Higgins JPT, White IR, Wood AM. Missing outcome data in meta-
analysis of clinical trials: development and comparison of methods, with
suggestions for practice. (Clinical Trials)
52. No information on study characteristic
for heterogeneity analysis
No information on study characteristic
for heterogeneity analysis
53. General recommendations for dealing
with missing data
General recommendations for dealing
with missing data
• Whenever possible, contact original investigators to
request missing data
• Make explicit assumptions of methods used to address
missing data
• Conduct sensitivity analyses to assess how sensitive
results are to reasonable changes in assumptions that
are made
• Address potential impact of missing data (known or
suspected) on findings of the review in the Discussion
section