Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Meta-Analysis 2.0 #248

Merged
merged 127 commits into from
Nov 6, 2024
Merged

Meta-Analysis 2.0 #248

merged 127 commits into from
Nov 6, 2024

Conversation

FBartos
Copy link
Collaborator

@FBartos FBartos commented Oct 11, 2024 •
edited
Loading

A complete rewrite of the classical meta-analysis. Altogether, this PR adds:

  • Effect Size Computation
  • Funnel Plot
  • Classical Meta-Analysis
  • Classical Meta-Analysis (Multilevel/Multivariate)
    (checkbox indicates whether the portion was reviewed)

fixes: https://github.com/jasp-stats/INTERNAL-jasp/issues/2316
fixes: https://github.com/jasp-stats/INTERNAL-jasp/issues/2267
fixes: jasp-stats/jasp-issues#2381
fixes: jasp-stats/jasp-issues#2371
fixes: jasp-stats/jasp-issues#2210
fixes: jasp-stats/jasp-issues#1804
fixes: jasp-stats/jasp-issues#1543
fixes: jasp-stats/jasp-issues#1542
fixes: jasp-stats/jasp-issues#373

Remaining issues in this PR:

  • Spatial and AR random effects options in Multilevel Multivariate Meta-Analysis not fixed yet
  • Dataset loading not updated (also Dropdown variable selection broken: https://github.com/jasp-stats/INTERNAL-jasp/issues/2662)
  • Unit tests not updated
  • Analysis update not possible messages

This is a large PR, I'm sorry Don.

@FBartos FBartos requested a review from vandenman October 11, 2024 15:35
@tomtomme tomtomme mentioned this pull request Oct 14, 2024
Open
@FBartos FBartos requested a review from juliuspfadt October 21, 2024 21:35
@FBartos
Copy link
Collaborator Author

FBartos commented Oct 21, 2024

@juliuspfadt - Could you please review the "Funnel Plot" portion of the update?
It is quite a separate piece of code and it can be reviewed individually. I will still update the data loading & fix some bugs in the rest tomorrow.

@juliuspfadt
Copy link

why 3 "n" in funnnelplot.R

@FBartos
Copy link
Collaborator Author

FBartos commented Oct 22, 2024

funnnelplot

lol, thanks

@juliuspfadt
Copy link

Screenshot 2024-10-23 at 09 26 44
I would think the study label is too close to the dot

@juliuspfadt
Copy link

shapeFail.jasp.zip
I suppose a continuous variable for the shape makes little sense... error nonetheless

@juliuspfadt
Copy link

juliuspfadt commented Oct 23, 2024 •
edited
Loading

error.jasp.zip
Screenshot 2024-10-23 at 09 32 07

same happens for Paule-Mandel (HU) Method, not only Hunter Schmidt SSC
Also generalised Q-stat + (MU)

@juliuspfadt
Copy link

Screenshot 2024-10-23 at 09 36 28

juliuspfadt
juliuspfadt requested changes Oct 23, 2024
View reviewed changes
R/funnelplot.R Outdated Show resolved Hide resolved
R/funnelplot.R Outdated Show resolved Hide resolved
R/funnelplot.R Show resolved Hide resolved
R/funnelplot.R Show resolved Hide resolved
R/funnelplot.R Show resolved Hide resolved
inst/qml/FunnelPlot.qml Show resolved Hide resolved
@FBartos
Copy link
Collaborator Author

FBartos commented Oct 23, 2024

@juliuspfadt the 5 commits above should fix the issues mentioned, thanks!

@FBartos
Copy link
Collaborator Author

FBartos commented Oct 25, 2024

@juliuspfadt
The Effect Size Computation is now also ready for review. You need the newest (yesterday's) nightly to have all qml working though. This is quite a convenient dataset for testing:
gibson2002.csv

@FBartos
Copy link
Collaborator Author

FBartos commented Oct 25, 2024

And meta-analysis is also ready for review :)

@FBartos FBartos requested review from juliuspfadt and removed request for vandenman October 25, 2024 12:31
@juliuspfadt
Copy link

remove saveRDS stuff

@FBartos
Copy link
Collaborator Author

FBartos commented Oct 27, 2024

@juliuspfadt completely ready for a review! :)

@FBartos FBartos requested a review from vandenman October 28, 2024 09:12
@FBartos
Copy link
Collaborator Author

FBartos commented Oct 28, 2024 •
edited
Loading

So, here's just one last reorganization; sorry for the confusion.

@juliuspfadt reviews

  • Effect Size Computation
  • Funnel Plot

@vandenman reviews

  • Classical Meta-Analysis
  • Classical Meta-Analysis (Multilevel/Multivariate)

I will add unit tests, data library examples, and documentation in a later PR.

vandenman
vandenman requested changes Oct 29, 2024
View reviewed changes
Copy link
Contributor

@vandenman vandenman left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Some minor issues, see below and attached.
bugs_metaanalysis.jasp.zip

(rename and drop .zip).

For how many changes there are it works surprisingly well.

R/classicalmetaanalysis.R Outdated Show resolved Hide resolved
R/classicalmetaanalysis.R Outdated Show resolved Hide resolved
R/classicalmetaanalysiscommon.R Outdated Show resolved Hide resolved
@FBartos FBartos requested a review from vandenman October 30, 2024 11:59
@juliuspfadt
Copy link

I think you fixed the errors. Nice is also the error message at the top for old jasp files, but there seems to be some spaces or line break missing. I also still think the study labels are too close to the dots.
Screenshot 2024-10-30 at 17 58 48

@juliuspfadt
Copy link

would it not be nice of these boxes were aligned... for (ES):
Screenshot 2024-10-30 at 18 14 57

@juliuspfadt
Copy link

ES: I think in this table it would make sense to have "Note. Error etc." instead of the "Note." being all the way down
Screenshot 2024-10-30 at 18 17 51

@juliuspfadt
Copy link

juliuspfadt commented Oct 30, 2024 •
edited
Loading

Screenshot 2024-10-30 at 18 43 57

I think this scrollbar is unnecessary. But I get that qml is annoying. If it is an easy fix... @boutinb

@juliuspfadt
Copy link

I open the analysis and enter nothing but a dataset, and I already see a table with an error. that does not seem very user friendly.
Screenshot 2024-10-30 at 18 48 08

@juliuspfadt
Copy link

juliuspfadt commented Oct 30, 2024 •
edited
Loading

ESfail.jasp.zip
Screenshot 2024-10-30 at 19 06 00

@juliuspfadt
Copy link

Cronbach should be "Coefficient" and alpha as greek letter.

@juliuspfadt
Copy link

Seems like the "Error in Step 1" for this should already be caught in qml.
Screenshot 2024-10-30 at 19 33 08

FBartos and others added 26 commits November 6, 2024 09:50
@FBartos
fix multivariate meta-analysis: Baujat not available for rma.mv
7d5f897
@FBartos
fix multivariate meta-analysis: profile likelihood plot
1d8b26a
@FBartos
fix meta-analysis: weights and GENQ estimators
6e295d5
@FBartos
fix meta-analysis: omnibus moderation test specification
14a1247
@FBartos
fix multivariate meta-analysis: fix spatial random effects & known co…
e377126 
…rrelational structures
@FBartos
add testing data sets
5ce0170
@FBartos
fix multivariate meta-analysis: add inclusion test for random effects
8c6d712
@FBartos
hide confidence intervals for random effects option
c382d64
@FBartos
fix multivariate meta-analysis: optimizer settings
2da8d27
@FBartos
add metafor code injection
30bbd76
@FBartos @vandenman
Apply suggestions from code review
26666a1 
Co-authored-by: Don van den Bergh <[email protected]>
@FBartos
typos
046a9d3
@FBartos
fixing .maExtendMetaforCallFromOptions
2104707
@FBartos
make nicer error messages
14b1066
@FBartos
fixing forest plot with heterogeneity test
609f302
@FBartos
fixing pooled effect size estimates (under no moderation but heterosk…
fcae175 
…edasticity)
@FBartos
fix empty weights
d4cb595
@FBartos
fix funnel plot: allow offsetting x-axis label position + larger default
3d205f0
@FBartos
fix effect size computation: no errors on no input
1eef1bf
@FBartos
fix effect size computation: better handling of observed effects
8854086
@FBartos
fix effect size computation: cronbachsAlpha -> coefficientAlpha
ff097d0
@FBartos
fix effect size computation: fixing incorrect variance type
0a09a2e
@FBartos
fix meta-analysis: prevent issues with dropping the intercept
32fd0b1
@FBartos
Update effectsizecomputation.R
304d88a
@FBartos
computedcolumnsNamesEffectSize -> computedColumnsNamesEffectSize
7a10472
@FBartos
add documentation (part 1/2)
f873fdb
@FBartos FBartos merged commit 96281c7 into jasp-stats:master Nov 6, 2024
1 of 5 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@juliuspfadt juliuspfadt juliuspfadt approved these changes

@vandenman vandenman vandenman approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
3 participants
@FBartos @juliuspfadt @vandenman