diff --git a/DESCRIPTION b/DESCRIPTION index 0f34bfc6..ba9960a1 100644 --- a/DESCRIPTION +++ b/DESCRIPTION @@ -34,7 +34,7 @@ Authors@R: Description: Implementation of the fast algorithm for wild cluster bootstrap inference developed in Roodman et al (2019, STATA Journal) for linear regression models - https://journals.sagepub.com/doi/abs/10.1177/1536867X19830877?journalCode=stja, + , which makes it feasible to quickly calculate bootstrap test statistics based on a large number of bootstrap draws even for large samples - as long as the number of bootstrapping clusters diff --git a/README.Rmd b/README.Rmd index e085abfe..cb1f21e7 100644 --- a/README.Rmd +++ b/README.Rmd @@ -41,7 +41,7 @@ The package's central function is `boottest()`. It allows the user to test univa The `fwildclusterboot` package currently supports multi-dimensional clustering and one-dimensional hypotheses. It supports regression weights, multiple distributions of bootstrap weights, fixed effects, restricted (WCR) and unrestricted (WCU) bootstrap inference and subcluster bootstrapping for few treated clusters [(MacKinnon & Webb, (2018))](https://academic.oup.com/ectj/article-abstract/21/2/114/5078969). -If you are interested in the wild cluster bootstrap for IV models [(Davidson & MacKinnon, 2010)](https://www.tandfonline.com/doi/abs/10.1198/jbes.2009.07221]) or want to test multiple joint hypotheses, you can use the [wildboottestjlr](https://github.com/s3alfisc/wildboottestjlr) package, which is an R wrapper of the [WildBootTests.jl](https://github.com/droodman/WildBootTests.jl) Julia package. While `fwildclusterboot` is already quite fast (see the benchmarks below), +If you are interested in the wild cluster bootstrap for IV models [(Davidson & MacKinnon, 2010)](https://www.tandfonline.com/doi/abs/10.1198/jbes.2009.07221) or want to test multiple joint hypotheses, you can use the [wildboottestjlr](https://github.com/s3alfisc/wildboottestjlr) package, which is an R wrapper of the [WildBootTests.jl](https://github.com/droodman/WildBootTests.jl) Julia package. While `fwildclusterboot` is already quite fast (see the benchmarks below), the implementation of the wild bootstrap for OLS in `WildBootTests.jl` is - after compilation - orders of magnitudes faster, in particular for problems with a large number of clusters. diff --git a/README.md b/README.md index b194c833..0a62aeea 100644 --- a/README.md +++ b/README.md @@ -47,8 +47,8 @@ subcluster bootstrapping for few treated clusters [(MacKinnon & Webb, If you are interested in the wild cluster bootstrap for IV models [(Davidson & MacKinnon, -2010)](https://www.tandfonline.com/doi/abs/10.1198/jbes.2009.07221%5D) -or want to test multiple joint hypotheses, you can use the +2010)](https://www.tandfonline.com/doi/abs/10.1198/jbes.2009.07221) or +want to test multiple joint hypotheses, you can use the [wildboottestjlr](https://github.com/s3alfisc/wildboottestjlr) package, which is an R wrapper of the [WildBootTests.jl](https://github.com/droodman/WildBootTests.jl) Julia diff --git a/cran-comments.html b/cran-comments.html new file mode 100644 index 00000000..629930c4 --- /dev/null +++ b/cran-comments.html @@ -0,0 +1,415 @@ + + + + + + + + + + + + + +cran-comments + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ + + + + + + +
+

Re-submission of 0.7

+

Fixed both issues below:

+ +

The closing %5D = “]” should likely be removed.

+
    +
  • The Description field contains
    • +
+

https://journals.sagepub.com/doi/abs/10.1177/1536867X19830877?journalCode=stja,

+

Please write DOIs as doi:10.1177/1536867X19830877 in the Description field without the rest of the URL.

+
+
+

Version 0.7

+
    +
  • checks pass on github actions: windows, mac, ubuntu
    • +
  • checks pass local cmd check
    • +
  • checks pass on r-develop
    • +
+
+
+

Version 0.5.1

+

Please apologize the quick re-submission. Version 0.5.1 fixes a bug introduced in version 0.5.

+
    +
  • checks pass on github actins
    • +
  • checks pass local cmd check
    • +
  • checks pass on r-develop - note regarding fast re-submission
    • +
+
+
+

Version 0.5

+
    +
  • checks pass on github actins
    • +
  • checks pass local cmd check
    • +
  • checks pass on r-develop; note regarding codecov link has been addressed
    • +
+

Changes:

+
    +
  • Version 0.5 fixes an error for the bootstrap with weighted least squares introduced with version 0.4. All unit tests that compare fwildclusterboot with weighted least squares results from boottest.stata pass. In particular, enumerated cases pass with exact equality (in such cases, the bootstrap weights matrices are exactly identical in both R and Stata).
    • +
  • boottest() now stops if fixest::feols() deletes non-NA values (e.g. singleton fixed effects deletion) and asks the user to delete such rows prior to estimation via feols() & boottest(). Currently, boottest()'s pre-processing cannot handle such deletions - this remains future work.
    • +
  • To align fwildclusterboot with Stata’s boottest command (Roodman et al, 2019), Mammen weights are no longer enumerated in fwildclusterboot::boottest().
    • +
  • boottest() no longer sets an internal seed (previously set.seed(1)) if no seed is provided as a function argument.
    • +
  • Sampling of the bootstrap weights is now powered by the dqrng package, which speeds up the creation of the bootstrap weights matrix. To set a “global” seed, one now has use the dqset.seed() function from the dqrng package, which is added as a dependency.
    • +
+
+
+

Version 0.4 first submission - re-submission

+

Apparently, I did not fix the openBLAS error appropriately. The test tolerance level has again been reduced, from 1e-04 to 1e-02. The relative test difference found in the openBLAS tests is 0.003568575, so the test should now pass. If this does not fix the issue, I will remote all tests in test_type from CRAN.

+

The error stated " test_type.R………………. 139 tests 1 fails test_type.R………………. 140 tests 2 fails test_type.R………………. 140 tests 2 fails test_type.R………………. 140 tests 2 fails test_type.R………………. 140 tests 2 fails test_type.R………………. 140 tests 2 fails test_type.R………………. 140 tests 2 fails test_type.R………………. 141 tests 2 fails test_type.R………………. 142 tests 2 fails test_type.R………………. 143 tests 2 fails test_type.R………………. 144 tests 2 fails test_type.R………………. 145 tests 2 fails test_type.R………………. 146 tests 2 fails test_type.R………………. 147 tests 2 fails test_type.R………………. 148 tests 2 fails test_type.R………………. 149 tests 2 fails test_type.R………………. 150 tests 2 fails test_type.R………………. 151 tests 2 fails test_type.R………………. 152 tests 2 fails test_type.R………………. 153 tests 2 fails test_type.R………………. 154 tests 2 fails test_type.R………………. 155 tests 2 fails test_type.R………………. 156 tests 2 fails test_type.R………………. 157 tests 2 fails test_type.R………………. 158 tests 2 fails test_type.R………………. 159 tests 2 fails test_type.R………………. 160 tests 2 fails 4.6s —– FAILED[data]: test_type.R<345–345> call| expect_equivalent(boot_fixest_c\(conf_int, boot_felm_c\)conf_int, call| –> tol = 1e-04) diff| Mean relative difference: 0.003568575 —– FAILED[data]: test_type.R<346–346> call| expect_equivalent(boot_felm_c\(conf_int, boot_lm\)conf_int, tol = 1e-04) diff| Mean relative difference: 0.00357477 Error: 2 out of 868 tests failed In addition: Warning messages: 1: There are only 1024 unique draws from the rademacher distribution for 10 clusters. Therefore, B = 1024 with full enumeration. Consider using webb weights instead. 2: Further, note that under full enumeration and with B = 1024 bootstrap draws, only 2^(#clusters - 1) = 512 distinct t-statistics and p-values can be computed. For a more thorough discussion, see Webb Reworking wild bootstrap based inference for clustered errors (2013). 3: There are only 1024 unique draws from the rademacher distribution for 10 clusters. Therefore, B = 1024 with full enumeration. Consider using webb weights instead. 4: Further, note that under full enumeration and with B = 1024 bootstrap draws, only 2^(#clusters - 1) = 512 distinct t-statistics and p-values can be computed. For a more thorough discussion, see Webb Reworking wild bootstrap based inference for clustered errors (2013). Execution halted "

+
+
+

Version 0.4 re-submission

+

The openBLAS error on CRAN has been fixed (point for in version 0.4 submission. Some tests for exact equality failed with relative difference e-05 on openBLAS. In consequence, all exact tests are set to reltol = 1e-04. This should fix all CRAN errors.

+
+
+

Version 0.4 submission

+
    +
  • New feature I: boottest() now allows for univariate tests that involve multiple variables. E.g. one can now test hypothesis as ´var1 + var2 = c´ where c is a scalar. More details on the syntax can be found in the vignette. All methods of for objects of class boottest have been updated.
    • +
  • New feature II: boottest() now also supports “equal-tailed” p-values and one-sided hypotheses. For one-sided tests, confidence intervals are currently not supported.
    • +
  • Internal changes: To allow for multivariable tests, the boot_algo2() function has slightly been modified. invert_p_val2() is superseded by invert_p_val().
    • +
  • Further, a CRAN error is fixed - some tests for exact equality failed with relative difference e-05 on openBLAS. In consequence, all exact tests are set to reltol = 1e-04. This should fix all CRAN errors.
    • +
+
+
+

Version 0.3.7 submission

+ +
+
+

Version 0.3.6 submission, Resubmission 2

+
    +
  • Another error due to “noSuggests” / vignette - package fixest was used, but not loaded
    • +
+
+
+

Version 0.3.6 submission, Resubmission

+
    +
  • Adressed errors related to suggests described below (reason for error; unprotected library(lfe) statements in tests)
    • +
  • noSuggests error due to missing library(“pkg”) statements in vignette
    • +
  • lfe now install on ubuntu dev
    • +
+
+
+

Version 0.3.6 submission

+ +

Please correct before 2021-08-08 to safely retain your package on CRAN.

+

Packages in Suggests should be used conditionally: see ‘Writing R Extensions’. This needs to be corrected even if the missing package(s) become available."

+

Solution: add if(requireNamespace(“pkg”)) for suggested packages in vignette, examples, tests

+

I have checked package installation via github actions on windows, mac, ubuntu & ubuntu devel - installation fails on ubuntu devel because the “lfe” package cannot be installed.

+

Further, the packages passes the checks via devtools::check_win_devel()

+
+
+

Version 0.3.5 submission

+
    +
  • fixes two bugs introduced with version 0.3.3
    • +
+
+
+

Version 0.3.4 submission

+
    +
  • fix an error in the vignette that caused multiple erros on CRAN
    • +
+
+
+

Version 0.3.3 re-submission

+
    +
  • updated date in description
    • +
+
+
+

Version 0.3.3 submission

+
    +
  • implements full enumeration for rademacher and mammen weights if 2^k < B, where k is the number of clusters and B the number of bootstrap iterations
    • +
+
+
+

Version 0.3.2 re-submission

+
    +
  • update date in description
    • +
+
+
+

Version 0.3.2 submission

+
    +
  • Fixes a CRAN test error message for Oracle Solaris.
    • +
+
+
+

Another resubmission

+
    +

  • ** running tests for arch ‘i386’ … [296s] OK ** running tests for arch ‘x64’ … [239s] OK

    +

    Please reduce the test timings by using

    +
      +
    • small toy data only
      • +
    • few iterations
      • +
    • or by running less important tests only conditionally if some environment variable is set that you only define on your machine?
      • +
    • +
+

Long-running tests now only run on developing versions (4-digit) - this submission is 3-digit (0.3.1)

+
+
+

This is a third re-submission

+
    +
  • Please remove the single quotes from function names in your description. e.g.: ‘lm()’ –> lm()
    • +
+

DONE

+
    +
  • Please add to .Rd files regarding exported methods and explain the functions results in the documentation. Please write about the structure of the output (class) and also what the output means. (If a function does not return a value, please document that too, e.g. or similar) Missing Rd-tags: boottest.Rd: check_set_nthreads.Rd: cpp_get_nb_threads.Rd: create_data_2.Rd: dot-onLoad.Rd: getBoottest_nthreads.Rd: invert_p_val2.Rd: p_val_null2.Rd:
    • +
+

DONE. Added return values to all functions OR functions are no longer exported.

+
    +
  • You have examples for unexported functions. Please either omit these examples or export these functions. Used ::: in documentation: man/boottest.felm.Rd: felm_fit <- felm(proposition_vote ~ treatment + ideology1 + log_income | Q1_immigration, data = fwildclusterboot:::create_data_2(N = 1000, N_G1 = 10, icc1 = 0.91, N_G2 = 10, icc2 = 0.51, numb_fe1 = 10, numb_fe2 = 10, seed = 12345)) man/boottest.fixest.Rd: feols_fit <- feols(proposition_vote ~ treatment + ideology1 + log_income, fixef = “Q1_immigration”, data = fwildclusterboot:::create_data_2(N = 1000, N_G1 = 10, icc1 = 0.91, N_G2 = 10, icc2 = 0.51, numb_fe1 = 10, numb_fe2 = 10, seed = 12345)) man/boottest.lm.Rd: lm_fit <- lm(proposition_vote ~ treatment + ideology1 + log_income + Q1_immigration, data = fwildclusterboot:::create_data_2(N = 1000, N_G1 = 10, icc1 = 0.91, N_G2 = 10, icc2 = 0.51, numb_fe1 = 10, numb_fe2 = 10, seed = 12345))
    • +
+

DONE. deleted fwildclusterboot::: call from examples

+
+
+

This is a second re-submission

+

Thanks, we see: + Found the following (possibly) invalid URLs: URL: https://www.tidyverse.org/lifecycle/#experimental (moved to https://lifecycle.r-lib.org/articles/stages.html) From: README.md Status: 200 Message: OK

+

Please change http –> https, add trailing slashes, or follow moved content as appropriate.

+

Moved URL https://www.tidyverse.org/lifecycle/#experimental to https://lifecycle.r-lib.org/articles/stages.html

+
+
+

This is a re-submission

+ +
+
+

Test environments

+
    +
  • GitHub Actions: windows-latest (release)
    • +
  • GitHub Actions: GitHub Actions: macos-latest (release)
    • +
  • GitHub Actions: ubuntu-20.04 (release)
    • +
  • GitHub Actions: ubuntu-20.04 (devel)
    • +
  • win-builder: devel
    • +
+
+
+

R CMD check results

+

0 errors | 0 warnings | 1 note

+
    +
  • This is a new release.
    • +
+
+ + + + +
+ + + + + + + + + + + + + + + diff --git a/cran-comments.md b/cran-comments.md index 5f26e627..5f6b1b41 100644 --- a/cran-comments.md +++ b/cran-comments.md @@ -1,3 +1,24 @@ +## Re-submission of 0.7 + +Fixed both issues below: + ++ Found the following (possibly) invalid URLs: + URL: https://www.tandfonline.com/doi/abs/10.1198/jbes.2009.07221%5D + From: README.md + Status: 404 + Message: Not Found + +The closing %5D = "]" should likely be removed. + + + ++ The Description field contains + +https://journals.sagepub.com/doi/abs/10.1177/1536867X19830877?journalCode=stja, + +Please write DOIs as in the Description +field without the rest of the URL. + ## Version 0.7 - checks pass on github actions: windows, mac, ubuntu