SampleSelectR Pre-Submission

[szimmer]

Submitting Author Name: Stephanie Zimmer
Submitting Author Github Handle: @szimmer
Repository: https://github.com/RTIInternational/SampleSelectR
Submission type: Pre-submission
Language: English

---

Package: SampleSelectR
Title: Randomly select samples for various probability-based methods
Version: 1.0.0
Authors@R: c(
    person("Stephanie", "Zimmer", , "sazimmer@rti.org", role = c("aut", "cre")),
    person("John David", "Bunker, Jr.", role = "aut"),
    person("Thomas", "Burkett", role = "aut"),
    person("Philip", "Lee", role = "aut"),
    person("Haby", "Sow", role = "aut"),
    person(, "RTI International", role = "fnd")
  )
Description: Randomly select samples with SRS, systematic, and various PPS methods. Also includes functionality to select within strata and allocate sample sizes.
License: GPL (>= 3)
Encoding: UTF-8
Roxygen: list(markdown = TRUE)
RoxygenNote: 7.3.2
Depends: 
    R (>= 4.1.0)
Imports: 
    data.table,
    stats,
    rlang,
    tidytable
LazyData: true
URL: https://github.com/RTIInternational/SampleSelectR, https://rtiinternational.github.io/SampleSelectR/
BugReports: https://github.com/RTIInternational/SampleSelectR/issues
Suggests: 
    testthat (>= 3.0.0)
Config/testthat/edition: 3

Scope

• Please indicate which category or categories from our package fit policies or statistical package categories this package falls under. (Please check one or more appropriate boxes below):

  Data Lifecycle Packages

  • data retrieval
  • data extraction
  • data munging
  • data deposition
  • data validation and testing
  • workflow automation
  • version control
  • citation management and bibliometrics
  • scientific software wrappers
  • field and lab reproducibility tools
  • database software bindings
  • geospatial data
  • translation

  Statistical Packages

  • Bayesian and Monte Carlo Routines
  • Dimensionality Reduction, Clustering, and Unsupervised Learning
  • Machine Learning
  • Regression and Supervised Learning
  • Exploratory Data Analysis (EDA) and Summary Statistics
  • Spatial Analyses
  • Time Series Analyses
  • Probability Distributions
  • Survey statistics - random sampling

• Explain how and why the package falls under these categories (briefly, 1-2 sentences). Please note any areas you are unsure of:

This did not fit exactly under any pre-defined category so I added my own. It implements statistical sampling algorithms

• If submitting a statistical package, have you already incorporated documentation of standards into your code via the srr package?

This has not been done

• Who is the target audience and what are scientific applications of this package?

Statisticians who are drawing random samples from sampling frames. Primary audience is those in official statistics and public opinion research.

• Are there other R packages that accomplish the same thing? If so, how does yours differ or meet our criteria for best-in-category?

Some methods of sampling are included in other packages but not the Chromy sequential PPS method. We also wanted to make a unified API for sampling that we can easily add additional methods of sampling in the future.

• (If applicable) Does your package comply with our guidance around Ethics, Data Privacy and Human Subjects Research?

Not applicable

• Any other questions or issues we should be aware of?:

In early discussions of submitting to rOpenSci and discussing whether it is appropriate.

[ropensci-review-bot]

Thanks for your pre-submission to rOpenSci, our editors will reply soon.

[mpadge]

Thanks for inquiring here @szimmer. This looks like a really interesting package, but unfortunately stats categories are literally hard-coded in the form of our series of category-specific standards. We do hope and plan to slowly expand categories, but can not do that in response to individual submissions.

That said, I have had a bit of a look through your package, and do see one possibility, in the Probability Distributions category. In current form, your package simply relies on direct uniform sampling throughout, or sampling in proportion to group sizes. I know nothing of the envisioned or typical area of application of these routines, and imagine your current approach is likely sufficient? Adaptation to accord with our Probability Distributions standards would require modifying the underlying sampling behaviour, to allow sampling from specific distributional forms. Sampling in proportion to group sizes, for example, can be considered a specific form of more general procedures like Poisson sampling where the group size is effectively an expectation value. There are likely many other ways that the sampling routines could be made more general and flexible. BUT as stated above, this may have no relevance to the typical or envisioned area of application of this package, in which case ... Modifying to accord with our standards is likely either infeasible or unproductive. Please let me know what you think.

[mpadge]

@szimmer Do you have any thoughts in response to my suggestions above?

[szimmer]

I'm sorry, on vacation through the end of next week and will review more thoroughly when I return

[szimmer]

I don't think it makes sense to modify what we are doing to fit in the box of probability distributions. There are MANY ways to sample proportional to size and unfortunately, it's sometimes done incorrectly. In this package, we implement a way that is one valid way to do this.

I'm not sure how to go forward with this. It seems that selecting random samples is clearly a statistical use-case but not one that is under the scope of ROpenSci.