Methodology decision aid — methodology_decision

Helps researchers choose between the three methodology modes (reflexive_scaffold, codebook_collaborative, framework_applied) by surfacing the decision-relevant differences. Operates in three modes:

interactive = TRUE (default in interactive R sessions): prompts the researcher with a short series of questions and returns a recommended mode plus reasoning.
interactive = FALSE with criteria supplied: returns a recommendation deterministically based on the supplied criteria.
Neither: prints a comparison of the three modes for the researcher to read; returns NULL invisibly.

Usage

methodology_decision_aid(
  interactive = base::interactive(),
  ta_family = NULL,
  has_apriori_framework = NULL,
  wants_irr = NULL
)

Arguments

interactive: Logical; whether to prompt for input. Defaults to interactive() – TRUE in console sessions, FALSE in scripts.
ta_family: Optional character: one of "reflexive", "codebook", "template", "framework", "content". Used in non-interactive mode.
has_apriori_framework: Optional logical: whether the researcher has a pre-existing theoretical framework to apply.
wants_irr: Optional logical: whether the researcher wants inter-rater reliability statistics in the output.

Value

A list with elements recommended_mode (character), reasoning (character), alternative (character or NA). Invisibly NULL when called in print-only mode.

Details

Per AC3, there is no default methodology mode in validate_config(); this function exists so users can make an informed choice rather than picking arbitrarily.

Examples

# Interactive (prompts in a console, so only run when one is attached):
if (interactive()) {
  result <- methodology_decision_aid()
}

# Non-interactive (deterministic):
result <- methodology_decision_aid(
  interactive = FALSE,
  ta_family = "reflexive",
  has_apriori_framework = FALSE
)

# Print-only comparison:
methodology_decision_aid(interactive = FALSE)
#> 
#> === pakhom methodology modes (multi-mode architecture) ===
#> 
#> Three modes; pick one per run. No default. Methodology is stamped on
#> every output and cannot be silently changed mid-run.
#> 
#> --- reflexive_scaffold ---
#>   Best for: Big-Q reflexive TA per Braun & Clarke 2022
#>   AI role:  Provocateur (questions, counter-narratives, absent voices)
#>             AI never outputs codes/themes as findings.
#>   Required: Reflexive memos at every pause point; positionality at
#>             multiple timepoints; quote provenance verification.
#>   IRR:      Disabled (incoherent for reflexive TA per RTARG 2024).
#> 
#> --- codebook_collaborative ---
#>   Best for: Codebook TA, template TA, applied health-services research
#>   AI role:  Collaborator. AI proposes codes; researcher gates each.
#>             Codebook ships as deliverable.
#>   Required: Researcher-articulated codebook (in researcher's words).
#>   IRR:      Available as quality diagnostic.
#> 
#> --- framework_applied ---
#>   Best for: Abductive coding, theoretical-framework analysis,
#>             framework analysis (Ritchie & Spencer), content analysis
#>             with a-priori categories.
#>   AI role:  Framework-applier with anomaly flagger. AI applies
#>             researcher-supplied framework; flags entries that resist.
#>   Required: theoretical_framework.yaml with constructs and citations;
#>             anomaly resolution log.
#>   IRR:      Available when framework is positivist.
#> 
#> Run methodology_decision_aid() (no args) in an interactive session
#> for a guided 3-question recommendation.
#>