nnpiv.diagnostics.relative_wellposedness_from_nested_npiv

nnpiv.diagnostics.relative_wellposedness_from_nested_npiv(A, D, B, C, *, mask_s=None, mask_t=None, **kwargs)[source]

Run the pre-estimation diagnostic for canonical nested NPIV blocks.

Parameters

  • A – Blocks from nested NPIV data generation where A is first-stage endogenous treatment, D first-stage instrument, B second-stage endogenous treatment, and C second-stage instrument. Diagnostic A uses A, C and D (as \(C'\)); B is accepted for interface consistency but not used directly.

  • D – Blocks from nested NPIV data generation where A is first-stage endogenous treatment, D first-stage instrument, B second-stage endogenous treatment, and C second-stage instrument. Diagnostic A uses A, C and D (as \(C'\)); B is accepted for interface consistency but not used directly.

  • B – Blocks from nested NPIV data generation where A is first-stage endogenous treatment, D first-stage instrument, B second-stage endogenous treatment, and C second-stage instrument. Diagnostic A uses A, C and D (as \(C'\)); B is accepted for interface consistency but not used directly.

  • C – Blocks from nested NPIV data generation where A is first-stage endogenous treatment, D first-stage instrument, B second-stage endogenous treatment, and C second-stage instrument. Diagnostic A uses A, C and D (as \(C'\)); B is accepted for interface consistency but not used directly.

  • mask_s (array-like, optional) – Optional subset masks/indices for stage-specific diagnostics.

  • mask_t (array-like, optional) – Optional subset masks/indices for stage-specific diagnostics.

  • **kwargs – Forwarded to relative_wellposedness_diagnostic().

Returns

Output dictionary returned by relative_wellposedness_diagnostic().

Return type

dict

Notes

This wrapper maps D to C_prime and ignores B.