Np-chart¶

Monitor a fixed-size sequence of samples by plotting each sample's count of nonconforming units against binomial center and control limits, separating common-cause fluctuation from special-cause signals.

Version: v1 · 2026-08-30 · History
Domain-specific #: 2398
Origin domain: statistics
Subdomain: attribute control charts
Aliases: Np chart, Number-nonconforming chart

Core Idea¶

An np-chart is an attribute control chart for the number \(D_i\) of nonconforming units in each sample \(i\), used when sample size \(n\) is constant and each inspected unit receives a conforming/nonconforming classification. Under a stable binomial reference with long-run nonconforming probability \(p\), the plotted count has mean \(np\) and variance \(np(1-p)\). The chart does not estimate product quality from one point; it organizes a time-ordered sequence to detect evidence that the process distribution has changed.^[1]

A Phase I baseline estimates \(\bar p\) from appropriately chosen historical samples or uses a justified standard value. The center line is \(n\bar p\), and conventional three-standard-deviation limits are \(n\bar p\pm3\sqrt{n\bar p(1-\bar p)}\), truncated to feasible counts where necessary. New fixed-size samples contribute integer counts. Points beyond limits or nonrandom run patterns trigger investigation under a declared signaling rule. Limits describe expected common-cause variation under the model; specification limits describe customer or engineering acceptability and are separate.^[2]

The np-chart is not a p-chart: both use binomial logic, but the p-chart plots a fraction and can accommodate varying sample sizes, while the np-chart plots a count and requires constant size for fixed limits. It is not a c-chart for defect counts, because one unit can have several defects while nonconforming-unit classification is binary. It is not a hypothesis test repeatedly applied without a monitoring plan, nor proof that a signaled cause is known. Autocorrelation, overdispersion, unequal opportunity, misclassification, and baseline contamination can invalidate nominal limits.^[3]

Structural Signature¶

Process stream. Time-ordered production or service units supply the monitored population.
Fixed subgroup size. Every plotted sample contains the same declared number n of inspected units.
Binary classification rule. Each unit is classified once as conforming or nonconforming under an operational definition.
Sample count. The integer D-i records how many units in subgroup i are nonconforming.
Stable baseline. A justified p or estimated p-bar supplies the common-cause reference.
Center and limits. Binomial mean and dispersion define the expected monitoring band.
Signaling rule. Point, run, or pattern criteria determine when investigation is initiated.
Response loop. A signal routes to cause analysis and documented process action rather than automatic tampering.

What It Is Not¶

Not a p-chart. A p-chart plots proportions and adjusts limits for changing subgroup sizes.
Not a c-chart. A c-chart counts defects or events, potentially several per inspection unit.
Not a specification chart. Control limits describe process behavior; specifications describe acceptability.
Not a defect-severity score. Every unit contributes only the declared binary nonconforming indicator.
Not proof of a special cause. A signal raises an investigation trigger but does not identify its generator.
Not a valid model under arbitrary dependence. Binomial limits require assumptions about opportunity, classification, and dependence.

Scope of Application¶

The abstraction is literal wherever practitioners can identify the same constitutive roles, apply the same boundary tests, and obtain the same kind of output. The following habitats are uses of Np-chart itself, not metaphors based only on resemblance.

Manufacturing inspection. Tracking numbers of rejected or nonconforming units in equal production samples.
Service quality. Monitoring a fixed number of transactions classified by a stable pass/fail rule.
Phase I analysis. Assessing whether historical samples provide a defensible stable baseline.
Phase II monitoring. Comparing future subgroup counts with frozen or carefully governed limits.
Model diagnostics. Checking overdispersion, dependence, and classifier consistency.
Investigation governance. Separating statistical signals from corrective-action authority.

Clarity¶

A clear account of Np-chart must preserve the recognition invariant stated in the Core Idea rather than rely on the title alone. State subgroup size, sampling cadence, unit, binary classification, and whether p is known or estimated. Report the baseline interval and screen it for instability before using it to set limits. Distinguish control limits, warning rules, and specification requirements. Check unequal sample size, overdispersion, autocorrelation, and classification drift before interpreting nominal false-alarm rates. These declarations are not editorial extras: each changes what observations count, which transformations are licensed, and what conclusion can be drawn. A reader should be able to reconstruct the input, the operative rule, the output, and at least one defeater from the account without consulting an implementation or guessing an unstated convention.

Manages Complexity¶

Np-chart manages complexity by replacing a diffuse field of observations or possible operations with a bounded role structure: process stream supplies time-ordered production or service units supply the monitored population.; fixed subgroup size supplies every plotted sample contains the same declared number n of inspected units.; binary classification rule supplies each unit is classified once as conforming or nonconforming under an operational definition.; sample count supplies the integer D-i records how many units in subgroup i are nonconforming.; stable baseline supplies a justified p or estimated p-bar supplies the common-cause reference.. The compression is useful because it localizes disagreement. One can ask whether the input was properly formed, whether a constitutive relation held, whether an alternative explanation defeats the inference, or whether the output was overinterpreted. The same compression can mislead when its discarded detail is exactly what the decision requires. A reference-grade use therefore reports both the invariant retained and the information intentionally lost.

Abstract Reasoning¶

Define the monitored unit and one mutually exclusive conforming/nonconforming rule.
Verify that each plotted subgroup has the same size n.
Select a defensible stable baseline and estimate p-bar if no standard p is supplied.
Compute the center and binomial limits, respecting the feasible integer range from zero to n.
Plot counts in time order and apply a prespecified signal rule.
Investigate data quality and process causes without treating every signal as a confirmed defect mechanism.
Re-estimate limits only through a governed change that preserves historical comparability.
Test the candidate interpretation against the nearest named confusable rather than accepting a shared surface feature.
State the conclusion at the same scope as the source conditions, and retain uncertainty or nonuniqueness where the construct does not remove it.

Knowledge Transfer¶

The strict upward abstraction is Monitoring. Np-Chart instantiates Monitoring because it repeatedly observes a process statistic against an expected band and routes unusual deviations to investigation. Within attribute control charts, the full mechanism transfers literally when the same roles and boundary tests recur. Beyond that domain, only the parent-level skeleton should travel. Reusing the label Np-chart after removing its constitutive vocabulary would hide a change of mechanism behind an analogy. The honest transfer rule is therefore two-stage: recognize the domain-specific pattern first, then lift only the parent relation that remains invariant under a substrate change.

Examples¶

Canonical¶

Twenty baseline subgroups each contain \(n=100\) units. Across them, 80 units are nonconforming, giving \(\bar p=0.04\) and center line \(4\). The conventional standard deviation is \(\sqrt{100(0.04)(0.96)}\), so the lower limit truncates at zero and the upper limit is derived from the three-sigma expression. A later count above that limit signals a distributional departure to investigate; it does not by itself show which machine, material, or inspection change caused it.

Mapped back: input and conventions → constitutive role test → bounded output → explicit interpretation and defeater check.

Applied / In Practice¶

A service team samples 50 completed cases each day and plots the number failing a documented completion standard. When staffing schedules create serial dependence and a new reviewer changes classifications, the nominal chart signals frequently. The team first repairs the measurement and sampling system, then establishes a stable reference. This prevents a chart intended to distinguish causes from merely visualizing a shifting denominator or classifier.

Mapped back: field observation or problem → candidate recognition → confusable and limit checks → appropriately scoped conclusion.

Structural Tensions¶

T1: Concrete counts versus variable opportunity. Counts are intuitive, but fixed limits cease to fit when subgroup size changes. Diagnostic: Verify identical n for every point or use a p-chart with point-specific limits.
T2: Statistical control versus conformance. A stable process may consistently violate specifications. Diagnostic: Display specifications separately and ask two different questions.
T3: Sensitivity versus false alarms. Additional run rules detect smaller shifts but increase signaling frequency. Diagnostic: Prespecify the rule set and evaluate its in-control behavior.
T4: Estimated baseline versus contaminated history. Special causes in Phase I inflate or distort limits. Diagnostic: Audit baseline stability and document any exclusions.
T5: Binomial simplicity versus extra variation. Dependence or heterogeneous probabilities widen the real count distribution. Diagnostic: Compare observed dispersion with \(np(1-p)\) before trusting limits.
T6: Autonomy versus Monitoring. Monitoring supplies observe-detect-respond, while the np-chart adds a fixed-size binomial count mechanism. Diagnostic: Remove count, binary classification, fixed n, and binomial limits and test whether the named chart remains.

Structural–Framed Character¶

The np-chart is structural under its probability and sampling assumptions, while subgroup design, baseline governance, signal rules, and cause response are operationally framed. The five framing criteria point in a consistent direction. Evaluative weight is limited to whether the defining conditions are met, not whether the outcome is desirable. Human practice matters to the extent that experts choose conventions, instruments, or reporting thresholds, but those choices do not make every verdict arbitrary. Institutional history explains the name and standard use; it does not replace the recognition rule. The operative vocabulary travels within the home field and closely adjacent subfields, while transfer farther away requires translation to the parent prime. Thus recognition remains disciplined even where interpretation is defeasible.

Structural Core vs. Domain Accent¶

What is skeletal. Np-Chart instantiates Monitoring because it repeatedly observes a process statistic against an expected band and routes unusual deviations to investigation. This is the part that can be expressed without the candidate's specialist nouns.

What is domain-bound. The irreducible accent is statistical process control, fixed subgroup size, number nonconforming, binomial center and variance, Shewhart limits, time order, and common-versus-special-cause interpretation. Remove those elements and the result is no longer Np-chart; it is only the parent relation or a loose analogy.

Why this does not clear the prime bar. The name does not recur with unchanged diagnostics across three independent domains. What transfers is already represented by prime:monitoring. The candidate remains autonomous because its in-domain recognition rule, failure modes, and consequences are stable, but its vocabulary and interventions do not float free of the home substrate.

Np-Chart instantiates Monitoring because it repeatedly observes a process statistic against an expected band and routes unusual deviations to investigation.

The prospective workspace queue contains one strict upward edge to prime:monitoring. No live DAG mutation is authorized.

Relationships to Other Abstractions¶

Current abstraction Np-chart Domain-specific

Parents (1) — more general patterns this builds on

Np-chart is a kind of Monitoring Prime

Np-Chart instantiates Monitoring because it repeatedly observes a process statistic against an expected band and routes unusual deviations to investigation.

Hierarchy paths (2) — routes to 2 parentless roots

Np-chart → Monitoring → Feedback

Show alternative path (1)

Neighborhood in Abstraction Space¶

Np-chart sits in a sparse region of the domain-specific corpus (90^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Statistical Process Control (14 abstractions)

Nearest neighbors

Natural Process Variation — 0.81
P-chart — 0.81
C-chart — 0.80
X-bar chart — 0.77
Shewhart individuals control chart — 0.77

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With¶

p-chart. Plots fraction nonconforming and supports unequal sample sizes.
c-chart. Plots counts of defects under a Poisson-style opportunity model.
u-chart. Plots defects per unit and adjusts for changing opportunity.
run chart. Shows time order without model-based control limits.
acceptance sampling. Makes lot disposition decisions rather than monitoring process stability over time.
specification limits. External tolerances are not estimated common-cause boundaries.

References¶

[1] NIST/SEMATECH. (2012). e-Handbook of Statistical Methods, §6.3.3.2, Proportions Control Charts. https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc332.htm registry ↩

[2] Montgomery, D. C. (2019). Introduction to Statistical Quality Control, 8^th ed. Wiley. ISBN 978-1-119-39930-8. registry ↩

[3] Ryan, T. P. (2011). Statistical Methods for Quality Improvement, 3^rd ed. Wiley. https://doi.org/10.1002/9781118058114 registry ↩