Quality Control and Assurance in Scanning

Quality control (QC) and quality assurance (QA) are the disciplines that separate a defensible digitization program from a pile of unreliable image files. When an organization scans paper records, it is not simply making pictures of pages; it is producing the authoritative version that will be relied upon for business, legal, and historical purposes, sometimes after the paper originals are destroyed. If the resulting images are illegible, incomplete, mislabeled, or improperly preserved, the organization may lose access to information it is legally required to retain. Rigorous quality programs exist to guarantee that the digital surrogate is a complete, accurate, and trustworthy representation of the source.

It helps to keep two related ideas distinct. Quality assurance is preventive: it shapes the process so that defects are unlikely to occur, through equipment calibration, documented procedures, operator training, and standards adoption. Quality control is detective: it inspects the output to catch defects that slipped through, through visual review, automated validation, and sampling. A mature program uses both, because preventing errors is cheaper than finding them later and finding them at all is cheaper than discovering, years afterward, that a critical record cannot be read.

Image Quality and Capture Standards

The most visible dimension of quality is the image itself. Reviewers assess resolution (typically expressed in pixels per inch or dots per inch), bit depth and color mode, tonal accuracy, and the absence of capture defects. Common defects include skew, cropped or clipped content, missing pages, scanner streaks, dropped detail in shadows or highlights, moire patterns, and over- or under-exposure that renders faint text unreadable. Rather than relying on subjective judgment, leading programs anchor their targets to published benchmarks. The Federal Agencies Digital Guidelines Initiative (FADGI) provides widely used star-rating performance levels and objective measurement methods, allowing institutions to verify capture devices against measurable criteria using calibration targets rather than eyeballing each scan. Tying acceptance thresholds to such standards makes quality reproducible across operators, vendors, and time.

Completeness and Fidelity Verification

A perfect-looking image of the wrong document, or a folder missing half its pages, is still a failure. Completeness checks confirm that every page of every record was captured, in the correct order and orientation, with nothing duplicated or omitted. Practical controls include reconciling page and document counts against batch manifests or preparation logs, verifying that multi-page documents are correctly assembled, and confirming that fragile inserts, double-sided pages, sticky notes, and oversized foldouts were all captured. Fidelity verification asks a deeper question: does the digital surrogate faithfully represent the legal and informational content of the original, including signatures, annotations, stamps, and marginalia that may carry evidentiary weight? Where originals are slated for destruction after scanning, this verification is the last opportunity to catch loss, so it warrants the most stringent review.

Metadata and Indexing Accuracy

Records are only as findable and trustworthy as the metadata attached to them. QC therefore extends to descriptive, administrative, and technical metadata: titles, dates, identifiers, record series, retention codes, and the fields used to retrieve documents. Errors in indexing, transposed identifiers, wrong dates, misfiled series, can render a perfectly captured image effectively lost. Where optical character recognition (OCR) is applied to make text searchable, programs should assess OCR accuracy against the source and account for low-confidence results on degraded or handwritten material. Technical metadata, such as capture date, device, resolution, and file format, supports authenticity and long-term management, and aligns with records principles emphasized in NARA’s records management guidance and in international standards such as ISO 15489.

Sampling, Inspection Methods, and Tolerances

Inspecting every image is sometimes feasible with automated tools but rarely practical for human review at scale, so programs define inspection strategies and acceptable defect tolerances in advance. A common approach combines automated validation of every file, checking file format, integrity, resolution, and presence of required metadata, with statistical sampling for human visual inspection. The program documents sample sizes, what constitutes a critical versus minor defect, and the rejection threshold that sends a batch back for rescanning. Defining tolerances before production begins prevents inconsistent, ad hoc judgments and gives vendors clear, contractual acceptance criteria. Critically, results should be recorded: inspection logs, defect rates, and corrective actions form the audit trail that demonstrates the program was conducted diligently, which matters if the records’ reliability is ever challenged.

File Format, Integrity, and Preservation

Quality does not end at acceptance; a record that degrades or becomes unreadable over time has failed just as surely as one scanned poorly. Sound programs select sustainable, well-documented file formats appropriate to the material and retention period, often a high-quality preservation master plus access derivatives, and validate that files conform to their declared format. Fixity checks, using checksums or hashes generated at capture and re-verified over time, detect silent corruption and confirm that files have not changed during storage, migration, or transfer. These practices connect digitization quality to the broader discipline of digital preservation described by the Library of Congress, ensuring that today’s investment in capture quality remains accessible for the full retention period.

Governance, Standards, and Documentation

Underpinning all of the above is governance: written specifications, roles and responsibilities, and the chosen standards baseline. Because requirements evolve, programs should track the current guidance landscape rather than relying on superseded benchmarks. Notably, NARA revoked its endorsement of the DoD 5015.2 standard in 2022 in favor of the Universal Electronic Records Management (ERM) Requirements developed through the Federal Electronic Records Modernization Initiative (FERMI), a reminder that authoritative requirements shift and that programs must periodically revalidate their standards baseline. Documenting the specification, the QA/QC procedures, the acceptance criteria, and the results creates the evidentiary record that the digitized files are trustworthy. For organizations building or refining such a program, the digitization and imaging hub gathers related guidance on capture, conversion, and the lifecycle management of digital records.