Practice with national-level exam (FACT, FACT Plus, NET, CUET, etc.) mocks, learn from structured notes, and get your doubts solved in one place.
The analysis-comparison-evaluation-verification methodology imported from latent print examination into modern document work, what each phase actually involves (analysis of the questioned writing in isolation, comparison against exemplars, evaluation of similarities and differences against population frequency, verification by an independent examiner), the linear ACE-V variant that prevents iterative bias, the order-of-evaluation protocols that prevent target-shifting, and how each step is documented to survive a Daubert challenge.
Last updated:
When a questioned document examiner appears in a US federal court and the opposing attorney rises for a Daubert challenge, one of the first questions is: what is the examiner's method, and does it satisfy the criteria for scientific reliability? For the last two decades, the answer that has organised the field's response to that challenge has been ACE-V: Analysis, Comparison, Evaluation, and Verification.
ACE-V did not originate in document examination. It was formalised in latent print examination by David Ashbaugh in the 1990s and adopted systematically by the International Association for Identification and by the Interpol fingerprint working group. Document examiners borrowed it because it addressed a structural problem the discipline shared with fingerprints: the examination was a skilled human judgment, but the steps of that judgment had never been articulated in a way that could be independently reviewed, replicated, or tested for error rates.
ACE-V is not a guarantee of accuracy. A methodology can be well-structured and still produce errors if the underlying data is poor or the examiner is insufficiently trained. But it is the current standard frame for court-defensible handwriting examination across the US, UK, and Australia, and it is increasingly the reference model for the Indian CFSL laboratory network's examination protocols as accreditation requirements align toward NABL T-126 and ISO 17025.
Understanding ACE-V at the level of what actually happens in each phase, where the bias risks sit, and what documentation each phase requires, is essential for anyone who will produce, review, or challenge a handwriting opinion in a professional or legal context.
The discipline that makes ACE-V structurally sound is that the examiner must complete Analysis before touching the exemplars. Breaking this rule is the single most common methodological failure in document examination.
The Analysis phase concerns the questioned writing only. The examiner has not yet looked at the comparison exemplars in any detail. The entire purpose of this sequencing is to prevent the examiner's perception of the questioned writing from being contaminated by what they already know about the known writer: a cognitive bias called target-shifting, where the examiner unconsciously perceives the questioned writing as looking more like the exemplar writer than it actually does.
During Analysis, the examiner documents the following across the questioned writing:
The scope of the sample. How many letters, words, and numerals are present? What proportion of the alphabet is represented? Are there capitalised letters? Are there numerals? The scope determines what comparisons are actually possible: an examination of a two-word signature is a fundamentally different evidentiary exercise from an examination of a three-page handwritten letter.
The condition of the sample. Has the document been photocopied, scanned, transmitted by fax, aged, wet, or folded? Each of these processes degrades specific features: photocopying flattens line quality, fax transmission loses fine pen-pressure detail, aging can obscure ink-flow and starting-stroke features. The examiner identifies which features can be reliably evaluated given the document's condition, and which cannot, before any comparison begins.
The class characteristics. Which copybook system or systems are represented? This places the writing in a population context that will govern the evaluation of similarities and differences in the Evaluation phase.
The individual characteristics. The examiner catalogues and diagrams the individual features present: connecting-stroke forms for each letter, terminal-stroke forms, proportion ratios across zones, pen-lift positions, pressure distribution patterns. These are recorded without reference to the exemplars. The SWGDOC guidelines and ASTM E2290 both specify that this documentation should be in the case notes before the comparison phase begins, so that a reviewer can reconstruct the examiner's pre-comparison assessment independently.
The limitations. Any factor that limits the examiner's ability to evaluate individual features: poor reproduction quality, limited text quantity, evidence of disguise, evidence of illness or age effect, abbreviation of letters. Limitations do not necessarily preclude a conclusion, but they define the boundaries of any conclusion that can be offered.
The Analysis phase is complete when the examiner has a documented, independent characterisation of the questioned writing that can stand on its own before the exemplars are introduced.
Comparison is not a visual matching exercise, it is a systematic feature-by-feature evaluation of how the questioned writing and the exemplar writing relate, both where they look similar and where they differ.
In the Comparison phase, the examiner places the questioned writing alongside the exemplars and systematically evaluates how the individual features identified in Analysis are represented in the exemplar set.
The first step is verifying that the exemplars are adequate for comparison. The Exemplars and Standards topic covers this in detail, but from the ACE-V perspective, the examiner must confirm at the outset of Comparison that the exemplars include the same letter combinations, the same script type (cursive vs print), contemporaneous exemplars from near the date of the questioned document, and sufficient quantity. If the exemplars are inadequate, the Comparison phase cannot yield a reliable result, and the examiner should document this before proceeding.
For each individual feature identified in Analysis, the examiner now evaluates its representation in the exemplar set: Is this feature present in the exemplars? If present, is it consistent in form across the exemplar samples? Does the feature's form in the questioned writing fall within the range of natural variation the examiner can observe across the exemplars, or does it fall outside that range?
This is not a counting exercise. The SWGDOC standard and the ENFSI Best Practice Manual both caution against point-counting systems (which claim to identify a writer based on reaching a threshold number of matching features) because they imply a precision that population-frequency data does not support and because they risk conflating class-characteristic matches with individual-characteristic matches. The examiner's judgment must weigh the significance of each feature against its estimated population rarity and against the quality of the representation of that feature in the available sample.
Differences between the questioned writing and exemplars are as important as similarities. A difference may indicate different authorship; or it may indicate disguise, different writing conditions, different materials, or aging. The examiner evaluates whether each observed difference is within the range of known intra-writer variation for that feature, or whether it represents a genuine discrepancy.
Evaluation is where the examiner translates a body of documented observations into a conclusion, and where the language of the conclusion matters as much as the science behind it.
The Evaluation phase is the examiner's synthesis of everything documented in Analysis and Comparison into a conclusion expressed on the applicable conclusion scale.
The standard conclusion scale in use across most US laboratories is the SWGDOC nine-point scale, ranging from "Identification" (the highest level of certainty for common authorship) through graduated probability levels to "Elimination" (the highest level of certainty for different authorship), with a central "No conclusion" category for cases where the evidence is insufficient. The UK equivalent uses similar gradations under the ENFSI framework, though European laboratories are increasingly moving toward likelihood-ratio expressions anchored to population-frequency data.
An "Identification" conclusion under the SWGDOC scale means that the examiner has found sufficient agreement in the individual characteristics of the questioned writing and the exemplar set to conclude, to a degree of scientific certainty, that they were produced by the same writer. The SWGDOC definition requires that the similarities are of such quality and quantity that the probability of their occurring by coincidence is considered to be beyond the realm of practical possibility. This is not a probabilistic statement tied to a specific numerical frequency; it is a professional judgment informed by the examiner's training and experience.
The Evaluation phase documentation must include: the number and nature of the individual features found to be in agreement, the number and nature of differences observed and the examiner's explanation for each (whether the difference is within natural variation, explicable by condition, or unexplained), and the examiner's conclusion on the applicable scale.
Crucially, the examiner must also record any contextual information that was available during the examination and assess its potential influence on the conclusion. A contextual influence, such as being told that the suspect has already confessed to writing the document, can produce unconscious bias toward an identification conclusion. The linear ACE-V variant and sequential unmasking protocols (discussed in the next section) are designed specifically to manage this risk.
Verification is not a rubber-stamp of the first examiner's conclusion; done correctly, it is a genuinely independent repeat of the ACE process by a second examiner who does not know the first examiner's conclusion.
The Verification phase requires a second, qualified document examiner to independently conduct the Analysis and Comparison phases without access to the first examiner's notes or conclusions, and then to form their own evaluation before comparing results.
The critical requirement is genuine independence. In laboratory practice, this means the verifier receives the questioned document and the exemplar set but not the case report or the first examiner's feature list. If the verifier's conclusion agrees with the first examiner's, this provides a meaningful check. If they disagree, the disagreement is documented, the two examiners discuss the features in contention, and the laboratory's protocol determines how the disagreement is resolved and whether it affects the final opinion.
Verification in the latent print world gained painful relevance after the Mayfield case in 2004, in which three FBI latent print examiners erroneously identified Brandon Mayfield as the source of a fingerprint found at the Madrid train bombing. Each verifier had confirmed the identification, but the verification had not been truly independent: later review established that the verifiers were all aware that the first examiner had made a positive identification before conducting their own review. The effect of that contextual knowledge was to bias each subsequent examiner toward confirmation of the existing opinion.
In document examination, the same risk exists. A verifier who knows that the senior examiner in the laboratory has concluded identification faces a social and cognitive pressure to confirm that conclusion, even when examining the evidence independently. Laboratory protocols that require blind verification (the verifier is not told the first examiner's conclusion before completing their own Analysis and Comparison) are standard in better-accredited laboratories in the UK (per the FSR Codes of Practice), US (per FBI QAS and ASCLD-LAB criteria), and Australia (per NATA accreditation requirements). Indian CFSL protocol is evolving toward this standard under the NABL T-126 revisions.
The circular form of ACE-V, where an examiner cycles between Analysis and Comparison, introduces a confirmation bias risk that the linear variant is designed to eliminate.
The classic description of ACE-V is a sequential, non-iterative process: complete Analysis, then complete Comparison, then form an Evaluation, then send to Verification. In practice, however, examiners sometimes cycle back: a feature noticed during Comparison prompts a second look at the Analysis phase, which is then revised in light of what was seen in the exemplar. This iterative cycling is called "circular ACE-V" in the methodological literature, and it introduces a specific bias risk.
The risk is as follows. If the examiner identifies a feature in the questioned writing during Comparison that matches a feature in the exemplar (and that was not catalogued in Analysis), and then revises the Analysis documentation to include that feature, the feature list in Analysis has been contaminated by knowledge of the exemplar. The Analysis is no longer an independent baseline; it is a list of features that the examiner already knows match the exemplar writer.
The linear ACE-V variant, formalised in the cognitive bias mitigation literature by Dror and colleagues (Dror 2011, Dror and Langenburg 2010) and referenced in the NIST 2020 handwriting report, requires that the Analysis documentation be completed and locked before Comparison begins. If the examiner subsequently notices a feature during Comparison that was not in the original Analysis, the correct action is to document the additional feature clearly as a Comparison-phase observation, not to revise the Analysis section retroactively.
The practical implementation differs slightly across laboratories. The FBI Questioned Documents Unit protocol, the Australian Federal Police document examination SOP, and the UK Forensic Access and Document Evidence Ltd SOPs all implement linear ACE-V with locked pre-comparison feature documentation; exact implementation detail is laboratory-specific. The CFSL examination protocol in India is moving toward this structure as part of the ongoing ISO 17025 alignment process.
A methodology is only as defensible as its documentation, and ACE-V was partly designed to produce case notes that a court can review.
The documentation required at each ACE-V phase is not administrative overhead; it is the primary mechanism by which the methodology is made reviewable, reproducible (in the sense of another examiner being able to reconstruct the reasoning), and challengeable by opposing parties in court.
Analysis documentation should include: scaled photographs or reproductions of the questioned writing with the features annotated; a written feature list identifying each individual characteristic by location (letter, word, line) and by type (connecting stroke, terminal form, proportion, etc.); a record of the condition of the document and any features that could not be evaluated due to condition; and the date and examiner identity.
Comparison documentation should include: the feature-by-feature comparison table showing each individual characteristic from the Analysis list, its presence or absence in the exemplar set, and whether the form in the questioned writing falls within the exemplar's range of natural variation; documentation of any differences and the examiner's assessment of each difference; and reference to the specific exemplar documents used.
Evaluation documentation should include: the conclusion on the applicable scale with the basis stated (the combination of individual characteristics that supports the conclusion, or the absence of sufficient individual characteristics, or the presence of inexplicable differences); identification of limiting factors; and any context disclosure (what case background the examiner was given, and when).
Verification documentation should include: the verifier's identity, their independent feature list from Analysis, their Comparison results, their independent conclusion, and, where the verifier's conclusion differs from the first examiner's, a record of the resolution discussion.
This four-part documentation package is what survives a Daubert hearing. Under the Daubert criteria (testing, peer review, known or potential error rate, general acceptance), the methodology can be defended on the first two grounds; the documentation is what addresses the third ground (error rate) by making the examiner's reasoning visible to a reviewing court. US federal courts since Kumho Tire (1999) have required that the examiner's specific application of the methodology, not just the methodology in the abstract, satisfies Daubert criteria, and the case notes are the primary evidence of that application.
In the UK, following R v. T [2010] EWCA Crim 2439 and subsequent criminal court decisions, forensic science evidence is increasingly required to be accompanied by a clear statement of the methodology, the basis for the conclusion, and the limits of the opinion. The ACE-V documentation package satisfies this requirement when completed correctly. Under the Indian BSA 2023 and IEA 1872 expert evidence framework, a court may examine the basis for an expert's opinion, and comprehensive ACE-V documentation provides that basis in a structured, reviewable form.
| Phase | Primary activity | Bias risk addressed | Key documentation output |
|---|---|---|---|
| Analysis | Characterise and catalogue features of the questioned writing in isolation, before examining exemplars | Prevents target-shifting: examiner cannot unconsciously shape their perception of the questioned writing to match the known writer | Feature list with annotated images; condition notes; scope statement; all locked before Comparison begins |
| Comparison | Feature-by-feature evaluation of questioned writing against exemplar set, including both similarities and differences | Requires explicit documentation of differences, preventing the examiner from ignoring inconvenient discrepancies | Comparison table; difference assessment; exemplar adequacy check |
| Evaluation | Form a conclusion on the applicable scale from the combined weight of the Comparison results | Conclusion scale forces graduated language, preventing overstatement beyond what evidence supports | Conclusion with stated basis; limiting factors; context disclosure |
| Verification | Independent repeat of ACE (blind) by a second qualified examiner | Prevents confirmation bias cascading through the laboratory; catches individual examiner error | Verifier's independent feature list; verifier's conclusion; record of any disagreement and resolution |
What is the primary reason that the Analysis phase of ACE-V must be completed and documented before the examiner views the comparison exemplars in detail?
Test yourself on Questioned Document with free, timed mocks.
Practice Questioned Document questions