Paternity Test Accuracy: What 99.9% Probability Actually Means

Medical & affiliate disclosure: CTX Stat provides educational laboratory information and is not a substitute for professional medical advice, diagnosis, or treatment. CTX Stat may receive compensation from some outbound provider links when an affiliate program is active; provider comparisons and rankings are based on consumer fit, effective cost, access, policies, and reputation—not commission.

Paternity test accuracy is usually very high when a qualified laboratory tests the correct people’s samples, but the number on the report is not a simple “percent sure” in everyday language. A result such as 99.9% probability of paternity means the tested man’s DNA profile is far more consistent with being the biological father than a random, unrelated man from the relevant reference population. It does not mean the lab has compared him with every possible man, and it does not make the result legally usable unless the test followed the right chain-of-custody process.

What “99.9% probability of paternity” actually means

Most modern paternity tests compare DNA markers from the child with DNA markers from the alleged father, and often the mother when her sample is available. The question is not “Do these two people share DNA?” because many people share common genetic markers. The question is more specific: How much more likely are these DNA results if the tested man is the biological father than if a random, unrelated man is the biological father?

That comparison is expressed as a paternity index at each marker. The lab then combines the marker-by-marker values into a combined paternity index, or CPI. The International Society for Forensic Genetics describes paternity testing statistics as a likelihood-ratio problem, and AABB offers relationship-testing statistical training around likelihood ratios, paternity index, combined likelihood ratio, and probability of relationship. See the ISFG Paternity Testing Commission publications and AABB relationship-testing calculation training for the professional framework behind these reports.

In the common simplified formula, when the prior probability is assumed to be 0.5, the probability of paternity is:

Probability of paternity = CPI ÷ (CPI + 1)

That means a probability of 99.9% corresponds to a CPI of about 999. A probability of 99.99% corresponds to a CPI of about 9,999. A probability of 99.999% corresponds to a CPI of about 99,999. Each extra “9” represents a large increase in statistical weight.

For a consumer reading a report, the practical meaning is: the DNA evidence strongly supports paternity compared with the alternative that a random unrelated man is the father. It is not a statement that “there is a 0.1% chance the lab made a mistake.” Lab error, sample mix-up, close-relative scenarios, and wrong-person collection are separate issues from the CPI calculation.

Paternity result language: what the common terms mean

Report term	Plain-language meaning	What to check
Probability of paternity	The percentage calculated from the CPI and the lab’s assumptions, often reported as 99.9%, 99.99%, or higher for an inclusion.	Look for the CPI and whether the report explains the assumed prior probability.
Paternity index (PI)	The statistical weight from one DNA marker.	One marker alone is not the full answer; the combined result matters.
Combined paternity index (CPI)	The combined weight of evidence across all tested markers.	A higher CPI generally supports a stronger inclusion.
Not excluded	The tested man’s DNA is consistent with being the biological father.	Read this with the probability of paternity, not by itself.
Excluded	The tested man has genetic mismatches with the child that rule him out under the lab’s criteria.	Confirm the samples were correctly collected and identified, especially if the result is unexpected.
Inconclusive	The lab could not issue a strong inclusion or exclusion based on available samples and data.	Ask whether adding the mother, another child, or additional markers could help.

How DNA paternity testing works

For a standard postnatal paternity test, the usual specimen is a buccal swab, which collects cheek cells from the inside of the mouth. The lab extracts DNA and examines multiple genetic markers, commonly short tandem repeats (STRs). STRs are useful because they vary from person to person and are inherited in predictable patterns from biological parents. NIST describes STR markers as the primary loci used in modern forensic DNA testing, and its STR resources are widely used in human identification science; see the NIST STRBase resource.

A child inherits one copy of each autosomal marker from the mother and one from the biological father. If the mother is tested, the lab can often identify which child allele came from her and which allele must have come from the biological father. If the alleged father has the expected paternal allele across many markers, the CPI rises. If he lacks the expected paternal allele at multiple markers, the result may support exclusion.

Many labs can test the alleged father and child without the mother. This is sometimes called a “duo” test. A trio test, involving the mother, child, and alleged father, is usually statistically cleaner because the mother’s DNA helps identify the paternal contribution. Some providers explicitly recommend including the mother when possible because it can reduce inconclusive results and help interpret possible mutations.

Why a paternity test is usually not reported as 100%

Accredited paternity laboratories generally avoid saying paternity is 100% proven because the calculation is based on a comparison against a population model, not a comparison against every person who has ever lived. AABB relationship-testing materials also emphasize probability and statistical calculation rather than absolute certainty; AABB’s current Standards for Relationship Testing Laboratories outline requirements for parentage and complex relationship testing laboratories.

There are several reasons for this cautious wording:

• Population allele frequencies are estimates. The CPI uses reference population data. Good labs apply accepted statistical methods, but the calculation still rests on population genetics.
• Close relatives can share more DNA than unrelated men. If the alleged father’s brother, father, son, or another close male relative could be the biological father, the lab needs to know. A standard probability may assume the alternative is a random unrelated man, which is not the same scenario.
• Mutations happen. A child may have a DNA marker that differs from the alleged father because of a real biological mutation rather than non-paternity.
• Testing depends on the samples provided. The best laboratory calculation is only about the specimens actually received and tested.

Exclusions: when the tested man is ruled out

An exclusion means the tested man’s DNA is inconsistent with being the biological father. In a straightforward case with properly identified samples, multiple genetic mismatches across tested markers provide strong evidence that the tested man is not the biological father.

However, a single mismatch does not always equal non-paternity. STR mutations can occur when DNA is passed from parent to child. Professional paternity-statistics recommendations specifically address possible mutation and null allele scenarios, including how the evidence should be evaluated when an apparent inconsistency appears. The ISFG recommendations on biostatistics in paternity testing discuss likelihood-ratio evaluation, possible mutation, population genetics, and non-paternity reporting.

In practical terms, labs generally look for a pattern. One inconsistency may prompt mutation analysis, additional markers, or a request for the mother’s sample. Several independent mismatches are much harder to explain by mutation and usually support exclusion.

How mutations affect paternity test accuracy

A mutation is a small genetic change that can occur between parent and child. In paternity testing, a mutation may make one marker appear inconsistent even when the tested man is the biological father. This is one reason labs do not base the conclusion on a single marker.

When a mutation is suspected, a qualified lab may:

• review the raw data and repeat testing if needed;
• test additional DNA markers to increase statistical power;
• request the mother’s sample if she was not included;
• apply mutation-aware statistical calculations;
• explain the inconsistency in the final report.

Mutation handling is one of the places where lab quality matters. A low-quality report may simply show a surprising mismatch without enough explanation. A better report or lab consultation should clarify whether the result is an exclusion, a mutation-supported inclusion, or an inconclusive case needing more testing.

Sample quality: why collection still matters

Most postnatal paternity tests use cheek swabs because they are painless, noninvasive, and suitable for infants, children, and adults. Labcorp DNA and HomeDNA both describe cheek-swab collection for paternity testing, and DDC notes that cheek swabs are the standard for at-home paternity testing. See Labcorp DNA’s legal vs. at-home testing comparison, HomeDNA’s paternity kit instructions, and DDC’s home paternity testing information.

Good DNA samples usually come from firm swabbing of the inside cheek, not from a large amount of saliva. Common collection problems include eating or drinking immediately before collection, touching swab tips, putting swabs in the wrong envelopes, collecting from the wrong person, or sending a damp swab in a sealed plastic bag where bacterial growth can interfere with analysis.

If a sample does not yield enough usable DNA, a reputable lab should request a recollection rather than forcing a questionable answer. A delayed report is frustrating, but it is better than a result based on poor-quality material.

Home vs. legal paternity testing: accuracy is not the only issue

Many people assume “legal” means the DNA analysis is more accurate. Usually, the key difference is collection and documentation, not necessarily the lab method. A home paternity test may use the same or similar laboratory analysis, but samples are collected privately. A legal paternity test requires identity verification, witnessed collection, documentation, and chain of custody.

Feature	At-home paternity test	Legal paternity test
Best for	Personal knowledge, private family questions, early information before deciding next steps.	Court, custody, child support, immigration, birth certificate changes, estate matters, or official records.
Collection	Self-collected cheek swabs at home.	Collected or witnessed by an approved professional with identity checks.
Accuracy of lab analysis	Can be highly accurate for the samples provided when performed by a qualified lab.	Can be highly accurate and also tied to verified identities.
Legal use	Usually not court-admissible.	Designed for legal or official use.
Main limitation	The lab may not be able to verify whose sample was submitted.	Costs more and requires appointments or approved collection.

Labcorp DNA states that legal testing involves cheek-swab collection at certified collection facilities and documented chain of custody, while at-home tests are for personal answers. Labcorp’s public DNA testing page lists paternity testing starting at $525 for legal testing and $210 for an at-home kit. HomeDNA states that its at-home paternity kit requires a separate $139 lab fee when registering and that at-home results are not court-admissible. DDC similarly states that home paternity results are for personal knowledge only and legal testing is needed for child support, custody, immigration, Social Security, or adding a name to a birth certificate. See Labcorp DNA testing options, HomeDNA Paternity, and DDC legal paternity testing.

Provider cost and ordering context

Prices change, and the right option depends on whether the result is for personal knowledge or legal use. The table below is not a ranking. It summarizes public provider information that may help you know what to verify before ordering.

Provider	Public paternity offering	Price or fee details to verify	Notable limitations
Labcorp DNA	At-home and legal paternity testing.	Public DNA testing page lists paternity starting at $210 for at-home and $525 for legal testing.	At-home testing is for personal knowledge; legal testing uses chain-of-custody collection.
HomeDNA	Retail at-home paternity kit with cheek swabs.	Public page states an additional $139 lab fee is required when registering; retail kit price may vary by store.	At-home results are not court-admissible; legal testing must be arranged separately.
DDC	Home, legal, prenatal, immigration, and relationship DNA testing.	Public pages emphasize phone ordering/consultation; verify total cost, shipping, collection, and optional expedited service before ordering.	Home results are for personal knowledge; legal uses require supervised collection and chain of custody.
AlphaBiolabs USA	Peace-of-mind and legal DNA testing options.	Public FAQ lists home paternity testing at $89 and other relationship-test prices; verify checkout total and any state-specific requirements.	AlphaBiolabs notes home paternity results are for personal knowledge and not court-admissible; New York has special collection rules.
AffinityDNA USA	Home, legal, immigration, prenatal, and discreet sample options.	Public paternity page lists home paternity from $119, legal from $299, immigration from $499, and prenatal from $1,395.	Verify whether optional express service, unusual samples, or professional collection add cost.

For New York residents, be especially careful. Some paternity providers note that New York has special requirements for paternity testing, including approved collection rather than ordinary self-swabbing. New York State laboratory standards for paternity/identity testing require systems for positive identification, chain of custody, rejection criteria when chain of custody is not acceptable, and written agreements for outside collectors. See the New York State Department of Health specialty requirements and AlphaBiolabs USA’s DNA testing FAQ.

Why lab quality matters more than marketing claims

Many paternity-test ads use similar language: “99.9% accurate,” “99.99% accurate,” or “results in 1–2 days.” Those claims are not enough by themselves. A consumer should also look for signs that the lab can handle the hard cases, not just the straightforward ones.

Important quality factors include:

• Appropriate accreditation and standards. AABB accreditation is widely referenced in U.S. relationship testing. AABB states that relationship-testing accreditation is voluntary in the United States, but many state laws require AABB accreditation for reports used in legal proceedings, and U.S. immigration DNA testing is accepted only when initiated directly with an AABB-accredited relationship-testing facility. See AABB’s relationship-testing accreditation FAQ.
• Validated methods. The lab should use established STR or SNP methods appropriate for the type of paternity test.
• Clear chain of custody for legal tests. Identity verification is essential if the result will be used outside a private family conversation.
• Enough markers for the case. Close-relative scenarios, missing mother samples, or mutations may require additional markers or a more complex analysis.
• Transparent reporting. The report should show the conclusion, CPI, probability of paternity, tested parties, specimen type or collection context, and any limitations.
• Access to support. If a result is unexpected or inconclusive, the provider should be able to explain what the result does and does not mean.

What can reduce confidence in a paternity result?

Modern DNA testing is powerful, but the result should be interpreted in context. Situations that may require extra care include:

• A close relative of the alleged father could be the father. Tell the lab before testing if a brother, father, son, uncle, or other close male relative is a realistic possibility.
• The mother is unavailable. Many tests can proceed without her, but her sample can improve statistical clarity.
• The result is for legal use. Do not rely on a home kit if you may need the result for court, custody, immigration, Social Security, inheritance, or a birth certificate.
• There was a sample mix-up risk. If participants were not collected together, envelopes were not labeled immediately, or someone else handled the kit, consider whether a legal recollection is more appropriate.
• The result is inconclusive. Ask whether additional markers, additional relatives, or a mother’s sample could help.
• The test is prenatal. Noninvasive prenatal paternity testing is a different type of test that uses fetal DNA in the pregnant person’s blood plus the alleged father’s DNA. It has different sample, timing, and interpretation issues than postnatal cheek-swab testing.

Paternity testing is not ancestry or medical genetic testing

A paternity test is a relationship test. It is designed to answer whether the tested man can be the biological father of the tested child. It is not the same as consumer ancestry testing, carrier screening, diagnostic genetic testing, newborn screening, or medical risk testing.

Consumer genetic tests raise privacy and interpretation issues because DNA can reveal information about biological relatives as well as the person tested. MedlinePlus Genetics notes that direct-to-consumer genetic testing may provide ancestry, trait, or health-risk information, but privacy can be compromised if data are shared, sold, or breached. The National Human Genome Research Institute also emphasizes that whether testing can be performed without a donor’s consent can depend on the test, use, provider, and state law. See MedlinePlus Genetics on direct-to-consumer testing and NHGRI on privacy in genomics.

Before ordering, read the provider’s privacy policy, consent language, data retention policy, and rules for testing a minor. If the result may affect custody, support, immigration, inheritance, or family safety, consider legal advice before collecting samples.

How to read a paternity test report

When your report arrives, do not focus only on the big percentage. Review the full report carefully:

1. Confirm the tested parties. Make sure the report matches the people and sample IDs you expected.
2. Identify the conclusion. Does it say excluded, not excluded, included, or inconclusive?
3. Find the CPI. This number shows the strength of the DNA evidence before conversion to a percentage.
4. Read the probability statement. A 99.9% result is strong; 99.99% or 99.999% is stronger, but still not written as 100%.
5. Check for notes. Look for language about mutation, missing mother sample, close relatives, additional testing, or sample limitations.
6. Match the report to your purpose. A home-test report may answer a personal question but still may not be acceptable for court or official agencies.

If a result is emotionally or legally significant, avoid making irreversible decisions based only on a quick glance at the percentage. Contact the laboratory for clarification, and consider legal or genetic-counseling support when the situation is complex.

Practical next steps

• Decide how the result will be used. If there is any chance you need court-admissible proof, choose a legal chain-of-custody test from the beginning.
• Tell the lab about close relatives. If another possible father is related to the tested man, disclose that before testing.
• Include the mother when possible. Her sample may strengthen the analysis and help resolve mutations or borderline results.
• Follow collection instructions exactly. Label envelopes immediately, avoid touching swab tips, and let swabs dry as directed.
• Verify total cost before ordering. Check kit price, lab fee, shipping, professional collection, legal documentation, expedited processing, and state-specific rules.
• Read the full report. The CPI, conclusion, and limitations matter as much as the percentage.

FAQs

Is a 99.9% paternity result the same as 100%?

No. A 99.9% probability of paternity means the DNA evidence strongly supports paternity under the laboratory’s statistical assumptions, but it is not reported as absolute certainty. The percentage is calculated from the CPI, usually with an assumed prior probability of 50%.

What does CPI mean on a paternity test?

CPI stands for combined paternity index. It is the combined statistical weight from the tested DNA markers. A CPI of about 999 corresponds to about 99.9% probability of paternity when a 50% prior probability is used. A CPI of about 9,999 corresponds to about 99.99%.

Can a paternity test be wrong?

A qualified lab’s DNA analysis is highly accurate for the samples tested, but wrong results can occur if the wrong person is swabbed, samples are mislabeled, identity is not verified, a close relative scenario is not disclosed, or the lab uses poor methods. Legal chain-of-custody testing helps control identity-related risk.

Why does the report say “not excluded” instead of “is the father”?

“Not excluded” is standard cautious scientific wording. It means the tested man’s DNA is consistent with being the biological father, and the report should provide the probability of paternity and CPI to show the strength of that conclusion.

Is the mother’s DNA required?

Often no, especially for a standard postnatal alleged father-child test. However, including the mother can make the analysis stronger and may help resolve an inconclusive result or possible mutation.

Are home paternity tests as accurate as legal tests?

The DNA analysis can be similarly accurate, but home tests do not verify identity or maintain legal chain of custody. If you need results for court, custody, child support, immigration, inheritance, Social Security, or a birth certificate, choose a legal paternity test.

What if the result is inconclusive?

Ask the lab what caused the inconclusive result. Common next steps may include adding the mother’s sample, testing additional DNA markers, recollecting a low-quality sample, or testing another alleged father or relative.

Can ancestry DNA results prove paternity?

Consumer ancestry tests can reveal close biological relationships, but they are not a substitute for a dedicated paternity test with appropriate consent, reporting, and chain of custody. They also usually are not accepted for legal paternity purposes.

Source Notes

• International Society for Forensic Genetics: Paternity Testing Commission
• AABB relationship-testing calculation training
• NIST STRBase resource
• AABB Standards for Relationship Testing Laboratories
• ISFG recommendations on biostatistics in paternity testing
• paternity kit instructions
• home paternity testing information
• DDC legal paternity testing
• New York State Department of Health specialty requirements
• AlphaBiolabs USA DNA Testing FAQs
• AABB Relationship Testing Accreditation FAQ
• MedlinePlus Genetics: Direct-to-Consumer Genetic Testing
• NHGRI on privacy in genomics
• AffinityDNA USA Paternity Tests