Forensic psychology

In short

Forensic psychology is the application of psychological knowledge, assessment and research to questions raised by the legal system — criminal responsibility, fitness to stand trial, credibility of a witness, dangerousness, suicide-versus-homicide, and the behavioural interpretation of a crime scene. It sits between behavioural science and law: the psychologist does not decide guilt but supplies an opinion a court may accept under the expert-evidence provision of the Bharatiya Sakshya Adhiniyam, 2023 (BSA). In India it is examined in B.Sc/M.Sc Forensic Science papers on "Forensic Psychology / Criminal Behaviour", in NFSU and university entrance tests, and in UGC-NET Psychology and Criminology. Employers include the Central and State Forensic Science Laboratories under DFSS (MHA), NFSU Gandhinagar, NIMHANS and IHBAS, State police and CID behavioural-science units, prisons and juvenile justice boards, and the private insurance-fraud and corporate-investigation industry.

Explained

What the subject covers, and what it is not

• Definition: the professional practice and research application of psychology to law, legal process and legal institutions — assessment, treatment, consultation and expert testimony.
• Forensic vs clinical psychology: the clinical psychologist's client is the patient and the goal is treatment; the forensic psychologist's client is usually the court or the investigating agency, the "patient" may have reason to lie, and confidentiality is limited and must be warned about at the start.
• Forensic vs criminal psychology vs criminology: criminal psychology studies why individuals offend; criminology studies crime as a social phenomenon; forensic psychology is the courtroom- and investigation-facing application of both.
• Core referral questions: criminal responsibility (insanity), fitness/competency to stand trial, risk of future violence, mental condition as a mitigating factor at sentencing, child custody and child witness capacity, credibility and suggestibility, disability certification, psychological injury in compensation claims.
• Roles: court-appointed examiner, defence or prosecution expert, police consultant (profiling, interview strategy), prison/probation assessor, researcher, trainer.
• Who may practise clinically in India: the Mental Healthcare Act, 2017 defines a "clinical psychologist" essentially as a person with an M.Phil in Clinical Psychology (or equivalent recognised qualification) registered with the Rehabilitation Council of India under the RCI Act, 1992. A B.Sc/M.Sc Forensic Science graduate may do crime-scene behavioural work, research and lab testing, but not clinical diagnosis or psychotherapy without that registration.

Criminal responsibility, competency and Indian law

• Mens rea link: the criminal law needs a guilty mind; mental disorder is relevant only where it destroys the capacity the law requires, not merely because a diagnosis exists.
• Insanity defence — Section 22, Bharatiya Nyaya Sanhita, 2023 (the provision that replaced Section 84 IPC): nothing is an offence done by a person who, at the time of doing it, by reason of unsoundness of mind is incapable of knowing the nature of the act, or that it is wrong or contrary to law.
• This codifies the McNaughton Rules (England, 1843): cognitive test only — knowledge of nature and wrongfulness. India does not recognise irresistible impulse, the Durham "product of mental disease" test, the ALI substantial-capacity test, or diminished responsibility as a general defence.
• Legal vs medical insanity: the courts repeatedly hold that only *legal* insanity at the moment of the act exonerates; a psychiatric history, treatment records or later illness are evidence from which it may be inferred, not a substitute. The burden of proving the exception lies on the accused (standard: preponderance of probabilities), while the prosecution must still prove the act.
• Behavioural indicators the court weighs: absence of motive, no preparation or concealment, acting in the open, no attempt to escape, immediate confession, bizarre behaviour before and after, documented prior treatment.
• Fitness to stand trial: an accused who cannot understand the charge, the proceedings, or instruct a lawyer is "unfit to make a defence". The Bharatiya Nagarik Suraksha Sanhita, 2023 carries forward the old CrPC Chapter procedure (ss. 328–339 CrPC) — the court must have the accused examined by a medical officer/psychiatrist, may postpone the trial, release on bail with care arrangements or order custody in a mental health establishment, and must resume when fitness is restored.
• Competency is present-tense and functional; responsibility is past-tense and retrospective. The two are separate assessments and must not be merged in a report.
• Mental Healthcare Act, 2017: governs capacity to consent to treatment, advance directives, nominated representatives, admission procedures, and court referrals for determination of mental illness; it also effectively decriminalised attempted suicide by presuming severe stress and directing care.
• Intoxication: BNS treats voluntary intoxication as no defence (involuntary intoxication is dealt with separately) — relevant because intoxication is far commoner than psychosis in Indian homicide files.
• Juvenile preliminary assessment: under the Juvenile Justice (Care and Protection of Children) Act, 2015, for a child aged 16–18 alleged to have committed a *heinous* offence, the Board must conduct a preliminary assessment of mental and physical capacity, ability to understand consequences and circumstances — assisted by a psychologist or psycho-social worker — to decide on trial as an adult. This is an assessment of capacity, not of guilt.
• Sentencing: mental condition is a recognised mitigating factor in the "rarest of rare" framework; the Supreme Court has treated severe post-conviction mental illness as a ground to commute a death sentence (Shatrughan Chauhan, 2014; Accused X v. State of Maharashtra, 2019).

Assessment: methods, tests and their psychometrics

• Multi-method rule: no forensic opinion rests on one source. Combine clinical interview, mental status examination, collateral interviews (family, neighbours, police diary, hospital records), behavioural observation and psychometric tests.
• Reliability (consistency) and validity (measuring the intended construct) must both be stated; forensic reports are cross-examined on them. Types: test–retest, inter-rater (kappa), internal consistency (Cronbach's alpha); content, criterion and construct validity.
• Intelligence/cognitive tests used in India: Binet–Kamat Test, Malin's Intelligence Scale for Indian Children (MISIC), Bhatia's Battery of Performance Tests of Intelligence, Seguin Form Board, Vineland Social Maturity Scale (Indian adaptation), WAIS-IV India where available.
• Personality and psychopathology: MMPI-2 (with validity scales L, F, K), 16PF, NEO-PI; projective tests — Rorschach Ink Blot, Thematic Apperception Test (Indian adaptation), Sentence Completion, Draw-a-Person. Projectives have weak courtroom standing because of low inter-rater reliability.
• Structured diagnostic interviews: SCID, IPDE; neuropsychological batteries (NIMHANS battery) for head injury, epilepsy, dementia claims.
• Malingering (feigning) and defensiveness: forensic settings have strong incentives to fake. Tools — TOMM, Rey 15-item test, M-FAST, SIRS-2, MMPI validity scales. Indicators: approximate answers, over-endorsement of rare symptoms, worse-than-chance performance on forced-choice tests, inconsistency between claimed deficit and observed function.
• Risk assessment, three generations: (1) unstructured clinical judgement — poor accuracy; (2) actuarial instruments (VRAG, Static-99R) — fixed items, statistical score; (3) structured professional judgement (HCR-20 V3 for violence, SAVRY for adolescents, PCL-R for psychopathy, 20 items scored 0/1/2, maximum 40).
• Honest limitation for India: most of these instruments are Western-normed, licence-restricted, and only partly validated on Indian samples; a report should say so rather than quote a foreign cut-off as fact.
• Psychological autopsy / equivocal death analysis: reconstruct the deceased's state of mind from diaries, messages, financial and medical records and informant interviews to help distinguish suicide, accident and homicide; it supports, never replaces, the post-mortem and scene findings. Used in dowry-death, custodial-death and farmer-suicide inquiries.

Criminal profiling and behavioural crime-scene analysis

• Definition: inferring probable offender characteristics (age band, likely occupation, social competence, familiarity with the area, prior record, likely post-offence behaviour) from crime-scene behaviour, victimology and forensic findings.
• Three logics: *deductive* (reasoning from this case's evidence), *inductive* (base rates from solved-case databases), and *geographic* (spatial pattern of offences).
• Classical FBI typology: organised offender — planned, weapon brought and removed, restraints used, body hidden, victim a stranger but targeted; disorganised — spontaneous, weapon of opportunity, body left at scene, depersonalisation, minimal concealment. Treat it as a continuum; the strict dichotomy has poor empirical support.
• Key concepts: *modus operandi* (learned, changes with experience) versus signature (psychological need, stable); staging (deliberate alteration of the scene to mislead); personation/undoing; escalation; linkage analysis (connecting series by behavioural consistency and distinctiveness).
• Geographic profiling: the *circle hypothesis* (marauder with an anchor point inside the offence circle versus commuter who travels out), distance-decay, buffer zone around the residence; implemented in software such as Rigel/Dragnet. Useful for prioritising a suspect list, not for identification.
• Investigative uses: suspect prioritisation, interview strategy, media strategy, risk to further victims, decision on linking cases across jurisdictions.
• Limitations: profiling is probabilistic and not evidence of identity; validation studies show professional profilers only modestly better than informed laypersons on some dimensions; Barnum effect makes vague profiles feel accurate; it is never admissible as proof that a particular accused committed the act. Indian courts have not accepted a profile as substantive evidence.

Detection of deception, and the Selvi line of law

• Polygraph: records respiration, electrodermal activity (skin conductance) and cardiovascular measures; it detects arousal, not lies. Question formats — Relevant/Irrelevant, Comparison (Control) Question Test, Guilty Knowledge/Concealed Information Test (strongest logic, needs facts only the offender could know). Countermeasures, anxious innocents and the base-rate problem all push false positives up.
• Narcoanalysis: slow intravenous sodium pentothal/thiopentone to a hypnotic-sedative stage so inhibition falls; the material produced is suggestible, mixes fantasy with fact, and is medically risky.
• BEOS / BEAP (Brain Electrical Oscillation Signature Profiling): multi-channel EEG recording of electrophysiological response to probes describing crime details, to infer experiential knowledge; an Indian development (C. R. Mukundan), used at FSLs in Gujarat and Maharashtra. It is not internationally peer-validated to evidentiary standard. P300-based Guilty Knowledge Test is the related international technique.
• Selvi v. State of Karnataka (2010): the Supreme Court held that compulsory administration of narcoanalysis, polygraph and BEAP violates Article 20(3) (no compulsion to be a witness against oneself) and Article 21 (personal liberty, mental privacy, substantive due process); it amounts to cruel, inhuman treatment. A test may be conducted only with free and informed consent following the NHRC Guidelines (2000) — consent before a Magistrate, lawyer present, explanation of consequences, recording at an independent agency. Even then the statement made during the test is not evidence; only material subsequently and independently discovered may be admissible on the s. 27 Indian Evidence Act principle (now carried into the BSA).
• Verbal and behavioural approaches: Statement Validity Assessment with CBCA criteria (quantity of detail, contextual embedding, unexpected complications, self-deprecation, unstructured production) plus a validity checklist; Reality Monitoring; Scientific Content Analysis (SCAN) — weakly validated; cognitive load approaches (unanticipated questions, reverse-order recall, strategic use of evidence) outperform the classical "nervous body language" cues, which research shows are near chance.

Interviewing, memory, testimony and false confession

• Memory is reconstructive, not a recording: encoding is affected by weapon focus, stress, exposure duration, cross-race effect; storage by time and post-event information (Loftus's misinformation effect); retrieval by question wording and social pressure.
• Cognitive Interview (CI/ECI): rapport building, free narrative, context reinstatement, report everything, change of order, change of perspective, witness-compatible questioning, closure. Yields substantially more correct detail without large increases in error.
• PEACE model (Planning and preparation, Engage and explain, Account, Closure, Evaluate) — non-accusatorial investigative interviewing; contrast with the accusatorial Reid technique, whose confrontation, minimisation and false-evidence steps are linked to false confessions.
• Question types, worst to best: leading → suggestive/forced choice → closed → open → free recall. Avoid "Did he have a knife?"; ask "Tell me about his hands."
• False confession types: voluntary, coerced-compliant (to escape the situation), coerced-internalised (comes to believe it). Risk factors: youth, intellectual disability, mental illness, sleep deprivation, long interrogation, suggestibility (measured by the Gudjonsson Suggestibility Scale). Legally, confessions to police are inadmissible in India (ss. 25–26 Indian Evidence Act, carried into BSA s. 23), a safeguard directly aimed at this risk.
• Identification parades: conduct through a Magistrate; fillers must resemble the suspect in description; sequential presentation and unbiased pre-parade instruction ("the person may or may not be present") reduce false picks; double-blind administration prevents cueing. Confidence is malleable and a poor index of accuracy after feedback.
• Child and vulnerable witnesses: POCSO Act, 2012 — child-friendly recording, support person, no aggressive cross-examination, in-camera and video-conferencing options; developmentally appropriate language, no repeated suggestive questioning, use of the NICHD interview protocol. Avoid anatomical dolls as a stand-alone diagnostic.

Report writing, expert testimony and ethics

• Admissibility: the opinion of a person specially skilled in a science is a relevant fact under the expert-evidence provision of the BSA, 2023 (s. 39, replacing s. 45 of the Indian Evidence Act, 1872). Expert opinion is *advisory* — it is weak evidence standing alone and needs corroboration; the court is not bound by it.
• Report structure: referral question and authority; date, place, duration and conditions of examination; limits-of-confidentiality warning given; sources of information listed; behavioural observations; test names, dates and validity-scale results; reasoning; opinion addressing only the question asked; limitations and alternative explanations.
• Ultimate-issue problem: write "his capacity to know the nature of the act appears to have been impaired" with the data, not "he is innocent under Section 22" — the legal conclusion belongs to the judge.
• Ethical duties: informed consent or statutory authority, competence within one's qualification and registration, no dual roles (do not treat and evaluate the same person), impartiality regardless of who pays, test security, data retention, and disclosure of base rates and error rates.
• Cross-examination attacks to expect: normative sample not Indian, no inter-rater check, malingering not tested, reliance on a projective test, retrospective opinion without contemporaneous records, hearsay-only history.

Key points

• Forensic psychology — psychology applied to legal questions; the court, not the patient, is the client.
• Section 22 BNS, 2023 — the insanity exception that replaced s. 84 IPC; cognitive incapacity at the time of the act.
• McNaughton Rules — knowledge of nature and wrongfulness; India follows this and rejects irresistible impulse and diminished responsibility.
• Legal vs medical insanity — a diagnosis is not a defence; the state of mind at the moment of the act decides.
• Burden — on the accused on a preponderance of probabilities; the act itself must still be proved by the prosecution.
• Competency to stand trial — present ability to understand proceedings and instruct counsel; BNSS carries forward the CrPC unsoundness-of-mind procedure.
• JJ Act, 2015 preliminary assessment — psychologist-assisted capacity assessment for 16–18-year-olds in heinous offences; not an inquiry into guilt.
• Selvi (2010) — involuntary polygraph, narcoanalysis and BEAP are unconstitutional (Arts. 20(3) and 21); consent before a Magistrate per NHRC guidelines, and the statement itself is not evidence.
• Polygraph — measures arousal, not deception; the Concealed Information Test has the soundest logic; low base rates destroy predictive value.
• BEOS — Indian EEG-based test for experiential knowledge; not internationally validated as evidence.
• MO vs signature — MO is learned and changes; signature expresses a stable psychological need and links a series.
• Organised/disorganised typology — a useful continuum, an empirically weak dichotomy; a profile prioritises suspects, it never identifies one.
• Circle hypothesis — marauder offends around an anchor point; commuter travels away from it.
• Cognitive Interview — context reinstatement, report everything, change order and perspective; more correct detail than standard interviewing.
• False confession — voluntary, coerced-compliant, coerced-internalised; youth, disability and long interrogation are the main risk factors.
• Malingering — must be explicitly tested (TOMM, MMPI validity scales, SIRS-2); below-chance forced-choice performance is a strong indicator.
• Risk assessment — actuarial and structured professional judgement (HCR-20, PCL-R, SAVRY) beat unstructured clinical judgement, but Indian norms are limited.
• Expert opinion (BSA s. 39) — relevant but advisory and corroboration-dependent; never state the ultimate legal conclusion.

Formulas and examples

Example 1 — why a "95% accurate" polygraph convicts the innocent. 1,000 employees screened for a theft; base rate of actual culprits 20% (200 guilty, 800 innocent). Test sensitivity 0.85, specificity 0.70.

• TP = 0.85 × 200 = 170; FN = 30. FP = 0.30 × 800 = 240; TN = 560.
• PPV = 170/(170 + 240) = 170/410 = 41.5% — a "deception indicated" result is wrong more often than right.
• NPV = 560/(560 + 30) = 94.9%; overall accuracy = (170 + 560)/1,000 = 73%.
• By Bayes: prior odds = 200/800 = 0.25; LR+ = 0.85/0.30 = 2.83; posterior odds = 0.708; P = 0.708/1.708 = 41.5% ✓.
• Conclusion for an exam answer: the test's value is driven by the base rate, which is why screening use is condemned and the Concealed Information Test is preferred.

Example 2 — IQ and the intellectual-disability cut-off. A 12.0-year-old boy scores a mental age of 9.0 years on the Binet–Kamat Test.

• Ratio IQ = (9.0/12.0) × 100 = 75.
• As a deviation score, z = (75 − 100)/15 = −1.67 → about the 4.7th percentile.
• Now the measurement error: with SD = 15 and reliability r_xx = 0.91, SEM = 15 × √(1 − 0.91) = 15 × 0.30 = 4.5. 95% CI = 75 ± 1.96 × 4.5 = 66 to 84.
• Interpretation: the true score straddles the conventional 70 cut-off, so the report must give the band and the functional (adaptive) assessment, not a bare "IQ 75".

Example 3 — inter-rater reliability of two profilers. 100 solved homicides classified organised/disorganised. Both said organised in 45, both disorganised in 30; A organised–B disorganised in 15; A disorganised–B organised in 10.

• Po = (45 + 30)/100 = 0.75.
• Marginals: A organised = 60/100 = 0.60; B organised = 55/100 = 0.55.
• Pe = (0.60 × 0.55) + (0.40 × 0.45) = 0.33 + 0.18 = 0.51.
• κ = (0.75 − 0.51)/(1 − 0.51) = 0.24/0.49 = 0.49 — only *moderate* agreement despite 75% raw agreement; this is the standard criticism of the typology.

Example 4 — fairness of an identification parade. Nominal size 6 (suspect + 5 fillers). 30 mock witnesses who have only the eyewitness's verbal description pick: suspect 12, filler A 7, B 5, C 4, D 2, E 0.

• Functional size = 30/12 = 2.5 against a nominal 6 — the parade is effectively a 2–3 person lineup and is biased, because the suspect alone matches the description.
• Expected chance picks if fair = 30/6 = 5; the suspect drew 12, more than double.
• Remedy: choose fillers from the description, not from availability, and use sequential double-blind presentation.

Example 5 — reading a risk-assessment tool's output. 200 released offenders followed for 3 years; 40 reoffended (base rate 20%). The tool flagged 60 as high risk, of whom 30 reoffended.

• Sensitivity = 30/40 = 75%; FP = 30; TN = 160 − 30 = 130 → specificity = 130/160 = 81.3%.
• PPV = 30/60 = 50%; low-risk reoffence rate = 10/140 = 7.1%.
• RR = 0.50/0.0714 = 7.0 — a seven-fold difference in group rates, yet half the "high risk" group did not reoffend, so the opinion must be worded as group probability, never as a prediction about the individual.

Where it is used

• Forensic Science Laboratories (DFSS/MHA, State FSLs): polygraph, BEOS and narcoanalysis divisions in Gujarat, Maharashtra, Karnataka, Delhi and other States; scientific assistant and scientific officer posts.
• Police and CID behavioural units: suspect prioritisation in serial offences, interview strategy, negotiation support, psychological autopsy in custodial and dowry deaths.
• Courts, prisons and juvenile justice: court-referred capacity and fitness assessments at NIMHANS and IHBAS, JJ Board preliminary assessments, prisoner risk and rehabilitation profiling, probation reports.
• Child protection: POCSO Special Courts, One Stop Centres, Childline and CWC teams using child-forensic interviewing protocols.
• Academia, training and research: NFSU Gandhinagar, LNJN-NICFS Delhi, university forensic psychology departments; UGC-NET and PhD research in profiling and eyewitness memory.
• Private sector: insurance-fraud and disability-claim evaluation, corporate internal investigations and integrity interviewing, HR background screening, expert consultancy in defence briefs.

Interview questions

• Q: What is the legal test of insanity in India? A: Section 22 of the Bharatiya Nyaya Sanhita, 2023 (formerly s. 84 IPC): incapacity, by reason of unsoundness of mind at the time of the act, to know its nature or that it is wrong or contrary to law — the McNaughton cognitive test.
• Q: Difference between legal and medical insanity? A: Medical insanity is any psychiatric diagnosis; legal insanity is the specific incapacity at the moment of the offence. Only legal insanity exonerates, and the burden of establishing it lies on the accused.
• Q: What did Selvi v. State of Karnataka decide? A: In 2010 the Supreme Court held involuntary narcoanalysis, polygraph and BEAP tests violate Article 20(3) and Article 21; such tests need free informed consent recorded before a Magistrate under the NHRC 2000 guidelines, and the statements produced are not admissible, though independently discovered material may be.
• Q: Does a polygraph detect lies? A: No — it records physiological arousal, which can be produced by fear, anger or the significance of a question in an innocent person; interpretation is inferential and highly base-rate dependent.
• Q: Which deception-detection question format is methodologically strongest and why? A: The Concealed Information (Guilty Knowledge) Test, because it uses details only the offender could know, so a strong response to the correct detail has a specific rather than generic explanation.
• Q: Difference between modus operandi and signature? A: MO is the functional method of committing the crime and evolves with experience; the signature is the behaviour that satisfies a psychological need, is unnecessary to complete the offence and stays stable across a series.
• Q: Can a criminal profile be used to convict? A: No. A profile is a probabilistic investigative aid that narrows a suspect pool; it is not evidence of identity and is not admitted as substantive proof in Indian courts.
• Q: What is the cognitive interview and why is it preferred? A: A memory-based interviewing method using rapport, free narrative, context reinstatement, "report everything", change of order and change of perspective; it elicits far more correct detail than standard questioning without greatly raising errors.
• Q: Name the types of false confession and one major risk factor. A: Voluntary, coerced-compliant and coerced-internalised; principal risk factors are youth, intellectual disability, high suggestibility and prolonged confrontational interrogation.
• Q: How would you test whether a claimant is malingering memory loss? A: Use forced-choice symptom-validity tests such as the TOMM along with MMPI-2 validity scales, and look for below-chance performance, symptom over-endorsement and inconsistency between claimed and observed functioning.
• Q: What is a psychological autopsy and where is it used in India? A: A retrospective reconstruction of the deceased's mental state from records and informant interviews, used to help classify equivocal deaths — suicide, accident or homicide — in dowry-death, custodial-death and farmer-suicide inquiries.
• Q: Under which provision is your expert report relevant, and what weight does it carry? A: Under the expert-opinion provision of the Bharatiya Sakshya Adhiniyam, 2023 (s. 39, replacing s. 45 of the Indian Evidence Act); it is relevant but advisory, generally requires corroboration, and the court is not bound by it.

Common mistakes

• Equating a psychiatric diagnosis with the insanity defence, and ignoring that the test applies to the moment of the act.
• Confusing criminal responsibility with fitness to stand trial, or writing one assessment to answer both questions.
• Saying polygraph, narcoanalysis or BEOS results are "admissible in Indian courts" — the test needs consent and the statement is not evidence.
• Treating the organised/disorganised dichotomy, or a profile, as proof about a named accused; and quoting Western actuarial cut-offs as if Indian norms existed.
• Omitting malingering testing and the limits-of-confidentiality warning, then being dismantled on cross-examination.
• Reporting a bare IQ or test score without the standard error, confidence band and adaptive-functioning data.
• Using leading questions with children or repeating suggestive questions, which contaminates the account and can make genuine abuse unprovable.
• Stating the ultimate legal conclusion ("he is not guilty under Section 22") instead of the psychological findings and reasoning.

Related topics

Criminal behaviour and criminology · Victimology and crime-scene reconstruction · Forensic medicine and medico-legal autopsy · Questioned documents and handwriting examination · Indian criminal law: BNS, BNSS and BSA, 2023 · Juvenile justice and POCSO procedure