Filicide: Fifty Years of Trials, Verdicts & Mental-Illness Defenses

Filicide is a relationship category, not one explanation

Maternal filicide concerns a mother; paternal filicide concerns a father. Neonaticide generally concerns a newborn’s first day of life. Infanticide concerns infancy, although its legal meaning varies by jurisdiction. Familicide describes killings involving multiple family members, often a partner and children. These terms can overlap. [1]

Forensic literature describes several proposed pathways: a parent’s distorted belief that death will benefit the child; an acute psychotic episode; rejection of an unwanted child; fatal maltreatment; or revenge against a partner. Resnick’s classic categories organize research; they are not diagnoses or findings about a particular defendant. “Altruistic” in this literature describes the offender’s perceived rationale, not an endorsement that the act helped the child. [1]

The broader Exhibit J hub examines motive, psychiatric evidence, maternal and paternal patterns, and criminal responsibility. The trial sample here concerns mothers only. It cannot support conclusions about fathers or all forms of filicide.

Postpartum illness requires specific evidence

Postpartum psychosis is a severe psychiatric emergency that can involve hallucinations, delusions, marked mood changes and confusion. It is distinct from ordinary postpartum adjustment and postpartum depression. NHS guidance emphasizes urgent assessment and treatment. A postpartum date, a depression diagnosis or a report of unusual behavior is not sufficient to determine that a person was psychotic during a killing. [2]

A diagnosis and a legal defense answer different questions. Diagnosis concerns illness; criminal responsibility concerns the jurisdiction’s legal test at the time of the alleged offense. Competency concerns ability to participate in legal proceedings. Neither a competency finding nor a psychiatric hospitalization automatically resolves criminal responsibility.

How this comparison was built

The original research used a 1976–September 2, 2026 search window. The identified historical trial entries shown here run from 1984 through 2024; later rulings are recorded separately. The comparison contains 18 women and 19 jury-trial events, because Andrea Yates had two jury trials.

The original selection focused on U.S. cases involving a mother, her biological or adopted child, a death, an actual homicide jury trial, and an identifiable outcome. Guilty pleas, bench trials as the initial proceeding, competency-only hearings and cases involving someone else’s child were outside the primary jury comparison.

A later bench disposition may still be relevant to follow-up. Schlosser’s first proceeding was a jury trial; her subsequent judge-entered insanity verdict is tracked in a different column. That later finding must not be counted as the first jury’s verdict.

Unit of analysis: one woman, using her first recorded homicide jury trial in this compilation. Multiple children and multiple charges do not create multiple observations. Yates’s retrial is a follow-up event, not a second woman.

Source handling: the table is adapted from the supplied historical research and the master workbook’s filicide research section. Court opinions and contemporary reporting are linked for the legal distinctions and recorded outcomes. This page does not reproduce outdated “currently deliberating” language from the earlier research.

What did the first jury do?

Counting each of the 18 women once produces the following first-trial results:

Calculated from the 18 first-trial entries below. Percentages are rounded to one decimal; rounded values total 100.1%.

Casey Anthony accounts for the homicide acquittal in this list; Deanna Laney accounts for the first-jury NGRI result; Dena Schlosser accounts for the hung jury. The categories remain separate even when later proceedings change the case’s disposition.

The historical comparison: 18 women

The first-result column drives the calculation. The follow-up column explains selected later developments and charging distinctions; it is not a comprehensive account of every appeal, release decision or other prosecution.

Yates’s 2006 retrial is the nineteenth recorded jury event. Its NGRI result belongs in the follow-up analysis; it does not change the fact that her first jury returned a guilty verdict. [4] [5]

Victims and counts must remain distinct. Routier’s prosecution concerned Devon’s death, although both Devon and Damon died. Yates’s five children died, while the prosecution involved charges concerning three. A table that substitutes “children convicted of killing” for “children who died” would distort the record. [4] [11]

The seven-case insanity / criminal-responsibility subgroup

The original research separately identified Massip, Yates, Laney, Schlosser, Burchett, Schenecker and Sanders as cases in which an insanity or criminal-responsibility defense substantially shaped the trial. This is a preidentified subgroup, not a claim that other defendants had no psychiatric history.

Five guilty first verdicts show that a psychiatric defense does not prevent conviction. They do not establish that those five defendants lacked mental illness. Whether illness met the legal standard, what evidence was admitted, and whether the trial contained reversible error are separate questions.

The term “mental-health defense” also needs care. Evidence of depression, trauma, suicidality or medication exposure is not necessarily a formal insanity defense. Classification should follow the actual pleadings, jury instructions and disputed legal issue.

Following the cases changes the picture

Four of the seven selected women have a later NGRI disposition recorded in the comparison: Massip, Yates, Laney and Schlosser. Burchett and Schenecker have guilty outcomes. Sanders must be kept separate because her convictions were vacated and a new trial was ordered; the appellate court did not decide the insanity question.

“Final” can also mislead when a record has appeals, commitment reviews or later proceedings. Here, “later disposition” means the adjudicative outcome summarized in this comparison, not a claim that every possible legal issue has permanently ended.

Four procedural pathways worth understanding

Massip: a trial judge’s later ruling

Massip’s jury returned a second-degree murder verdict and a sanity finding. The trial judge subsequently reduced the offense and substituted an insanity finding. The appellate account documents this unusual sequence. It should not be summarized as “the jury found her insane,” and it does not establish that any trial judge can replace any jury verdict under any jurisdiction’s rules. [3]

Yates: reversal followed by a new jury

Yates’s first conviction was reversed after materially false expert testimony about a supposed television episode. That reversal addressed trial error; it did not itself acquit her by reason of insanity. A separate 2006 jury returned NGRI. Keeping those events separate shows why first-trial and later-disposition tables answer different questions. [4] [5]

Schlosser: deadlock followed by a judge’s decision

Schlosser’s first jury deadlocked. A later proceeding resulted in a judge’s NGRI determination. A deadlock therefore cannot be used as shorthand for either “almost convicted” or “almost acquitted.” A reported vote split, if reliably documented, describes that jury at that point; it cannot be transferred to another case. [6] [7]

Sanders: an evidentiary ruling, not an insanity acquittal

On August 6, 2026, the Massachusetts Supreme Judicial Court vacated Sanders’s convictions and ordered a new trial because exclusion of relevant psychiatric medical records was prejudicial. The opinion explains the importance of presenting evidence bearing on criminal responsibility. It does not establish that Sanders met the insanity standard. [10]

The trial was before Judge William F. Sullivan. Elliot Levine represented Sanders at trial; Robert F. Shaw Jr. represented her on appeal. Shared jurisdiction or courtroom personnel can make a procedural comparison useful, but they do not make Sanders’s evidence identical to Lindsay Clancy’s or predict another result. [10]

Burchett and Schenecker provide the other necessary side of the comparison: juries rejected their insanity defenses. A responsible discussion includes those outcomes alongside later NGRI dispositions. [8] [9]

What this research can—and cannot—tell us

The comparison establishes that guilty verdicts, insanity acquittals, deadlocks and subsequent changes all appear in the identified cases. It makes legal pathways visible. It cannot calculate the probability of a particular defendant receiving one of those outcomes.

• Selection bias: famous cases and cases with accessible records are easier to include than obscure or poorly documented ones.
• Disposition bias: excluding pleas and bench trials changes the population under study. Parents who died by suicide before prosecution never reach a homicide jury trial.
• Different laws: states use different insanity standards, burdens and procedures. Outcomes across states are not interchangeable measurements of the same test.
• Different evidence: diagnoses, expert opinions, prior symptoms, contemporaneous behavior and admissibility rulings vary.
• Timing: a first verdict, a reversal, a retrial and a commitment decision are distinct events.
• Small denominators: in the 18-woman sample, one case represents about 5.6 percentage points. In a seven-case subgroup, it represents about 14.3 points.

Research drawn from psychiatric hospitals, prisons or homicide records can also produce different clinical profiles. The forensic literature cautions that study populations influence apparent maternal and paternal patterns. Results from a hospital-based sample should not be treated as the psychiatric profile of every filicide defendant. [1]

An evidence framework for each case

A useful comparison records the victims and charged deaths; jurisdiction and instructions; medical and psychiatric history; symptoms close to the event; medications and treatment; expert opinions and their basis; planning and concealment evidence; prosecution and defense theories; the verdict by count; and each subsequent ruling.

Planning, concealment, a calm appearance or an inconsistent statement can be relevant evidence. None is a stand-alone diagnostic test. Likewise, a diagnosis cannot replace analysis of the applicable legal standard. The question is what the full record supports.

The children remain individuals, not statistical units or rhetorical tools. Comparing outcomes should illuminate the evidence without minimizing their deaths or using the horror of an act to bypass the criminal-responsibility question.

Continue through the case files

The historical sample is fixed for these calculations. Developing cases belong in the active-watch record and their individual pages; adding or reclassifying a case would require recalculating and relabeling the sample.

For the clinical and legal distinction, read Clinical Lens: Psychosis Is Not a Verdict. For a paternal/familicide comparison outside this maternal jury sample, see Chris Watts.