Coordinated Stalking Testimonies: A Coded Reading of 88 California Accounts, 2019

Each person's surname, ZIP code and first name were combined using the U.S. Census method known as BISG (Bayesian Improved Surname Geocoding). The result is a probability for each Census group. Compared with the population of their own ZIP codes, the writers lean white: an estimated 62%, against 47% in their neighborhoods. They lean less Hispanic, 20% against 31%, and somewhat less Asian or Pacific Islander, 7% against 12%. Black writers match their neighborhoods, at 8% against 6%. The white and Hispanic gaps are about two standard errors, suggestive but borderline once six groups are tested.

The most likely explanation is how the testimonies were gathered. Collections like this spread through an organization's own contacts, support groups and referrals, and one writer (#51) runs such groups in several states. The form was also in English. Recruitment that follows existing ties can leave whole communities out. Not one writer has a Chinese or Korean surname, although Chinese Americans alone are roughly 3–4% of Californians and more in the Bay Area. On that figure about three would be expected, and a sample with none would occur only about one time in 25. The Asian and Pacific Islander writers who do appear are Samoan, Filipino, Vietnamese, Lao and South Asian. These gaps describe who the organization reached, not who experiences this.

The method also has blind spots. Married names hide origin, and the Census counts Middle Eastern and Persian people as white.

Coded by language of origin, the surnames are mostly from the British Isles (44). Next come Spanish (15), German or Scandinavian (6), German or Ashkenazi Jewish (4) and Italian (4). Smaller groups are Persian, Samoan, Eastern European, Greek, Arabic, Filipino, Vietnamese, Lao, Portuguese and South Asian. Some first names show an origin the surname hides: Colombian, Japanese, Persian, Samoan and Lao. Only one writer states their own background (#81, Indian-Muslim).

Estimated ethnicity made no clear difference to outcomes. White-leaning and Hispanic-leaning writers reported to police at similar rates (66% and 71%).

Birth years were estimated from Social Security name counts for 1915–2001, adjusted for survival to 2019. This works for 79 of 89 people. As a group, the estimates match the US adult population almost exactly, with a median around 50, so first names alone reveal no age pattern here. The six writers who state an age, or give enough to work one out, all sit at or above the older end of their name's range. That hints the group is older than names suggest, but six people is too few to show it.

Several writers describe the harassment as having a purpose, not only effects. The most common is fear. Nineteen testimonies describe living in fear, and some name it as the aim: #51 says a confrontation’s “intent … was to invoke fear,” and #67 describes “scare tactics to keep me in fear of my life 24/7.” The second is isolation. Thirty-two testimonies describe losing friends or being cut off, ten say family members were also targeted, and #91 describes strangers who “want to become friends only to turn out as perpetrators gathering intel in order to influence old friends.” Four writers name religious faith as what sustains them; none says faith ended the harassment. These are the writers’ own interpretations, and the page reports them as such.

By first name, 44 people are women, 37 are men, and 8 names do not show gender; one of those (#7) is a man by his own text (“my son and wife”). With him counted, women are 44 of 82, or 53.7%. A one-sided binomial test against 50% gives p ≈ 0.29, with a 95% interval of 43–64%. H1 is not supported. Even if every unclear name were a woman (52 of 89, 58.4%), p ≈ 0.07.

This compares the 43 testimonies from women with the 38 from men. The biggest gap is fear and staying at home: 35% of women mention it, against 11% of men. Women also mention physical health decline, family strain and involuntary holds more often. But the sample is small and 27 comparisons were made. Once that is accounted for, none of the differences is statistically reliable. Treat them as leads for a larger study, not findings.

One hypothesis holds that targets are kept from knowing one another, and that if two do meet, at least one is recruited and told what happened. If this is correct, connections within a group of N people could bring the truth to N − 1 of them, leaving only the highest-priority target uninformed. The hypothesis can be tested one meeting at a time, provided the terms are fixed in advance: after each meeting between consenting participants, both record within a set period, such as 90 days, whether either was approached. Meetings after which both report no contact would count against the hypothesis. One limitation remains: if a recruited person is also instructed to conceal it, their report cannot be verified, and the test depends on participants’ honesty.

Readers who regard these accounts as delusion need not rely on the author’s analysis. Anyone who already knows a person making such an account, a relative, friend, or neighbor, can, with that person’s agreement, spend time in their surroundings and record what they observe. If the accounts are accurate, the visitor would expect to encounter some of the same patterns: repeated vehicles, unusual attention, or contact from officials. If nothing of the kind appears over a reasonable period, that is also evidence. Such observations are most useful when written down at the time, with dates, and compared against what the visitor would expect to see by ordinary chance.