The Tip Log

The log does not look like evidence. It looks like inventory.

In most departments it begins as a printed form, later a database screen: one row per tip, columns across the top. Date received. Time. Method, which means phone, walk-in, letter, email, and, in recent decades, the department's social media inbox. Caller name, or "anon." A summary field, rarely more than three lines. An assigned investigator. A priority code. A disposition: followed up, no further action, referred, duplicate. A file of this kind can hold hundreds of rows by the end of the first week and thousands by the end of the first year.

What follows is a composite. It is built from the structure of real tip logs described in published case reviews and court records, but the entries themselves are invented, and no real child or caller stands behind them. Read them as you would read the real thing.

Row 41. Day two, 9:14 p.m. Phone. Female caller, gives name. "Saw a girl matching description in a blue coat at the gas station on Route 9, about 4:30 on the day she went missing, with a man in a gray pickup." Priority B. Assigned. Disposition pending.

Row 42. Day two, 9:20 p.m. Phone. Anonymous. "Psychic vision. Water, a bridge, the number 7." Priority D. No further action.

Row 43. Day two, 9:31 p.m. Email. "The stepfather did it. Look at his face on the news." Priority C. Referred to lead.

You already know how to sort these, or you think you do. The psychic goes to the bottom. The anonymous accusation built on a televised face goes in the middle, because it contains no fact, only an opinion that happens to point at someone already in the file. The gas station sighting rises to the top: named caller, specific time, specific place, a vehicle. Investigators triage this way for good reason. A tip that can be checked against something outside itself, a camera, a receipt, a second witness, is worth more than one that can only be believed or disbelieved.

Most departments formalize this. Large investigations often use a lettered or numbered scale, with the top tier reserved for tips that are specific, verifiable, and tied to the critical hours. Major case management systems used in the United States and abroad are designed to force every tip into a row, so that nothing lives only in someone's memory. The procedure is sound. A tip log is the department's promise that it heard everyone.

So read row 41 the way the investigator would have.

It says a named adult saw a child who matched a description, at a time inside the window you circled at the end of the last chapter, in a place with likely video, beside a vehicle anyone could check against registrations. It misses whether the caller had seen the description on the news before she saw the girl, which matters, because a sighting reported after a broadcast is a memory that has already been handed its outline. It forces a question you can answer in an afternoon: is there footage from that gas station at 4:30?

Now turn the page. The follow-up sheet is stapled behind the log.

Day four. Investigator's note, handwritten: "Station camera overwrites at 72 hrs. Too late. Caller re-interviewed by phone. Unsure of coat color on reflection. Low value. Inconsistent w/ timeline."

You may feel relief. The system worked. The tip was ranked high, assigned, checked, and closed for a reason written in ink. A camera erased itself, a witness wavered. Every investigation loses tips this way, and you have already seen how quickly the living evidence ages.

Read the last three words again. Inconsistent w/ timeline.

Whose timeline? Not the camera's, which no longer existed. Not the caller's, who still placed the girl at the station at 4:30. The phrase refers to something the investigator was already carrying: a version of the afternoon in which the child never left her own street.

Go back to the spiral pad from Chapter 3, the one with the coffee ring. You marked the first assumption on its first page, the line written within an hour of arrival, before any interview was finished. In this composite it reads: "Likely domestic. Child did not leave area." At the time it was a reasonable hypothesis. You know the research on who most often kills young children, and an investigator who ignored it would be negligent. But a hypothesis written in the first hour has a way of hardening into a timeline by the fourth day, and a timeline becomes a filter.

Now sort the log again, this time by disposition rather than date. Do it slowly.

Every tip that placed the child near home was marked followed up. Every tip about the household was referred to the lead. Row 43, the opinion about a face on television, went upward because it fit. The tips that placed her elsewhere carry a different set of words in the disposition column: low value, unverifiable, does not match. A gray pickup appears in four separate rows, from four callers who did not know one another. None of the four was cross-referenced against the others. Each was closed alone.

Then, in a different ink, in the margin beside row 41, there is a note dated eleven months later, initialed by a sergeant reviewing the stalled file:

"This should not have been closed on timeline. Timeline was ours, not evidence. Pickup tips never linked. My error as supervisor."

That line is rare. Real files seldom contain a sentence in which someone writes down that the department's own belief did the sorting. But researchers have documented the pattern from the outside for years. Keith Findley and Michael Scott, in a 2006 article in the Wisconsin Law Review, examined wrongful convictions and described tunnel vision as a "compendium of common heuristics and logical fallacies" that leads investigators to focus on one suspect or theory, then select and filter evidence to build a case while ignoring or discounting evidence that points elsewhere. Confirmation bias, the tendency to seek and weigh information that supports what one already believes, appears in their account as an ordinary feature of human reasoning, present in careful people and amplified by pressure to resolve a case. Saul Kassin, Itiel Dror, and Jeff Kukucka, writing in 2013 in the Journal of Applied Research in Memory and Cognition, showed that the same distortion reaches the laboratory. Examiners who know the investigative theory can be pulled toward it when they interpret ambiguous results.

The tip log is where this becomes visible, because the log records both what came in and what was done with it. The incoming column is the public. The disposition column is the investigator's mind.

Then the noise arrives, and the noise is not neutral. After the first broadcast the calls double. After the stepfather is filmed leaving the station, the email column fills with rows like 43. When online forums take up the case, the volume can overwhelm a small department, and what overwhelms it is repetition: the same theory, pasted and rephrased, arriving from strangers who have read the same thread. A tip log has no column for source of belief. A hundred people who saw one video become a hundred rows, and a hundred rows pointing the same way look, on a busy night, like corroboration. The few callers who saw something with their own eyes are now a small fraction of the page, and the theory the crowd has adopted is often the theory the investigators already held, handed back to them with the weight of numbers.

Nobody lied. The caller at the gas station told what she remembered. The investigator closed a weak tip with a stated reason. The commenters believed what they posted. The sergeant, eleven months on, wrote the truth in a margin no one was required to read.

Go back to your own margin, the one with pencil hours, pen hours, and small circles. Add one more column, to the left of everything, and in it write the first assumption from the spiral pad. Then go down the list of hours and ask, for each one, whether what you know about it came from the evidence or from the sorting. Be strict here too. You will find hours you marked in pen because a tip was closed, and the tip was closed because it disagreed with the first line on the first page.

Those hours go back to pencil. Some of them were inside the window.

The four rows about the gray pickup are still in the file. The callers' names are written on them, along with the dates. One of the four mentions a partial plate, three characters, and the disposition field beside it is blank. Nobody ever wrote down what happened to that tip.