I-90 Median Rises to 42 Days on a Preliminary Early Read

At a preliminary sample of 5,908 cases this week, the I-90 median can reflect who happened to be approved as much as any underlying movement. The week's 42-day reading is 6 days above last week's 36, and the outlier flag fired at z = 2.12 against the prior 4-week trailing mean — not against the 12-week average, which sits at 60 days. Against that longer window, this week's reading is actually below average. The 8-week regression slope is –0.8 days per week with an R² of 0.02, well below the threshold for asserting any directional trend. The 12-week range runs from 15 to 167 days, placing 42 days comfortably within the recent scatter.

The median for I-90 final-stage waits opened the 12-week window at 167 days on May 4, across 10,288 approved cases that week, then fell sharply through May and June. By the week of June 15 the median had reached 15 days, with weekly sample sizes rising to 7,391 and then 10,216 to 12,938 cases in subsequent weeks. The scale of those completions — more than 10,000 cases cleared in a single week — appears to follow a period of concentrated adjudication (USCIS's decision step), working down cases that had accumulated at earlier stages. Since that mid-June low, the median has moved upward: 21 days (June 22), 21 days (June 29), 36 days (July 13), and now 42 days this week. At 42 days, the reading sits well below the May highs and roughly 18 days below the 12-week trailing average of 60 days.

The 8-week linear trend has a slope of –0.8 days per week and an R² of 0.02, a fit too weak to support any directional claim. The pattern is labeled "noisy_stable" in the data: weekly medians have ranged from 15 to 167 days over the 12-week window, with no consistent upward or downward line. The 12-week trailing average of 60 days serves as a better reference point than any slope. This week's 42-day reading falls below that average, as did five of the last seven weeks.

Cases with an RFE (Request for Evidence) response on file posted a median of 20 days this week, compared with 59 days for cases that went through plain processing with no RFE. That is a 2.93x speed ratio, and it triggered the rfe_fast flag in the data. The counterintuitive direction — RFE cases finishing faster than non-RFE cases — most likely reflects a batch-clearance effect: a group of RFE responses received around the same time moved through the adjudication (USCIS's case-decision step) queue together and cleared in a concentrated window. The data brief does not tell us why this particular week produced the inversion. What the data does not support is a conclusion that receiving an RFE improves an individual applicant's outcome; RFE cases in the broader 30-day window show a median of 33 days across 2,110 cases at the response stage, with a 25th percentile of 15 days and a 75th percentile of 68 days, consistent with normal variability.

RFE-response cases cleared nearly three times faster this week, most likely because a batch of responses received around the same time moved through together — not because an RFE itself speeds a case along.