How we calculate judge and hearing office approval rates

SSA counts by federal fiscal year, which runs from October 1 to September 30. Each disposition file is a running total from October 1 to the file's date. The September file holds the whole year; the next month's file starts again from zero.

That makes a single file uneven. Nationally, SSA's file through September 26, 2025 held 329,700 judge decisions; the file through October 31, 2025 held 28,206. A judge's rate in an October file rests on a few weeks of hearings, so it can sit far from the same judge's rate a month earlier for no reason but the calendar.

To give every judge and office the same, full-length window at any time of year, we add the rest of the previous fiscal year to SSA's year-to-date figure:

12 months = this fiscal year to date + all of last fiscal year − last fiscal year to the same point

The last term is SSA's file from the same month a year earlier. Subtracting it leaves only the part of last year that came after that point, so the result is the most recent 12 months. We do this separately for decisions, approvals, denials and dismissals, then divide. Here it is for Judge Maria Hodges, using SSA's file through July 31, 2026:

SSA's own file for the same date shows 48% for Judge Maria Hodges, from ten months. Our 47% covers twelve. The window starts the day after last year's file, and SSA's files close on different days of the month, so a window can run a few days over or under a calendar year. In a September file the last two terms cancel out and the window is simply the fiscal year.

• A different window. SSA's file is the fiscal year to date. Ours is the most recent 12 months. Other sites may show a fiscal year, a calendar year or a longer stretch.
• One judge, every office. SSA lists a judge separately at each hearing office where they closed cases. We add those rows together, so a judge who moved or helped another office has one figure across all of them.
• What counts as approved. We count fully and partially favorable decisions. A site that counts only fully favorable decisions will show a lower rate, and one that divides by cases closed, dismissals included, will too.
• Rounding. We round to whole percentages from the exact counts, so a figure can differ by a point from one computed to a decimal.
• Revised counts. SSA occasionally revises an earlier file. If a revision would make one of our subtractions fall below zero, we count zero rather than a negative number. It is rare.

Active judges. A judge is active when they appear in SSA's newest disposition file. Their page leads with the 12-month figures. A judge who no longer appears keeps a page with every decision on record, labeled with the year of their last file, so an older rate never reads as current. A judge whose record starts this fiscal year has a shorter window, from October 1, and their page says so.

The bar. The judge comparison tool and the similar-judges lists on judge pages include only active judges with a full 12 months of records and at least 100 decisions in them. Similar judges are those within 5 points, nearest first.

Margins. A rate built on a few hundred decisions moves by chance. The table shows how far, at the national approval rate of about 58%: the range that holds 95% of the time if a judge's underlying rate stayed the same.

Dismissals are left out of the rate because they say nothing about how a judge rules. In the example above, Judge Maria Hodges closed 619 cases in the 12 months: 563 decisions and 56 dismissals. The 47% is 263 of the 563 decisions.

A judge's page also shows a fiscal-year trend and every decision on record since SSA's files begin; those use the same definitions over different periods. Wait times are SSA's own monthly figure and are not recalculated.