AREALME IQ Test Data: Completion Rates and Item Statistics

The AREALME IQ Test now publishes a transparency report built on the current 119-question bank. The public data page uses English-interface attempts from September 5 to 13, 2026. The previous release mapped a July-August sample onto the current order, so question 50 had no live stats.

This is a living dataset. When the bank changes, or when a later window is stronger, we will publish a new release and keep the older files available.

AREALME IQ Test data disclosure showing 5,327 attempts, 2,162 completions, and 40.59% completion for the September 2026 sample

Overview of the September 2026 English-interface sample. Country labels and question previews are on the data page.

Key Numbers from the September 2026 Sample

Metric	September 2026 sample	Notes
Total attempts	5,327	English interface only
Completions	2,162	40.59% completion rate
Question 50 (new item)	30.38% correct	First live baseline; Complete-path item
Question 57 (maze)	72% completed	Continuous-scored item; not binary correct/incorrect
Question 119 (maze)	67.43% completed	Final maze; same completion-rate rule

Question 50 is the 2026 replacement item. Its classical item-difficulty index (correct answers divided by recorded answers) is 30.38%. The two maze items use completion rate instead, because their scores are continuous. Among people who finished those mazes, mean scores were 89.33% on question 57 and 94.63% on question 119.

Quick and Complete share the same bank, but they are not the same path. Quick is 49 questions. Only Complete reaches question 50 and the later mazes, so attempt totals are not a straight funnel from start to question 119.

Split	Attempts	Share
Complete	3,527	66.21%
Quick	1,800	33.79%
Mobile	3,661	68.73%
Desktop	1,546	29.02%
Tablet and other devices	120	2.25%

What’s in the Public IQ Dataset

The data page publishes:

Attempt, completion, and timing totals
Item-level stats for all 119 current questions, including how many people reached each item, how many answered, the published rate, and mean active time
Country groups with at least 20 attempts
CSV and JSON downloads, plus a file manifest, for academic reuse

The item table is the core of the release. For conventional questions, the published rate is the classical item-difficulty index. Each row also shows item drop-off: how many people reached the question versus how many answered it.

This release does not include mean IQ, median IQ, standard deviation, or a score histogram. Those belong to the live scoring model, which uses a rolling calibration window. The public files are attempt, sample-composition, and item-level aggregates.

Why We Keep Publishing New Samples

A transparency report only stays honest if it follows the live test. We add items, retire items, and rebuild the bank when the evidence says we should. This is the first release collected entirely on the current 119-question bank. If a later update changes a question, the next sample should include that question instead of leaving a blank row.

Methodology, Sampling Limits, and Data Reuse

Use the page as a record of this sample, not as a verdict on a person or a country.

The sample is self-selected. Country counts show who found the English test, not a census. That demographic skew is why the United States (1,499 attempts) and the United Kingdom (788) are the two largest groups. Those figures are not national IQ estimates.
English is the interface language, not a verified native-language field.
Resume attempts count in the totals. They are excluded from item statistics and from scoring-calibration inputs. Someone who leaves and comes back is still an attempt; their later answers are not mixed into the item rates. Continue-to-Complete after a Quick result is not a resume.
Academic reuse of this release is CC BY-NC 4.0. Credit AREALME, keep the license, and do not use the files for commercial products.

If you want the longer case for why public IQ data matters, read Online IQ Tests Have a Trust Problem.

Take the Test, or Download the Files

For people taking the test

Take the 119-question test. Finish Quick or Complete, then compare what you saw with the public item table if you want the evidence behind the questions.

For researchers and educators

Open the data page, download the CSV or JSON for release 2026-09-16-en-current-bank, and cite the September 5 to 13, 2026 window. Earlier mixed-bank files stay in the archive. When we publish the next sample, that older file will still be there.