Detector accuracy declined significantly as text length increased, with both detectors performing best on short texts¶
Claim · All claims
Evidence · 1 study · designr2·q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q2 i? Both detectors performed best on short texts (300–330 words) with accuracy declining for medium and long texts; the length–accuracy association was statistically significant for both (Turnitin χ²(2)=8.41, p=0.0149; Originality χ²(2)=13.17, p=0.0014). → Mohammad Hadra 2026
Evidence¶
Mohammad Hadra 2026¶
Mohammad Hadra, Karleen Cambridge and Mostefa Mesbah. (2026). Evaluating the accuracy and reliability of AI content detectors in academic contexts. International Journal for Educational Integrity. https://doi.org/10.1007/s40979-026-00213-1
q2 · i? · design · r2
Chi-square tests of independence on classification correctness across three word-count ranges (300–330, 450–550, 900–1,100 words) in the 192-text dataset. Table 3 shows Turnitin accuracy 0.87 for short versus 0.56 for medium texts; the article reports "Turnitin results yielded χ² (2) = 8.41, p = 0.0149".
"Pearson’s chi-square tests confirmed that these differences were statistically significant for both detectors. Turnitin results yielded χ² (2) = 8.41, p = 0.0149, and for Originality χ² (2) = 13.17, p = 0.0014."
Discussion¶
Related Claims¶
- EchoPrompt's AUROC rises with text length, from an average of 75.8% in the shortest 1-40 word bin to 98.9% in the 321-360 bin on DetectRL Length — reports the opposite
- Originality showed a borderline non-significant trend toward higher accuracy on professional than EFL student writing, while Turnitin showed no significant difference — related
- EchoPrompt maintains cross-proxy robustness, averaging 87.06% AUROC across seven proxy models and outperforming DNA-DetectLLM by 2.70% AUROC on average — reports the opposite
- Originality outperformed Turnitin in overall accuracy and macro-average recall on a balanced EFL-relevant dataset, but both showed limited consistency across classes — related
- Both detectors achieved substantially lower accuracy on scientific writing than on humanities texts — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
detector-accuracy-declines-with-text-length |
title |
Detector accuracy declined significantly as text length increased, with both detectors performing best on short texts |
description |
Detector accuracy declined significantly as text length increased, with both detectors performing best on short texts |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2026 |
generated |
by process:wiki-ingest on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — mohammad-hadra-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.