Internet Addiction Test: 20 questions about how you use the internet
Kimberly Young published the IAT in 1998, adapting the diagnostic criteria for pathological gambling into 20 questions about going online. The items ask about staying online longer than you intended, losing sleep, hiding how much time you spend, and feeling restless when you are offline. The wording shows its age, checking email included, and it is still the most widely translated questionnaire of its kind.
Each item is rated 1 (rarely) to 5 (always). The total runs 20 to 100, higher means more problems, and nothing is reverse scored. An earlier format scored 0 to 5 with the same 100-point ceiling, so cutoffs published for one version do not transfer to the other.
Reading a score
| Total | Reading |
|---|---|
| 20-30 | Mild or no problematic use |
| 31-49 | Mild |
| 50-79 | Moderate |
| 80-100 | Severe |
The line most studies use is 50, where moderate problems begin. Another widely cited set of bands puts 20 to 39 as an average online user, 40 to 69 as excessive use, and 70 to 100 as significant problems. The two schemes disagree in the middle, which is a fair summary of how settled this field is. Treat a score near a boundary as a score near a boundary.
How common is it
Estimates range widely, and the range itself is the finding. Greenfield (1999) put it near 6% of internet users, Scherer (1998) found about 14% among college students, and Weinstein and Lejoyeux (2010) settled on 6 to 11%. In an Italian student sample, 36.3% landed in the excessive band and 1.9% in the significant problems band. A Korean sample of 279 students averaged 32.9 (SD 9.4). A Chinese validation reported 1.40% severe and 12.30% moderate.
Part of the spread comes from where the cutoff sits, and part from samples that differ in age, culture and how much of life happens online. Any claim that a fixed share of the population is addicted to the internet should be read with that in mind.
Reliability and structure
Internal consistency is consistently good: 0.91 in an Italian sample of 485 students (67.7% women, mean age 24.05), 0.91 in a Korean sample of 279 students (62.8% women, mean age 19.9), and 0.93 with a split-half coefficient of 0.94 in the Chinese validation. Over two weeks, the Korean retest correlation was 0.73.
What the items measure, beyond severity, is unresolved. Factor analyses have found anywhere from one factor to six. In the Italian sample, a one-factor solution explained 36.18% of the variance and a two-factor solution 42.15%, both fitting well (chi-square over degrees of freedom 2.86 and 2.68, CFI 0.99, RMSEA 0.02). A 2025 Chinese study of 1,042 students found three factors explaining 61.29% (chi-square over degrees of freedom 4.38, CFI 0.95, RMSEA 0.057). A Korean study found four factors explaining 59%. When a scale's dimensions depend that much on the sample, the total score is the only figure worth reporting.
Versions and limits
An adolescent version keeps the 20 items with reworded language, and a Korean translation went through a formal translation and back-translation process. The original items are copyrighted by the author and the publisher, and commercial use requires a licence.
The bigger limit is content. Several items assume a desktop era, and checking email as a symptom reads differently when the phone is the internet. Heavy use is also not the same as addiction: the scale asks about consequences, not hours, so someone who works online all day can score in the moderate range without having a problem. The opposite happens too, when the honest answers are about lost sleep and neglected relationships.
What this page is not
The IAT is a screening tool, not a diagnosis, and no questionnaire can decide whether your internet use is a disorder. If sleep, work, study or relationships are suffering, that is the useful signal regardless of the total, and it is worth talking to someone about it. This page is not a diagnosis and does not replace one.