The current study addressed some of the methodological shortcomings of previous studies on internet addiction. The main purpose of the study was to determine if two different internet addiction assessments would identify the same individuals as addicted to the internet. A total of 224 undergraduate internet users were surveyed using a stratified sampling plan based on the proportional allocation technique to procure as diverse a sample as possible. The assessments used were Young’s Internet Addiction Test (IAT), Caplan’s Generalized Problematic Use Scale (GPIUS), a demographic questionnaire, and a reasons-for-use questionnaire. Results showed that about 0.9% of the sample could be considered addicted to the internet according to both the IAT and GPIUS, which is a smaller percentage than found in previous studies. There were too few participants identified as addicted to the internet to determine if these two assessments identified the same individuals as addicted, however, it was shown that over a third of the sample was identified as “at risk” for addiction by one assessment and not the other. These results lead to the conclusion that the assessment measure used is of extreme importance when diagnosing internet addiction. Also, more robust sampling procedures may lead to fewer internet addicts identified, which could be a more accurate reflection of internet addiction in the target population. Regression results indicated that demographic and psychological predictor variables could more successfully explain the variance in IAT scores over GPIUS scores, although a similar pattern of direction and effects were shown for both criterion variables. Hierarchical regression revealed that the demographic variable “age” may be particularly important when attempting to predict internet addiction scores.