The Strength and Difficulties Questionnaire (SDQ) is one of the most common screening instruments for emotional and behavioral problems in children and adolescents. Although exploratory factor analyses support the originally proposed 5-factor structure of SDQ as well as a 3-factor model, the evidence from confirmatory factor analyses is more mixed. Some of the difficulties items in SDQ are positively worded and it has been proposed that this leads to method effects, i.e. these items share variance that is due to the method used rather than to a substantive construct. Also, there seems to be minor factors in some subscales. This study tests a series of 3- and 5- factor models pertaining to the factor structure of SDQ, also considering method effects and minor factors. The sample consists of a nationally representative cohort of Swedish adolescents born in 2001 (n = 5549). Results show a relatively better fit of the 5-factor model compared with the 3-factor model although neither of these had a satisfactory fit. Model fit was improved when specifying cross-loadings of the positively worded difficulties items on the prosocial scale as well as minor factors on the hyperactivity scale. Although no model provided a completely satisfactory fit to the data, the results show that the 5-factor model performs better than the 3-factor model and has an acceptable fit. We conclude that for the purposes of epidemiological research, SDQ has acceptable factorial validity, provided that researchers consider method effects and minor factors.