Clinicians and researchers have identified a pattern of "mixed" symptoms that are sometimes exhibited by individuals with bipolar spectrum disorders (BSDs). However, the criteria for these mixed states as outlined by the American Psychological Association have been criticized for being too stringent for most individuals that experience impairing episodes of mixed symptomatology (e.g., Akiskal, 1996). Mixed states are associated with a more impairing course of illness and suicidality. More research is needed on mixed states, and there is particularly little evidence for risk factors in recent-onset samples. The aims of this study were to 1) examine the prevalence of mixed states in a sample of individuals with recent-onset bipolar spectrum disorders, 2) examine the symptom structure of hypomanic and depressive episodes, and 3) examine some of the risk factors associated with mixed states. Participants in sample 1 were adolescents, initially aged 14-19, selected for exhibiting either moderate or high Behavioral Approach System (BAS) risk for first onset of BSDs. Participants in sample 2 were 18-24 year old undergraduates from Temple University recruited for having a bipolar spectrum diagnosis. Mixed states captured 37.10% of all episodes examined in sample 1, yet only 13 (10.48%) of these episodes met available research criteria for mixed mood episodes. Factor analysis yielded two adequate models that fit the data; one model had two factors that aligned with traditional depressive and hypomanic symptomatology, and another model had three factors that aligned with hypothesized overactivity, inhibited depression, and irritable risk taking components of bipolar disorder. Latent class analysis allowed for examining observed patterns of responses within individuals, and then grouping heterogeneous groups of individuals into classes based on similarities on dimensions of interest, performance within dimensions, or both (Nylund, Bellmore, Nishina, & Graham, 2007). The latent class analysis showed that three classes best defined bipolar spectrum individuals in sample 1: low impulsivity, aggressive, and substance problems classes were obtained. The `aggressive' class was significantly more likely than the `low impulsivity' class to experience any mixed symptomatology, although a continuous measure of mixed symptoms did not yield significant differences between classes. Overall, the results from the current study support findings suggesting that mixed mood states are more commonly experienced than originally believed. These results extend previous studies to include individuals with bipolar spectrum disorders, not solely bipolar I and II disorders. These findings also suggest that non-treatment-seeking samples may have different types of mixed mood states than those seeking treatment. These findings add support to the literature that individuals with BSDs and comorbid substance use diagnoses are at increased risk for chronic illness, and show that these individuals are also more likely to experience mixed mood states than those without comorbid substance use diagnosis. Treatment providers should be aware of the complications that are inherent in bipolar individuals with comorbid substance diagnoses, as they are more likely to experience more episodes as well as mixed symptoms. / Psychology