Aims To identify socioeconomic, behavioral and clinical factors that are associated with prediabetes according to different prediabetes definition criteria. Methods Analyses use pooled data of the population-based Cooperative Health Research in the Region of Augsburg (KORA) studies ( n = 5312 observations aged ≥ 38 years without diabetes). Prediabetes was defined through either impaired fasting glucose (IFG), impaired glucose tolerance (IGT) or elevated HbA1c according to thresholds of the American Diabetes Association. Explanatory variables were regressed on prediabetes using generalized estimating equations. Results Mean age was 58.4 years; 50% had prediabetes (33% had IFG, 16% IGT, and 26% elevated HbA1c, 10% fulfilled all three criteria). Age, obesity, hypertension, low education, unemployment, statutory health insurance, urban residence and physical inactivity were associated with prediabetes. Male sex was a stronger risk factor for IFG (OR = 2.5; 95%–CI: 2.2–2.9) than for IGT or elevated HbA1c, and being unemployed was a stronger risk factor for IGT (OR = 3.2 95%–CI: 2.6–4.0) than for IFG or elevated HbA1c. Conclusions The overlap of people with IFG, IGT and elevated HbA1c is small, and some factors are associated with only one criterion. Knowledge on sociodemographic and socioeconomic risk factors can be used to effectively target interventions to people at high risk for type 2 diabetes. Electronic supplementary material The online version of this article (10.1007/s00592-020-01573-x) contains supplementary material, which is available to authorized users.