The experimental conditions that produce analyte peak splitting in micellar electrokinetic capillary chromatography (MEKC) have been systematically investigated. The system studied was a neutral phosphate buffer and sodium dodecyl sulfate (SIDS) micelles as pseudostationary phase. A number of analytes showing a wide variety of hydrophobicity values and several organic solvents as sample diluents have been tested. Peak splitting phenomena are mainly due to the presence of organic solvent in the sample solution. They increase with the hydrophobicity of the analyte and decrease with the increase of the surfactant concentration. When hydrophobic compounds are analyzed the suggested ways to avoid split peaks are: (i) the use of 1-propanol or 1-butanol as sample diluent instead of methanol or acetonitrile or (ii) the use of high concentration of surfactant in the separating solution when the analyte must be dissolved in pure methanol or acetonitrile.