Background and Objectives: Nasometry is an easy, noninvasive method to obtain objective data regarding the function of velopharynx. However, because articulation errors may affect the results of nasometry, the examiner should interpret the nasalance score based on appropriate speech stimuli. The purpose of this study is to examine the difference of nasalance score between glottal and oral articulations in patients with velopharyngeal insufficiency (VPI). Materials and Method: Nineteen children between 3.4 and 12.1 years of age (mean age 5.7 years) with a confirmed VPl showing hypernasality and articulation errors (glottal stops) were included. Nasalance scores were obtained for two speech patterns of glottal and oral stops. In addition, the velopharyngeal functions were analyzed in four subjects using video nasopharyngoscopy. Results: The $mean{pm}S.D$ nasalance scores of the glottal stops and oral stops were $42.54{pm}16.26%$ and $25.47{pm}16.51%$ respectively (p=.000). Six of 19 patients achieved normal nasalance scores when glottal stops changed to oral stops by the trial speech therapy. Video nasopharyngoscope confirmed that large velopharyngeal gaps can be decreased into tiny gaps or complete closure when compensatory articulations were corrected for some cases. Conclusion: Compensatory articulation errors must be corrected for the reliable interpretation of the nasalance scores that are obtained in children with velopharyngeal insufficiency, which would facilitate to make a better decision for further management of these patients.