The Marlowe-Crowne Social Desirability Scale (MC-SDS, Crowne & Marlowe, 1960) was designed as a unidimensional measure of social desirability for the purpose of controlling for this response bias in self-reports. Subsequent research has not supported the unidimensionality of the original 33 item scale (i.e. Loo & Thorpe, 2000). Several short forms of the MC-SDS have been developed and tested, producing mixed results of their adequacy (Ballard, 1992; Barger, 2002; Reynolds, 1982; Strahan & Gerbasi, 1972). Previous research on the MC-SDS has relied on data from paper-and-pencil administrations of the scale. However, since there is evidence that mode of administration affects measures of social desirability (Kaufmann & Reips, 2008; Martin & Nagao, 1989), findings regarding the psychometric properties of the MC-SDS scale may not hold in internet surveys. The purpose of the current research was to evaluate the reliability and factor structure of the full length and short forms of the MC-SDS scale using internet data. Participants (n=743) were recruited through communication sites and e-mail by snowball sampling. The questionnaire contained two personality tests, background questions and an Icelandic version of the MC-SDS (original data from Ragnarsson, Ólason & Halldórsson, 2011). Results from reliability analysis were in line with results obtained in previous research (α=.74 for the full scale and α ranging from .48 to .67 for short forms). Confirmatory factor analysis of the full scale revealed some very low factor loadings and multiple error covariance between items. An adequate fit was obtained for the MC-SDS short forms (apart from significant chi square values, which can be expected for a sample of this size). Modification indexes indicated that psychometric properties could be improved. In sum the results for the Icelandic internet version of the MC-SDS were for the most part compatible with prior results from paper-and-pencil administration of the English version.