Social surveys are widely used as a major quantitative research technique in social sciences. Whilst the impact of survey design on people’s answers has been extensively discussed and investigated for paper questionnaires, the widespread use of the Internet in our society highlights the importance of testing more closely the impact of web survey design features. The objective of this research was twofold. First objective was to investigate whether social presence has an impact on quality of answers in the context of an online questionnaire. Second goal was to investigate order effect of five-point rating and frequency scales in web surveys. Four versions of an online questionnaire were designed and hosted by the server of the Sociology Department, University of Surrey. These differed from each other on two lines: the presence of an introductory picture and the order of five-point scales. Questionnaire one and four were introduced by the researcher’s picture, while questionnaire two and three were not. At the same time, in questionnaire one and two the five-point scale order was inverted compared to survey three and four. Consistently with the hypothesis of a motivational impact of social presence, this research finds a positive effect of the researcher‘s picture on reducing items nonresponse. Nevertheless, the results show little evidence of an additional impact on more complex survey tasks such as adding more comments in an open-ended question. With regards to rating and frequency scales, evidence of an asymmetrical effect between Agreeing and Disagreeing types of answers was found. Finally, the research unveiled a significant gendered pattern in answers to five-point scales, with men more likely to minimise hand/eye co-ordination by choosing options on the  right of the screen and women more likely to choose options closer to the question text.