One of the most important issues regarding survey research is the design of questionnaires. Its correct definition constitutes an essential point in the process of obtaining valid, reliable and quality data. For this reason, it is necessary to pay full attention to aspects such as wording or questions and answers order. A split-ballot experiment was conducted to test the hypotheses that slight changes in wording, question placement on the questionnaire and reading of answers could have in the results of a political survey. The results show that some variables were especially sensitive to a variation of placement on the questionnaire and changes in the order in which the answers were read. At the same time, a different wording of the same concept did not show significant differences.