The results of three previously published experiments are described where in each case the positions of (1) the positive and negative labels and (2) the values on different Likert-type scales were systematically varied. In each study different scales were used and different participants took part. The results showed that in each of these three studies positioning the positive label on the left-hand side of the scale (and starting with a high value) lead to higher mean scale scores than positioning the positive label (and ending with a high value) on the right. Some possible explanations for this effect are discussed, as well as some of the implications.