Two groups of year 6 pupils (age 10-11 years) each experienced science practical lessons that were essentially identical but for one difference: one group (theory-led) were told by the teacher what result they should expect, and the other group (hypothetico-deductive) were not. The theory-led group demonstrated experimental bias, recording results and drawing conclusions from phenomena that should not have occurred. The hypothetico-deductive group appeared to behave as neutral experimenters, and so were the better scientists. From the findings of this study, a useful pedagogy is described involving a cautionary tale related to allowing bias to influence science experimentation. (Contains 3 figures and 1 box.)