Nudge and bias in subjective ratings? The role of icon sets in determining ratings of icon characteristics