To explain, if an individual was equally extraverted and conscientious and was assessed on a Likert-type scale, each trait would be evaluated singularly, i.e. respondents would see the item "I enjoy parties" and agree or disagree with it to whatever degree reflected their preferences.
[3] However, ipsative measures may, especially among testing-naïve individuals exhibiting high levels of conscientiousness and/or neuroticism, decrease test validity by discouraging response and/or encouraging non-response.
Even a previously presented guideline "Choose the answer that [best/better] describes you" may be unhelpful in such a situation to responders who worry that endorsing one item or the other will still involve stating it to be accurate or "well"-descriptive to some positive degree.
One place where this might be implemented is in reference to tests used with K-12 students in the United States, where value-added modeling of teacher performance is currently popular.
Encouraging pupils to beat their previous scores can take peer pressure out of situations and eliminates the competitive element associated with norm-based referencing.