We report a study of 10-14 year old children's use of additive strategies while solving ratio and proportion tasks. Rasch methodology was used to develop a diagnostic instrument that reveals children's misconceptions. Two versions of this instrument, one with "models" thought to facilitate proportional reasoning and one without were given to a sample of 303 children. We propose a methodology for examining systematically the pupils' additive errors, their effect on ratio reasoning and how contingent on "model" presentation this is. First we provide a measure on which pupils, item-difficulty and additive errors can be located. We then construct a new measure, which we name "tendency for additive strategy". Finally, we find that the presence of "models" affects this new measure and draw inferences for choices of items in assessment and teaching. (Contains 3 tables.) [For complete proceedings, see ED500857.]