This study evaluates the difficulties in concept reasoning, changes in response patterns, and item misconception hydrolysis patterns using Rasch modeling. Data were collected through the development of 30 distractor-based diagnostic test items, measuring ten levels of conceptual reasoning ability in three types of salt hydrolysis compounds: Na[subscript 5]P[subscript 3]O[subscript 10], NaOCl and (NH[subscript 4])[subscript 2]SO[subscript 4]. These 30 written test items were completed by 849 students in Gorontalo, Indonesia. The findings show empirical evidence of the reliability and validity of the measurement. Further analysis found that the students' reasoning difficulty levels of the concept of saline solutions were varied; the calculation of saline solution's pH level is the most difficult construct to reason. In particular items, changes in response patterns were found; the misconception curve showed a declining trend and disappeared along with the increase of comprehension along the spectrum of students' abilities. The item misconceptions pattern was found repeatedly in similar items. This finding strengthens the conclusion that resistant misconceptions potentially tend to cause students' conceptual reasoning difficulties and are difficult to diagnose in conventional ways. This study contributes to developing ways of diagnosing resistant misconceptions and being a reference for teachers and researchers in evaluating students' chemical conceptual reasoning difficulties based on Rasch modeling.