This study aims to describe the quality of the quantitative items in multiple choice tests in the form of validity, reliability, difficulty index, different power and deceptive effectiveness of the Final Semester Tests for Class X Indonesian Subjects at SMA Negeri 1 Sokaraja 2016 academic year 2017. This research is quantitative descriptive. Data collection methods used in the study are documentation. Subjects in this study were all items of multiple choice questions End Semester Repetition Even Class X Indonesian Subjects at SMA Negeri 1 Sokaraja in 2016-2017. Multiple choice questions End Semester Repetition Even Indonesian Language Lessons for 2016-2017 school year are said to be inappropriate, so improvements need to be made. Such impropriety can be seen from the five quality criteria items. In terms of validity, after being analyzed in content and construct, having an alpha value of 1.00 is declared valid, but at the level of difficulty the problem is in the unfavorable category. This is because the item has a level of difficulty in easy criteria (0.71-1.00), namely 24 (48%) items, 15 (30%) items categorized as moderate (0.31-0.70) and 11 ( 22%) items are categorized as difficult (0.00-0.30). Then, the different power of multiple choice items has a low power difference (<0.20). This is because there are 37 (74%) items included in the low category (<0.20). The item cannot be accepted or with revisions must be made, then there are 9 items that are in the medium category (0.20-0.30) with a percentage of 18%. The item can be received with repairs, and 4 or 8% of items included in the category are high (> 0.30) so that the questions can be accepted. Furthermore, the effectiveness of the deception of multiple choice items does not function properly because there are 8 (16%) items that have met the criteria or there is no need to revise the alternative answers, while the items that need to be revised are 42 (84%) items. . Last seen in terms of reliability level, the problem has a high level of reliability (> 1.00).