Allard G, Faust D
University of Rhode Island, USA.
Assessment. 2000 Jun;7(2):119-29. doi: 10.1177/107319110000700203.
Given the paucity of previous research, we examined the occurrence of scoring error on widely used objective personality tests and examined its possible relation to two factors: scoring procedure complexity (SPC) and commitment to accuracy (CTA). We double-checked the scoring of three tests (MMPI, Beck Depression Inventory, Spielberger State/Trait Anxiety Inventory) across three settings. Each of the tests were misscored at a surprisingly high rate in at least one setting, and some such errors altered major interpretive implications. Tests of higher SPC showed greater error rates, but high CTA greatly reduced the occurrence of error across levels of SPC. Unexpected sources of error were also uncovered, such as commercial computer scoring errors and disagreement in scoring standards among test publishers. Practical suggestions for improving scoring accuracy are offered.
鉴于以往研究的匮乏，我们考察了广泛使用的客观人格测试中评分错误的发生率，并考察了其与两个因素的可能关系：评分程序复杂性（SPC）和对准确性的重视程度（CTA）。我们在三种情境下对三项测试（明尼苏达多项人格调查表、贝克抑郁量表、斯皮尔伯格状态/特质焦虑量表）的评分进行了复核。至少在一种情境下，每项测试的评分错误率都高得出奇，而且有些此类错误改变了主要的解释意义。SPC较高的测试显示出更高的错误率，但高CTA在SPC的各个水平上都大大降低了错误的发生率。还发现了意想不到的错误来源，如商业计算机评分错误以及测试出版商之间评分标准的不一致。文中提供了提高评分准确性的实用建议。
