Unverified50% confidenceBenchmarkExact time
Behaviorally Anchored Rating Scales (BARS) can improve inter-rater reliability from 0.4–0.6 for unstructured assessments to 0.7–0.9.
1
Sources
50%
Confidence
Long-term
Relevance
8/15/2026
First Seen
Sources
Related Claims
Unverified功效分析需要预先指定显著性水平α（通常0.05）、期望统计功效（通常0.8或0.9）和预期效应量三个核心参数66% similarUnverified效应量的判断标准是：0.2属于小效应，0.5属于中等效应，0.8以上算大效应64% similarUnverified模板匹配的置信度阈值通常取0.8～0.95来判断是否找到目标64% similarUnverified评估基准测试质量的关键指标是假阳性率和假阴性率，理想基准应将两者控制在较低水平63% similarUnverified当所有商家评分集中在4.5-4.9的狭窄区间时，评分系统的信息熵急剧下降，用户从评分中获得的有效决策信息趋近于零63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/754405API
curl https://kongchang.com/api/v1/knowledge/claims/754405MCP
get_claim(id=754405)