基于动态权重的百分位数指标在期刊评价实践中的应用研究

【目的】 将动态权重百分位数(Dynamic Weight Percentile Rank Score,DPRS)指标应用于跨学科期刊的评价,并与现有的文献计量指标进行对比。【方法】 以《期刊引证报告》数据库中5个学科的期刊作为评价对象,对期刊的DPRS、百分位数(Percentiles Rank Scores,PRS)、5年影响因子、篇均被引频次、平均百分位、来源标准化篇均影响力(Source Normalized Impact per Paper,SNIP)等指标进行相关性分析,在此基础上针对DPRS和PRS 2个指标在期刊评价中的效果与区别进行了实证研究。【结果】 DPRS与篇均被引频次、5年影响因子、平均百分位、SNIP指标均高度相关,可以消除学科不同带来的引文差异性,并且能够充分综合期刊在各百分位区间“相对数量”和“相对质量”2个方面的信息,且具有比其他类似指标更高的区分度和精准度。【结论】 DPRS指标能合理综合论文数量和被引频次分布2个方面的信息,可以实现跨学科期刊的评价,并且能够揭示期刊学术影响力的更多细节。     

The effectiveness of supplemental phonics instruction employing constant time delay instruction for struggling readers

This study examined the provision of supplemental instruction using constant time delay for struggling readers who had fallen behind in developing word‐blending skills. Additionally, the study examined the utility of adding a material reward contingency to the instructional program with a contingency derived from percentile‐shaping procedures. Procedures were compared in a withdrawal design applied across blocks of letters with word blending as the target response. Participants exhibited substantial learning gains when provided constant time delay instruction in phonics skills as compared to baseline demonstrating clear experimental control. Results for the constant time delay plus reward contingency conditions were clearly differentiated from baseline but undifferentiated from the other instructional conditions. The implications of the findings for intervention for struggling readers and future research relevant to more comprehensive reading interventions are discussed.     

Tutorial: Calculating Percentile Rank and Percentile Norms Using SPSS

Practitioners can benefit from using norms, but they often have to develop their own percentile rank and percentile norms. This article is a tutorial on how to quickly and easily calculate percentile rank and percentile norms using SPSS, and this information is presented for a data set. Some issues in calculating percentile rank and percentile norms are discussed, and a comparison between percentile rank and percentile norms is presented. In summary, percentile rank and percentile norms can be quickly and easily developed by using SPSS.     

Additional Revised Push-Up Test Norms for College Students

This study was designed to examine the impact of silhouette randomization on the responses to rating scales developed to rate the perceived current and ideal body shape, as well as body dissatisfaction. Seventy students (30 men and 40 women), ages 18 to 43 (M ± SD = 22.1 ± 5.7) years, completed the Stunkard, Sorensen, and Schulsinger (1983) Figure Rating Scale twice, approximately 2 weeks apart in a randomized, counterbalanced order. On one occasion, the traditional scale was completed, and another scale with a randomized set of silhouettes was completed. Results indicated a significant relationship between the traditionally and randomly presented scales for perceived current and ideal body shapes and for body dissatisfaction (all p < .01). Wilcoxon signed-ranks test revealed no significant mean rank difference between the two scales for perceived current and ideal body shapes and body dissatisfaction (all p > .05). Randomization of figures does not appear to have an impact on results from the Figure Rating Scale, and the sequentially presented scale appears to be appropriate for body image assessment.     

基于GAMLSS模型的我国青少年体质健康评价指标百分位数分布曲线及参考值

目的：建立我国10~17岁青少年体质健康评价指标百分位数分布曲线及参考值。方法：通过文献资料、专家访谈、问卷调查、测试法对我国16个省份32个城市的288所中小学5 048名儿童青少年的各项体质健康指标进行数据采集,其中男性2 779名,女性2 269名。测试内容依次由反映身体形态、心肺耐力、肌肉力量、速度、灵敏、柔韧性的6个一级指标和9个二级指标构成,运用基于位置、尺度、形状的广义可加模型（GAMLSS）,在模型拟合中考虑年龄、性别等因素,建立青少年体质健康指标的参考百分位数曲线。结果表明,我国10~17岁青少年体质健康评价指标总体随年龄的增长显著提升,各相同年龄段男生的体质健康指标普遍优于女生,各项测试指标的变化呈现出一定程度的波动性特征且与年龄显著相关（P<0.05）。研究尝试构建了我国10~17岁青少年体质健康指标P5-P95百分位分布曲线与参考值。结论：研究所构建的百分位数曲线与参考值对于我国中小学生体质健康评价标准与常规模型参照体系的修订、体质健康数据库建立、学生身体素质的长期系统监测以及人才的选拔、培养与分流具有一定的借鉴价值。     

新能源示范城市建设对区域环境污染治理的影响

新能源示范城市试点是促进中国深入推进生态文明建设的一项重要举措,客观评估其对环境污染的治理效果,对于这一试点政策的进一步推广具有重要意义。本文基于中国261个地级市2002—2016年的面板数据,应用倾向得分匹配法和双重差分法评估了新能源示范城市建设对环境污染的治理效应,并在此基础上引入偏差校正的非参数百分位Bootstrap法,对效应的内在机制进行了解析。研究结论表明：第一,新能源示范城市建设会显著减少废水和废气的排放,进而有效降低城市环境污染;第二,新能源示范城市可以通过能源技术创新效应、产业结构优化效应和财政资金扶持效应3条中介路径实现对环境污染的治理;第三,新能源示范城市建设过程中产生的污染治理效应存在异质性,主要体现为：一方面,经济增长水平较低的城市在进行示范城市建设的过程中往往会产生更大的污染治理效应;另一方面,示范城市的人力资本水平越高,其越能有效地抑制污染物的排放。本研究为中国进一步推进新能源示范城市建设战略、提高新能源示范城市建设质量、总结与推广新能源城市建设经验以及深入研发与推广新能源技术提供了依据。     

百分位数指标的修正及评价分析

针对相对影响和百分位数指标的局限,提出修正百分位数指标。从引文量分布特征、同区引文量差异、引文量与发文时长关联性等方面,对相对影响和百分位数指标进行了对比分析。在此基础上,将相对影响和百分位数指标融合并引入时效性参数,构建了修正百分位数指标并以图书情报类核心期刊及高校图书馆作为测评对象进行了案例分析。结果表明,修正百分位数指标可将位于同一百分位区间但发文时长及引文量不同的论文纳入到同一尺度下进行对比评价,测评结果对于引文数量特征和内在规律性的反映更为客观。     

影响因子百分位在学术期刊跨学科评价中的局限性

【目的】 探讨影响因子百分位(Journal Impact Factor Percentile,JIFP)在学术期刊跨学科评价中的局限性。【方法】 通过JCR数据库获取相关数据,结合JIFP计算公式进行分析,揭示JIFP的局限性。【结果】 JIFP的数值过于依赖学科规模;JIFP过分强调影响因子排序而掩盖了影响因子数值的差异;同一期刊在不同学科的JIFP差异明显。【结论】 JIFP应用于学术期刊跨学科评价具有明显局限性,应慎重使用JIFP进行期刊跨学科评价。     

关于百分位数精细计算方法的探讨

在区分百分位数类型、分析原始数据的基础上,对百分位数精细计算方法及其应满足的性质作了探讨,给出了一套根据原始数据计算的秩算法。     

原始资料百分位数计算方法探究

通过比较四种类型的百分位次的求法和基于四种类型的6种百分位数算法,探究适宜的原始资料百分位数计算方法,从统计学角度,最终建议使用第四种方法计算原始资料的百分位数.