水稻二化螟灾变风险分析多水平模型研究
|
摘要
本研究是由国家自然科学基金项目“水稻主要害虫灾变风险分析多水平模型研究”(项目编号为:30370914)资助,以水稻二化螟为例,对影响水稻二化螟发生的关键因子进行分析、辨识,并建立了影响水稻二化螟发生的多层线性模型。
本研究从生态系统的观点出发,考虑水稻二化螟发生处于不同地区,受不同地区地理、生态因子影响,使得水稻二化螟发生因素存在层次结构。本文作者应用多层线性模型统计分析技术,参考已发表文献资料以及现有数据,建立了水稻二化螟发生和地理、生态及气象因子关系的两水平线性模型,以期找出影响水稻二化螟发生关键因子、探讨不同生态系统层次对水稻二化螟发生的影响。本研究主要结果如下:
1.建立了水稻病虫、气象资料数据库,及其数据管理、查询系统
(1) 水稻病虫资料数据库的建立本研究分析所用病虫资料来源于浙江省34个县市水稻主要病虫区划资料统计表1981～2000二十年间数据(部分资料有缺失,实际用的只有26个县市资料)。对病虫数据资料进行整理并输入计算机,建立Access数据库并按病虫害种类建立若干相应病虫害发生情况数据表,包括稻飞虱、二化螟、三化螟、稻纵卷叶螟、稻瘟病、纹枯病、白叶枯病。
(2) 分析所需气象、地理、生态资料数据库的建立分析所需的气象资料来源于浙江省气象局1981～2000二十年间气象数据资料,将气象资料按旬进行整理并输入计算机,建立Access数据库并分别建立各个地区逐句气象资料数据表,包括温雨系数、温湿系数、最高温度、最低温度、平均温度、雨日、湿度、日照时数、降雨量、生态因子。数据库的建立为进一步运用HLM软件进行多层线性模型分析提供基础数据。
(3) 建立了水稻病虫和气象资料数据管理、查询系统,可自动地从数据库重检索数据,以用于水稻病虫多水平建模分析。
2.水稻二化螟两水平线性模型研究
应用多层线性模型原理,首先采用DPS统计软件筛选主要因子,然后应用HLM6.0分析软件,根据筛选的因子构建水稻二化螟两水平线性模型。经DPS软件分析、筛选,模型选入的第一层次的影响因子主要有:4月下旬最低温度、4月下旬湿度、5月上旬降雨量、5月中旬温雨系数,以及上年二化螟晚稻发生程度。应用HLM6.0分析软件组建的回归模型中的5个预测变量对二化螟的发生程度都存在显著影响:4月下旬最低温度(P=0.043)、4月下旬湿度(P=0.026)、5月上旬降雨量(P=0.011)、5月中旬温雨系数(P=0.007)以及上年晚稻发生程度(P=0.008)。
模型选入的第二层,即地理、生态因子有化肥用量、丘陵比及林地比。结合第一层次的影响因子,构成的模型对早稻二化螟的影响,其显著水平大多小于0.01(即作用极显著)。
根据上述因子建立的多层线性模型,对早稻水稻二化螟发生与生态、气象及地理因子关
This research is subsidized by National Natural Science Fund, "Research on Multilevel Models of Risks Analysis for Rice Insect Pests" (No.30370914) . This research analyzed and distinguished the key factors that influence the occurrence of striped stem borer, Chilo suppressalis. Applying the Hierarchical Linear Modeling (HLM) statistical technology, a two-level HLM model of the outbreaking of the striped stem borer over a large area was established. The mode was used to explain the key factors which influence the occurrence of Chilo suppressalis and probe into the influence of each ecosystem level upon the occurrence of Chilo suppressalis from the point of view of the Agricultural Ecosystem. The main results of this research are as follow:1. Establishment of information management system for rice insect pests and meteorological factors(1) Establishment of database for rice insect pests. The data of rice insect pests are come from 34 countries of Zhejiang province during the 20 years from 1981 to 2000 in this research. We sort out the data and put data into the computer. Microsoft Access was used to develop database management system for rice insect pests.(2) Establishment of database for meteorological, geographic, ecological factors. The data of meteorological factors are come from weather bureau of Zhejiang Province during the 20 years from 1981 to 2000 in this research. We tidy the data and put them into the computer. The database included the temperature-rainfall coefficient, the hygrothermal coefficient, maximum temperature, minimum temperature, average temperature, rain day, humidity, sunlight time, rainfall and other ecological factors.(3) We build the information management system for management and inquire of pest and meteorological data. The system can help us to get data from database convenient for the modeling.2. Study on the two-level hierarchical linear modeling for Chilo suppressalisConsidering the hierarchical structure of agricultural ecological system and applying the HLM statistical technology, a two-level HLM model for Chilo suppressalis was built. In the first level, 4 climatic factors and one pest factor was introduced into the model. They are minimum temperature in the late April, humidity in late April, rainfall in the early May, temperature-rainfall coefficient in the middle May and the occurrence degree of last-season rice. The level-1 regression model included 5 key factors which have significantly influence on the occurrence of Chilo suppressalis, such as minimum temperature in the late April (P=0.043 ) , humidity in late April (P=0.026) , rainfall in the late May (P=0.011) , temperature-rainfall coefficient in the middle
May (.P=0.007) , the occurrence degree of last-season rice (jP=0.008) .In the second level, three factors, the level of chemical fertilizer, percentage of hills cover, and percentage of forest land was introduced into the model. Combining the influence of the level-1 key factors, the level-2 key factors have significantly influence (/K0.01) on the occurrence of Chilo suppressalis of early-season rice in the final two-level linear model.Based on the hierarchical linear model, author carried out the analysis of the relationship between the occurrence of Chilo suppressalis and ecological, meteorological and geographic factors, and made clear several points as below:First, the results indicate that the most important factors that influence the occurrence of striped stem borer are still from the individual meteorological factors, which take about 56% of the total variance. But the inter-regional level factors can not be ignored, which take about 44% of the total variance.Second, between the intra-regional key factors (the level-1 key factors), the number of pest sources is the most significant, and the temperature-rainfall coefficient in middle May is the most significant in meteorological factors.Finally, between the inter-regional key factors (the level-2 key factors ) , the area percentage of hills land is likely most significant (the frequency is 3/6) , and the other factor is the percentage of forest land (the frequency is 2/6) , and the level of chemical fertilizer (the frequency is 2/6) . The result indicated that the regional difference is mainly come from geographic factors.
本研究从生态系统的观点出发,考虑水稻二化螟发生处于不同地区,受不同地区地理、生态因子影响,使得水稻二化螟发生因素存在层次结构。本文作者应用多层线性模型统计分析技术,参考已发表文献资料以及现有数据,建立了水稻二化螟发生和地理、生态及气象因子关系的两水平线性模型,以期找出影响水稻二化螟发生关键因子、探讨不同生态系统层次对水稻二化螟发生的影响。本研究主要结果如下:
1.建立了水稻病虫、气象资料数据库,及其数据管理、查询系统
(1) 水稻病虫资料数据库的建立本研究分析所用病虫资料来源于浙江省34个县市水稻主要病虫区划资料统计表1981～2000二十年间数据(部分资料有缺失,实际用的只有26个县市资料)。对病虫数据资料进行整理并输入计算机,建立Access数据库并按病虫害种类建立若干相应病虫害发生情况数据表,包括稻飞虱、二化螟、三化螟、稻纵卷叶螟、稻瘟病、纹枯病、白叶枯病。
(2) 分析所需气象、地理、生态资料数据库的建立分析所需的气象资料来源于浙江省气象局1981～2000二十年间气象数据资料,将气象资料按旬进行整理并输入计算机,建立Access数据库并分别建立各个地区逐句气象资料数据表,包括温雨系数、温湿系数、最高温度、最低温度、平均温度、雨日、湿度、日照时数、降雨量、生态因子。数据库的建立为进一步运用HLM软件进行多层线性模型分析提供基础数据。
(3) 建立了水稻病虫和气象资料数据管理、查询系统,可自动地从数据库重检索数据,以用于水稻病虫多水平建模分析。
2.水稻二化螟两水平线性模型研究
应用多层线性模型原理,首先采用DPS统计软件筛选主要因子,然后应用HLM6.0分析软件,根据筛选的因子构建水稻二化螟两水平线性模型。经DPS软件分析、筛选,模型选入的第一层次的影响因子主要有:4月下旬最低温度、4月下旬湿度、5月上旬降雨量、5月中旬温雨系数,以及上年二化螟晚稻发生程度。应用HLM6.0分析软件组建的回归模型中的5个预测变量对二化螟的发生程度都存在显著影响:4月下旬最低温度(P=0.043)、4月下旬湿度(P=0.026)、5月上旬降雨量(P=0.011)、5月中旬温雨系数(P=0.007)以及上年晚稻发生程度(P=0.008)。
模型选入的第二层,即地理、生态因子有化肥用量、丘陵比及林地比。结合第一层次的影响因子,构成的模型对早稻二化螟的影响,其显著水平大多小于0.01(即作用极显著)。
根据上述因子建立的多层线性模型,对早稻水稻二化螟发生与生态、气象及地理因子关
This research is subsidized by National Natural Science Fund, "Research on Multilevel Models of Risks Analysis for Rice Insect Pests" (No.30370914) . This research analyzed and distinguished the key factors that influence the occurrence of striped stem borer, Chilo suppressalis. Applying the Hierarchical Linear Modeling (HLM) statistical technology, a two-level HLM model of the outbreaking of the striped stem borer over a large area was established. The mode was used to explain the key factors which influence the occurrence of Chilo suppressalis and probe into the influence of each ecosystem level upon the occurrence of Chilo suppressalis from the point of view of the Agricultural Ecosystem. The main results of this research are as follow:1. Establishment of information management system for rice insect pests and meteorological factors(1) Establishment of database for rice insect pests. The data of rice insect pests are come from 34 countries of Zhejiang province during the 20 years from 1981 to 2000 in this research. We sort out the data and put data into the computer. Microsoft Access was used to develop database management system for rice insect pests.(2) Establishment of database for meteorological, geographic, ecological factors. The data of meteorological factors are come from weather bureau of Zhejiang Province during the 20 years from 1981 to 2000 in this research. We tidy the data and put them into the computer. The database included the temperature-rainfall coefficient, the hygrothermal coefficient, maximum temperature, minimum temperature, average temperature, rain day, humidity, sunlight time, rainfall and other ecological factors.(3) We build the information management system for management and inquire of pest and meteorological data. The system can help us to get data from database convenient for the modeling.2. Study on the two-level hierarchical linear modeling for Chilo suppressalisConsidering the hierarchical structure of agricultural ecological system and applying the HLM statistical technology, a two-level HLM model for Chilo suppressalis was built. In the first level, 4 climatic factors and one pest factor was introduced into the model. They are minimum temperature in the late April, humidity in late April, rainfall in the early May, temperature-rainfall coefficient in the middle May and the occurrence degree of last-season rice. The level-1 regression model included 5 key factors which have significantly influence on the occurrence of Chilo suppressalis, such as minimum temperature in the late April (P=0.043 ) , humidity in late April (P=0.026) , rainfall in the late May (P=0.011) , temperature-rainfall coefficient in the middle
May (.P=0.007) , the occurrence degree of last-season rice (jP=0.008) .In the second level, three factors, the level of chemical fertilizer, percentage of hills cover, and percentage of forest land was introduced into the model. Combining the influence of the level-1 key factors, the level-2 key factors have significantly influence (/K0.01) on the occurrence of Chilo suppressalis of early-season rice in the final two-level linear model.Based on the hierarchical linear model, author carried out the analysis of the relationship between the occurrence of Chilo suppressalis and ecological, meteorological and geographic factors, and made clear several points as below:First, the results indicate that the most important factors that influence the occurrence of striped stem borer are still from the individual meteorological factors, which take about 56% of the total variance. But the inter-regional level factors can not be ignored, which take about 44% of the total variance.Second, between the intra-regional key factors (the level-1 key factors), the number of pest sources is the most significant, and the temperature-rainfall coefficient in middle May is the most significant in meteorological factors.Finally, between the inter-regional key factors (the level-2 key factors ) , the area percentage of hills land is likely most significant (the frequency is 3/6) , and the other factor is the percentage of forest land (the frequency is 2/6) , and the level of chemical fertilizer (the frequency is 2/6) . The result indicated that the regional difference is mainly come from geographic factors.
引文
[1] 张雷,雷雳著.多层线性模型应用.教育科学出版社,2003.1.
[2] 雷雳,张雷.多层线性模型的原理及应用.首都师范大学学报(社会科学版),2002.2.110-114.
[3] 李小松等译.多水平分析模型.四川科学技术出版社,2000.24-30.
[4] 张志涛,辛文.二化螟种群动态及管理研究进展昆虫知识 2001,04.
[5] 张风雨,王海东.多水平模型及其在人口科学研究中的应用.中国人口科学.1995,6(51):1-7.
[6] 盖笑松,张向葵.多层线性模型在纵向研究中的应用.心理科学.2005,28(2):429-431.
[7] 何忠全,张志涛,陈志谊.我国水稻病虫害综合防治技术研究新进展.植物保护.2000,1:30-32.
[8] 唐启义,李绍石,病虫测报应验程度判定模式,病虫测报,1988.(1):1-5.
[9] 肖悦岩.预测预报准确度评估方法的研究.植保技术与推广.1997,17(4),3-6.
[10] 周惟敏,刘家骧,李月明,陈阳.农作物病虫害预报质量评定方法初探.植保技术与推广.1996,16(4),13-14.
[11] 何清波,苏炳华,钱亢.医学统计学及其软件包.上海:上海科学技术文献出版社,2002.191-193.
[12] 张波.Kappa一致性检验在流行病学中的应用实例.宁夏医学院学报.1995,17(4),336-337.
[13] 冯怡,张燕敏,薛筷明.Kappa一致性检验在护理研究中的应用.2004,18(3A),462-463.
[14] 唐启义,冯明光.实用统计分析及其DPS数据处理系统.北京:科学出版社,2002.110-111.
[15] 南京农学院主编.昆虫生态及预测预报.农业出版社,1985.424-428.
[16] Hai-Bo Chert, Xiang-Yu Zeng, et al. The advantages and agreement and sensibility analysis of motor dysfunction rating scale for Parldnson's disease. Chinese Journal of Clinical Rehabilitation. 2004, 8(4), 762-763.
[17] Stout JC, Ready RE, Grace J, et al. Factor analysis of the frontal systems behavior scale (FrSBe) . Assement 2003, 10, 79-85.
[18] Anthony S B, Stephen W R. Hierarchical Linear Models: Applications and Data Analysis Methods, Sage Publication, 1992.
[19] Rasbash J, Browne W G H, Yang M P I, Healy et al. Reference of Users Guide: Mlwin, London, Institute of Education, 1999. 1-150.
[20] 刘红云,孟庆茂.教育和心理学研究中的多层线性模型.心理科学进展.2002.10(2).213-219.
[21] 吴晓云,曾庆,周燕荣.多水平模型的最新进展.数理医药学杂志.2003.16(2).152-154.
[22] Goldstein H. Multilevel Statistical Models. Second Edition.London: Edward Arnold. 1995.
[23] Langford I, Leyland A, Rasbash J, et al. Multilevel Modelling of The Geographical Distributions of Diseases. Journal of Royal Statistical Society, Series C, 1999, 48 Part 2: 253-268.
[24] Barbosa M F, Goldstein H. Discrete Response Multilevel Models for Repeated Measures;an application to voting intentions data. Quality and Quantity. 2000, 34(3): 323-330.
[25] 杨珉.多元分析的发展一多水平模型简介.中国卫生统计.1994.11(5) 32-35.
[26] Goldstein,H. (1986) Multilevel mixed linear model analysis using iterative generalised least squares. Biometrika, 73: 43-56.
[27] 蒋学辉,章强华.植保技术与推广,1997,17(6):15-17.
[28] 王传全,王翔,施文等.安徽农业大学学报,1997,24(4):336-339.
[29] 汪恩国.植物保护,1999,25(1):14-17.
[30] 吴蔚文,蔡坤喜.西南农业大学学报,1992,14(1):19-21.
[31] 章士美,陈东.江西农业大学学报,1993,15(1):33-35.
[32] 雷国明.植保技术与推广,1995,(3):4-6.
[33] 马飞,程遐年.害虫预测预报研究进展.安徽农业大学学报.2001,28(1):92-97.
[34] 汪四水,张孝羲.害虫预测建模中的因子综合.南京农业大学学报 2000,23(2):35-38.
[35] 谷晓平,田兰,李波.水稻病虫害预测系统的研究.贵州气象.2002.26(2):3-7.
[36] 葛道阔,金之庆等.水稻栽培模拟优化决策系统(RCSODS)中的水稻病虫害预测模型及其应用.江苏农业学报.2003,19(2):70-74.
[37] 张孝羲,周立阳.害虫预测预报的理论基础.昆虫知识.1995,32(1):55-60.
[38] 陈钦勇.多层线性模型在大学英语影响因素中的初步应用研究.江西师范大学.学位论文.
[39] 桂子山.班主任之师生互动风格与小学生自我概念的关系研究.硕士学位论文.
[40] 吴曙雯.中国南方稻区褐飞虱灾变分析与预警系统的研究及应用.浙江大学博士学位论文.
[41] 秦厚国,罗任华,叶正襄,汪笃栋,付志飞.二化螟大发生原因及控制对策.华东昆虫学报.2005,14(1):91-93.
[42] 蔡之华,洪要文,曹炳宏,杨杰,白善军,王林.沿淮稻区二化螟重发原因与综防对策.安徽农学通报,2002,8(4):52.
[43] 黎贤伟.水稻二化螟持续大发生的原因及综合治理对策.湖北植保,2005年第1期.
[44] 《中国水稻病虫害综合防治策略与技术》,农业出版社.256-275.
[45] 《昆虫生态及预测预报》,南京农学院主编,农业出版社.300.
[46] 粮油作物病虫鼠害预测预报浙江省农作物病虫测报站,上海科学技术出版社.
[47] Jason W. Osbome. Advantages of Hierarchical Linear Modeling. Practical Assessent, Research, and Evaluation, 7(1) .
[48] J. Douglas Willms. Basic Concepts in Hierarchical Linear Modeling with Applications for Policy Analysis,
[49] Willms, J. D., & Paterson, L. (1995) . A multilevel model for community segregation. Journal of Mathematical Sociology, 20(1) 23-40.
[50] BEck, A. S., & Raudenbush, S.W. (1992) . Hierarchical linear models for social and behavioral research: Applications and data analysis methods. Newbury Park, CA: Sage.
[51] Raudenbush, S. W., & Bryk, A. S. (1986) . A hierarchical model for studying school effects. Sociology of education. 59,1-17.
[52] Raudenbush, S. W., & Willms, J. D. (1995 ) . The estimation of school effects. Journal of Educational and Behavioural Statistics, 20, 307-335.
[2] 雷雳,张雷.多层线性模型的原理及应用.首都师范大学学报(社会科学版),2002.2.110-114.
[3] 李小松等译.多水平分析模型.四川科学技术出版社,2000.24-30.
[4] 张志涛,辛文.二化螟种群动态及管理研究进展昆虫知识 2001,04.
[5] 张风雨,王海东.多水平模型及其在人口科学研究中的应用.中国人口科学.1995,6(51):1-7.
[6] 盖笑松,张向葵.多层线性模型在纵向研究中的应用.心理科学.2005,28(2):429-431.
[7] 何忠全,张志涛,陈志谊.我国水稻病虫害综合防治技术研究新进展.植物保护.2000,1:30-32.
[8] 唐启义,李绍石,病虫测报应验程度判定模式,病虫测报,1988.(1):1-5.
[9] 肖悦岩.预测预报准确度评估方法的研究.植保技术与推广.1997,17(4),3-6.
[10] 周惟敏,刘家骧,李月明,陈阳.农作物病虫害预报质量评定方法初探.植保技术与推广.1996,16(4),13-14.
[11] 何清波,苏炳华,钱亢.医学统计学及其软件包.上海:上海科学技术文献出版社,2002.191-193.
[12] 张波.Kappa一致性检验在流行病学中的应用实例.宁夏医学院学报.1995,17(4),336-337.
[13] 冯怡,张燕敏,薛筷明.Kappa一致性检验在护理研究中的应用.2004,18(3A),462-463.
[14] 唐启义,冯明光.实用统计分析及其DPS数据处理系统.北京:科学出版社,2002.110-111.
[15] 南京农学院主编.昆虫生态及预测预报.农业出版社,1985.424-428.
[16] Hai-Bo Chert, Xiang-Yu Zeng, et al. The advantages and agreement and sensibility analysis of motor dysfunction rating scale for Parldnson's disease. Chinese Journal of Clinical Rehabilitation. 2004, 8(4), 762-763.
[17] Stout JC, Ready RE, Grace J, et al. Factor analysis of the frontal systems behavior scale (FrSBe) . Assement 2003, 10, 79-85.
[18] Anthony S B, Stephen W R. Hierarchical Linear Models: Applications and Data Analysis Methods, Sage Publication, 1992.
[19] Rasbash J, Browne W G H, Yang M P I, Healy et al. Reference of Users Guide: Mlwin, London, Institute of Education, 1999. 1-150.
[20] 刘红云,孟庆茂.教育和心理学研究中的多层线性模型.心理科学进展.2002.10(2).213-219.
[21] 吴晓云,曾庆,周燕荣.多水平模型的最新进展.数理医药学杂志.2003.16(2).152-154.
[22] Goldstein H. Multilevel Statistical Models. Second Edition.London: Edward Arnold. 1995.
[23] Langford I, Leyland A, Rasbash J, et al. Multilevel Modelling of The Geographical Distributions of Diseases. Journal of Royal Statistical Society, Series C, 1999, 48 Part 2: 253-268.
[24] Barbosa M F, Goldstein H. Discrete Response Multilevel Models for Repeated Measures;an application to voting intentions data. Quality and Quantity. 2000, 34(3): 323-330.
[25] 杨珉.多元分析的发展一多水平模型简介.中国卫生统计.1994.11(5) 32-35.
[26] Goldstein,H. (1986) Multilevel mixed linear model analysis using iterative generalised least squares. Biometrika, 73: 43-56.
[27] 蒋学辉,章强华.植保技术与推广,1997,17(6):15-17.
[28] 王传全,王翔,施文等.安徽农业大学学报,1997,24(4):336-339.
[29] 汪恩国.植物保护,1999,25(1):14-17.
[30] 吴蔚文,蔡坤喜.西南农业大学学报,1992,14(1):19-21.
[31] 章士美,陈东.江西农业大学学报,1993,15(1):33-35.
[32] 雷国明.植保技术与推广,1995,(3):4-6.
[33] 马飞,程遐年.害虫预测预报研究进展.安徽农业大学学报.2001,28(1):92-97.
[34] 汪四水,张孝羲.害虫预测建模中的因子综合.南京农业大学学报 2000,23(2):35-38.
[35] 谷晓平,田兰,李波.水稻病虫害预测系统的研究.贵州气象.2002.26(2):3-7.
[36] 葛道阔,金之庆等.水稻栽培模拟优化决策系统(RCSODS)中的水稻病虫害预测模型及其应用.江苏农业学报.2003,19(2):70-74.
[37] 张孝羲,周立阳.害虫预测预报的理论基础.昆虫知识.1995,32(1):55-60.
[38] 陈钦勇.多层线性模型在大学英语影响因素中的初步应用研究.江西师范大学.学位论文.
[39] 桂子山.班主任之师生互动风格与小学生自我概念的关系研究.硕士学位论文.
[40] 吴曙雯.中国南方稻区褐飞虱灾变分析与预警系统的研究及应用.浙江大学博士学位论文.
[41] 秦厚国,罗任华,叶正襄,汪笃栋,付志飞.二化螟大发生原因及控制对策.华东昆虫学报.2005,14(1):91-93.
[42] 蔡之华,洪要文,曹炳宏,杨杰,白善军,王林.沿淮稻区二化螟重发原因与综防对策.安徽农学通报,2002,8(4):52.
[43] 黎贤伟.水稻二化螟持续大发生的原因及综合治理对策.湖北植保,2005年第1期.
[44] 《中国水稻病虫害综合防治策略与技术》,农业出版社.256-275.
[45] 《昆虫生态及预测预报》,南京农学院主编,农业出版社.300.
[46] 粮油作物病虫鼠害预测预报浙江省农作物病虫测报站,上海科学技术出版社.
[47] Jason W. Osbome. Advantages of Hierarchical Linear Modeling. Practical Assessent, Research, and Evaluation, 7(1) .
[48] J. Douglas Willms. Basic Concepts in Hierarchical Linear Modeling with Applications for Policy Analysis,
[49] Willms, J. D., & Paterson, L. (1995) . A multilevel model for community segregation. Journal of Mathematical Sociology, 20(1) 23-40.
[50] BEck, A. S., & Raudenbush, S.W. (1992) . Hierarchical linear models for social and behavioral research: Applications and data analysis methods. Newbury Park, CA: Sage.
[51] Raudenbush, S. W., & Bryk, A. S. (1986) . A hierarchical model for studying school effects. Sociology of education. 59,1-17.
[52] Raudenbush, S. W., & Willms, J. D. (1995 ) . The estimation of school effects. Journal of Educational and Behavioural Statistics, 20, 307-335.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |