《W12 描述性分析 Descriptive Analysis》由会员分享,可在线阅读,更多相关《W12 描述性分析 Descriptive Analysis(63页珍藏版)》请在金锄头文库上搜索。
1、 管理学院 School of Management . University of Science and Technology of China 中国科学技术大学 社会科学研究方法社会科学研究方法 Social Science Research Methods 刘和福刘和福 LIU Hefu W12: 描述性分析描述性分析Descriptive Analysis 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 2 - 抽样术语抽样术语 Sampli
2、ng Terminology 总体Population; 总体要素 population element; 普查census; 样本sample; 框架frame 抽样设计基本步骤抽样设计基本步骤 Steps in Sampling design 目标总体 target population; 抽样框架 sampling frame; 抽样方法 sampling method; 样本量 sample size; 抽样误差sampling bias 抽样方法抽样方法 Sampling method 概率抽样概率抽样Probability sampling: 简单随机 simple random;
3、 系统随机 systematic random sampling; 群集抽样 Cluster sampling; 分层抽样 stratified sampling 非概率抽样非概率抽样 Nonprobability sampling: 便利抽样 convenience; 判断抽样 judgment; 配额抽样quota; 雪球抽样snowball sampling Last week 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 3 - Data E
4、xploration, Examination, and Analysis in the Research Process 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 4 - 原始数据的准备 Prepare the raw data 编码 Coding 编辑 Editing 通过SPSS打开数据文件 Open the data with SPSS 数据处理 Process the data 缺省值 Missing value 数据分析 Analyz
5、e the data Preparation of Data 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 5 - 编码手册 You need a code book/scoring guide 每个样本需有一个编码号 You give ID number for each case (NOT real identification numbers of your subjects) if you use paper survey. 如果是在线调研,
6、需有特定手段来确定样本 If you use online survey, you need something to identify your cases. 运用Excel, SPSS或其它软件进行数据录入 You also can use Excel, SPSS or other software to do data entry. Before the data entry 5 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 6 - Examp
7、le of a code book 6 编码手册需界定 A code book is about how you code your variables. What are in code book? 1.变量名称 Variable names 2. 对应的应对选项的内涵Values for each response option 3. 如何对变量重新编码 How to recode variables 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: -
8、 7 - 不合法变量名示例及分析不合法变量名示例及分析 变量名变量名 错误原因错误原因 NAME_123 变量名超过变量名超过8个字符(个字符(11.0及以前版本)及以前版本) NAME_1. 以句点结尾以句点结尾 _12NAME 第一个字符不是字母第一个字符不是字母 NA ME?1 变量名中有?和空格变量名中有?和空格 ALL 用了不能用作变量的关键词用了不能用作变量的关键词 * *输入不合法的变量名时,会自动显示出错信息输入不合法的变量名时,会自动显示出错信息. . 变量的相关操作变量的相关操作 刘和福 中国科学技术大学管理学院 School of Management, Universi
9、ty of Science and Technology of China E-Mail: - 8 - 导入导入Excel 数据文件数据文件Import an Excel file 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 9 - 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 10 - 数据视角数据视
10、角 Data view 数据录入 The place to enter data 列:变量; 行:样本Columns: variables Rows: records SPSS interface 10 变量视角变量视角Variable view 变量的录入The place to enter variables 所有变量特征的列表List of all variables Characteristics of all variables 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of
11、 China E-Mail: - 11 - Enter data in SPSS 16.0 列:变量 行:样本 数据视角 11 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 12 - Enter variables 1. 变量视角 2. 变量名 3. 变量类型:字符;数据 4. 变量描述 12 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of
12、 China E-Mail: - 13 - Enter variables 13 2. 右击 1. 双击改变 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 14 - 输入每个变量的数值和标注Key in values and labels for each variable 数据排序 Sort the data 分析每个变量的频次 Run frequency for each variable 查看是否有错误数值 Check see if you
13、have variables with wrong values. 查看缺失值 Check missing values and physical surveys if you use paper surveys, and make sure they are real missing. 重新将字符变量变成数值变量 Sometimes, you need to recode string variables into numeric variables Clean data after import data files 14 刘和福 中国科学技术大学管理学院 School of Manage
14、ment, University of Science and Technology of China E-Mail: - 15 - Sorting the data 1. 选择变量 2. 点击 3. 点击” 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 16 - Missing values 缺失值 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technolog
15、y of China E-Mail: - 17 - 检查缺失值Check for missing value 用频次分析分类变量 Use Frequency for categorical variable 用描述统计分析指标变量 Use Descriptive Stat. for measurement variable 分类变量 For categorical variables: 如缺失值小于5%, 运用List-wise选项 如缺失值大于于5%,增加一个新类别 If =5%, 指标变量变量 For measurement variables: 如缺失值小于5%, 运用List-wis
16、e选项 如介于5%和15%间,用替代法来替代缺失值 如大于15%,剔除样本或变量 Missing values-methods 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 18 - SPSSMissing value Transform: 缺失值缺失值Missing Value -均值替代均值替代 series mean -近点有效数值的均值mean of nearby point -邻近点有效数值的中位数值median of nearby poi
17、nt -线性插值法linear interpolation -线性拟合方式Linear trend at point 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 19 - Missing value 1. 选择有缺失值的变量 3. 选一方法 4. 点击 2. 如替代原有变量,可将名字改成原有变量名,后点击”change” 刘和福 中国科学技术大学管理学院 School of Management, University of Science and
18、 Technology of China E-Mail: - 20 - 20 4. 点击 2. 新变量名 3. 设定公式 1. 计算生产新变量 5. 创建语法文件 语法文件可记录相关公式,并提供修改的便利方式 Compute Variable 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 21 - SPSS- Analyze Analyze: Descriptive Statistics Compare means Correlate Data R
19、eduction Scale Regression 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 22 - Descriptors of Research Design- Purpose of the Study Descriptive Who What Where When How much Not WHY & HOW 刘和福 中国科学技术大学管理学院 School of Management, University of Science and
20、Technology of China E-Mail: - 23 - Weber, L., Mayer, K.J., and Macher, J.T. (2011), An Analysis of Extendibility and Early Termination Provisions: The Importance of Framing Duration Safeguards, Academy of Management Journal, Vol. 54, 182-202 刘和福 中国科学技术大学管理学院 School of Management, University of Scie
21、nce and Technology of China E-Mail: - 24 - Sample: Angst et al. 2010 Angst, C.M., Agarwal, R., Sambamurthy, V., Kelley, K., 2010. Social Contagion and Information Technology Diffusion: The Adoption of Electronic Medical Records in US Hospitals. Management Science 56, 1219-1241 刘和福 中国科学技术大学管理学院 Scho
22、ol of Management, University of Science and Technology of China E-Mail: - 25 - Tallon, P.P., and Pinsonneault, A. (2011), Competing Perspectives on the Link between Strategic Information Technology Alignment and Organizational Agility: Insights from a Mediation Model, MIS Quarterly, Vol. 35, 463-48
23、4. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 26 - Descriptive Statistics_ Frequencies 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 27 - Descriptive- Frequencies 刘和福 中国科学技术大学管理学院 School of Managem
24、ent, University of Science and Technology of China E-Mail: - 28 - Statistics身高9991160.639.2659160.100161.08.403870.6235.226.077.405.15569.1136.4205.5160478.0154.700160.100166.500ValidMissingNMeanStd. Error of MeanMedianModeStd. DeviationVarianceSkewnessStd. Error of SkewnessKurtosisStd. Error of Ku
25、rtosisRangeMinimumMaximumSum255075PercentilesMean 均数均数 Std. Error of Mean 标准误标准误 Median 中位数中位数 Mode 众数众数 Std. Deviation 标准差标准差 Variance 方差方差 Skewness 峰度系数峰度系数 Std. Error of Skewness 峰度系数的标准误峰度系数的标准误 Kurtosis 偏度系数偏度系数 Std. Error of Kurtosis 偏度系数的标准误偏度系数的标准误 Range 全距全距 Minimum 最小值最小值 Maximum 最大值最大值 Su
26、m 合计合计 Percentiles 25 25%位数位数 50 50%位数位数 75 75%位数位数 Descriptive- Frequencies 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 29 - Frequency table 29 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 30 - 描述
27、性分析描述性分析 Descriptive- descriptives 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 31 - 选择标准数据 描述性分析描述性分析 Descriptive- descriptives 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 32 - 同时进行多变量的基础性描述统计分析,只
28、需将它们同时从左边选入右边框即可. Descriptive- descriptives 对原有变量标准化之后形成的新变量 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 33 - Descriptive- Explore 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 34 - 增加 “INDS”变量后,结果将
29、分成两部分 - 一组对应制造业,一组对应服务业 Descriptive- Explore 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 35 - 在”crosstabs”分析中,可输入一个或更多变量作为频数表中的“行”,其它变量作为“列”The “Crosstabs” dialog box lets you enter one variable (or more) as “rows” in a frequency table, and another
30、 variable as the “columns” in the same table. Descriptive- Crosstabs 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 36 - Descriptive- Crosstabs 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 37 - Descri
31、ptive- Crosstabs 基于双变量分组后每组的样本数The total number of firms that have each individual combination of the various levels of the two variables. 基于某一变量分组的样本数和占比The far-right column shows you the total for each diagnosis, as a percent of all firms. 基于另一变量分组的样本数和占比The bottom-most row slices the data the oth
32、er way, showing you the total for each type of industry, as a percent of all firms. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 38 - This can be used to test the non-response bias Zhang, Q. Y., Vonderembse, M. A., & Lim, J. S. (2003). Manufacturing
33、 flexibility: defining and analyzing relationships among competence, capability, and customer satisfaction. Journal of Operations Management, 21(2), 173-191. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 39 - Presentation and Discussion 刘和福 中国科学技术大学管
34、理学院 School of Management, University of Science and Technology of China E-Mail: - 40 - SPSS的均值检测的均值检测(Compare Means) MeansMeans:平均数:平均数 OneOne- -samples T testsamples T test:单一样本检测:单一样本检测 IndependentIndependent- -samples T testsamples T test:独立样本检测:独立样本检测 PairedPaired- -samples T testsamples T test
35、:成对样本检测:成对样本检测 OneOne- -way ANOVAway ANOVA: 单因子变异数分析单因子变异数分析 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 41 - 平均数平均数(Means) 分析基于产业类别分类的特定构念的均值差异分析基于产业类别分类的特定构念的均值差异 Analyze the difference of mean of a specific construct based on Industry type 刘和福 中
36、国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 42 - 平均数平均数(Means) 1. 将感兴趣的变量选入因变量框 Move the variables of interest A to the Variables list box. 2. 将分组变量选入自变量框To subdivide the dataset into groups, move the grouping variable INDS to the Grouping Variable(s) l
37、ist box. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 43 - 平均数平均数(Means) The mean of A is close between INDS1 and INDS0 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 44 - 单一样本平均数检测单一样本平均数检测(One-Sampl
38、e T Test) 比较样本均值与某一特定值的差异。特定值通常是总体的均值compares the mean score of a sample to a known value. Usually, the known value is a population mean. Hypotheses: Null: 样本均值和总体均值不存在样本均值和总体均值不存在显著差异显著差异There is no significant difference between the sample mean and the population mean. Alternate:样本均值和总体均值不样本均值和总体均
39、值不存在显著差异存在显著差异There is a significant difference between the sample mean and the population mean. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 45 - 3. 置信区间 单一样本平均数检测单一样本平均数检测(One-Sample T Test) 1. 目标变量 2. 特定值 刘和福 中国科学技术大学管理学院 School of Management, Un
40、iversity of Science and Technology of China E-Mail: - 46 - 单一样本平均数检测单一样本平均数检测(One-Sample T Test) 两者之间有显著差异,即样本均值3.5869 显著高于总体均值3. 样本均值为3.588, 略高于总体均值3 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 47 - 独立样本检测独立样本检测 IndependentIndependent- -samples T
41、 testsamples T test 比较某一变量在两组数据中的均值compares the mean scores of two groups on a given variable Hypotheses: Null: 均值在两组中不存在显著均值在两组中不存在显著差异差异The means of the two groups are not significantly different. Alternate:均值在两组中存在显均值在两组中存在显著差异著差异The means of the two groups are significantly different. 刘和福 中国科学技术
42、大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 48 - 独立样本检测独立样本检测 IndependentIndependent- -samples T testsamples T test 2. 分组变量 0: Manufacturing 1: Service 1. 分析变量 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 49
43、 - 独立样本检测独立样本检测 IndependentIndependent- -samples T testsamples T test “Manufacturing”组高于 “Service” 组. the Levenes Test for Equality of Variances: Significant (.05) = significantly different. Not significant= are equal we have met our second assumption. Independent Samples T Test: 如方差显著差异,则看第一行数据 如方差
44、无显著差异,则看底行数据. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 50 - Paired Samples T Test 分析两个变量的均值。比较每个样本的两个变量间的差异,分析其差异相比0值是否有显著差异 Hypothesis: Null: 两个变量的均值没有显著差两个变量的均值没有显著差异异There is no significant difference between the means of the two variables. A
45、lternate:两个变量的均值有显两个变量的均值有显著差异著差异There is a significant difference between the means of the two variables. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 51 - Paired Samples T Test 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Techn
46、ology of China E-Mail: - 52 - A1的均值更高 两个变量的相关性显著the correlation between the two variables is significant 两个变量的差异的描述统计 the descriptive statistics for the difference between the two variables 如果显著值小于0.05,则表示差异显著 If the significance value is less than .05, there is a significant difference. Paired Sam
47、ples T Test 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 53 - OneOne- -Way ANOVA Way ANOVA 单因素方差分析单因素方差分析 分析多组自变量分组的均值是否有显著差异is used to determine whether there are any significant differences between the means of three or more independent (unrelated
48、) groups. Hypotheses: Null: 组内均值没有显著差异组内均值没有显著差异There are no significant differences between the groups mean scores. Alternate:组内均值有显著差异组内均值有显著差异There is a significant difference between the groups mean scores. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Ma
49、il: - 54 - OneOne- -Way ANOVA Way ANOVA 单因素方差分析单因素方差分析 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 55 - OneOne- -Way ANOVA Way ANOVA 单因素方差分析单因素方差分析 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 56
50、- 如果 the Levenes Test 显著,两个变量的方差差异显著。如不显著,两者方差则相同is significant, the two variances are significantly different. It is not significant, the two variances are approximately equal. 两组间不存在显著差异。因此,表明A在不同所有制类型中不存在显著差异 There is not a significant difference between the two groups. Therefore, we can say that
51、 there is a significant difference between the four ownership types on A. OneOne- -Way ANOVA Way ANOVA 单因素方差分析单因素方差分析 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 57 - 可分析每组样本的均值的显著差异明细of groups are significantly differen. 对有显著差异的值会通过*显示 SPSS notes
52、a significant difference with an asterisk (*). 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 58 - The role of “compare mean” The statistical difference between the results of a survey in which the sample includes only those who responded (answered th
53、e questions) and a survey that would include those who failed to respond. Bias that occurs when those who respond tend to answer questions in a way that misrepresents the truth consciously (deliberate falsification) or unconsciously (unconscious misrepresentation) 刘和福 中国科学技术大学管理学院 School of Manageme
54、nt, University of Science and Technology of China E-Mail: - 59 - ANOVA_ Non-response bias_ Example Wu, L.-Y., 2010. Applicability of the resource-based and dynamic-capability views under environmental volatility. Journal of Business Research 63, 27-31 刘和福 中国科学技术大学管理学院 School of Management, Universi
55、ty of Science and Technology of China E-Mail: - 60 - The role of “compare mean” Test hypotheses Hypothesis 1a: On the Clan culture dimension, POEs will score higher than FCBs. Hypothesis 1b: On the Clan culture dimension, pre-reform SOEs will score higher than FCBs. Hypothesis 1c: On the Clan cultu
56、re dimension, POEs will score higher than desired SOEs. Ralston, D. A., Terpstra-Tong, J., Terpstra, R. H., Wang, X. L., and Egri, C., (2006), Todays state-owned enterprises of China: Are they dying dinosaurs or dynamic dynamos?, Strategic Management Journal, 27, 825-843. 刘和福 中国科学技术大学管理学院 School of
57、Management, University of Science and Technology of China E-Mail: - 61 - Descriptive Research Ralston, D. A., Terpstra-Tong, J., Terpstra, R. H., Wang, X. L., and Egri, C., (2006), Todays state-owned enterprises of China: Are they dying dinosaurs or dynamic dynamos?, Strategic Management Journal, 2
58、7, 825-843. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 62 - Descriptive Research Ralston, D. A., Terpstra-Tong, J., Terpstra, R. H., Wang, X. L., and Egri, C., (2006), Todays state-owned enterprises of China: Are they dying dinosaurs or dynamic dynamos?, Strategic Management Journal, 27, 825-843. 刘和福 中国科学技术大学管理学院 School of Management, University of Science and Technology of China E-Mail: - 63 -