研究方法总结.doc

第一章：
Qualitative research和quantitative research的含义、区别和共同点。
(1)含义
Qualitative research（定性研究）：指通过发掘问题、理解事件现象、分析人类的行为与观点以及回答提问来获取敏锐的洞察力。
Quantitative research（定量研究)：就是将问题与现象用数量来表示，进而去分析、考验、解释，从而获得意义的研究方法和过程。
(2）区别
1）概念不同
2)理论基础:定性研究主要是一种价值判断，它是建立在解释学、现象学和建构主义理论等人文主义的的方法论基础上；定量研究是一种事实判断，是建立在实证主义的方法论基础上。
3）特征不同:定性研究的研究者与研究对象密切接触，互相影响，研究者通过与研究对象的交往互动，通过移情作用来获取资料信息；定量研究中的研究者与研究对象相互独立，彼此分离。
4）研究工具和方法不同：定性研究（观察、访谈等方法获得描述性资料）；定量研究（量表、调查表工具进行测量，得到的资料可测量和统计)。
5)着眼点不同：定性研究着重事物质的方面；定量研究着重事物量的方面。
6）结论表述形式不同:定性结论多以文字描述为主;定量研究主要以数据、模型、图形等来表达。
（3）共同点
1）都属于量化研究。
2）都属于社会学方法.
3）定性研究是定量研究的基本前提，定量研究是定性研究的进一步深化。
3.名词解释,并举例:
Casual relationship： when variation in one factor is responsible for variation in another。（one factor influences another)
Empirical research： systematic study of relationships between scores obtained from cases on measures.
Cases： entities investigated in research。
Measures: instruments used to obtain scores from participants。
Scores （data）： numerical information about cases obtained on measures。
（3）Casual conceptual relationship： a relationship where variation(变化） in the dependent construct is responsible for variation in the dependent construct。
(4)operational relationships: variable with a measure to obtain scores from cases.
（5）empirical relationships: the correspondence between scores obtained from cases on measures。
(6)internal validity： 内部效度是因变量和自变量之间关系的确实性程度，是实验结论的真实性。
（7)internal validation(内部效化）: methods used to determine whether internal validity is likely.
（8）internal statistical validity： present when an empirical relationship is not due to chance。
（9)statistical validation： 统计效化，the use of probability theory to investigate internal statistical validity or statistical generalization validity。 利用概率论研究内部统计有效性或统计归纳有效性。
（10）statistical generalization: uses probability theory to generalize a relationship observed on a sample of cases to the relationship that applies to the broader population from which the sample was drawn。
（11）external generalization：
（12）external validation：外部效化，methods used to estimate external validity.
（13）external validity：外部效度是指在脱离研究情境后,研究结果还能成立的程度.
（14)Statistical generalization validity （统计通则化一般效度）: Present when an empirical relationship observed on a sample of cases provides a correct estimate of the relationship in the population of cases from which the sample was drawn.
(15)Construct validity：present when there is a high correspondence between case’s scores on a measure and the mental definition of the construct the measure is designed to represent。
(16）Construct validation :methods used to estimate a measure's construct validity.
2。 建构效度（construct validity）:
收敛（convergent validity）:是指运用不同测量方法测定同一特征时测量结果的相似程度，即不同测量方式应在相同特征的测定中聚合在一起。举例，比如分给10个人做你自己编制的智力测试和韦氏智力测验，如果两种测验的成绩相关比较高，说明你的智力测验聚合效度良好。
区别效度（discriminant validity)：是指当一个构念的多重指标相聚合或呼应时，则这个构念的多重指标也应与其相对立之构念的测量指标有负向相关。例如与「政治容忍」相关的多重指标应会与「政治不容忍」相关的多重指标间有负向相关.
Chapter3： Measurement foundations： Validity and Validation
1.名词解释：
(1）content validity（内容效度）: when a measure is judged to be construct valid, usually by individuals who are thought to be subject matter experts。
（2）Reliability：（信度）refers to the consistency of measurement。
(3)信度包括:
1）internal consistency： a form of reliability that addresses the consistency of scores from a set of item in a measure. （内部一致性信度)
2）interrater reliability: a form of reliability that address the consistency of scores from a set of observers。 （评分者间信度）
3）satbility reliability: a form of reliability that address the consistency of scores across time periods.（稳定信度)
(4)construct validity（建构效度）：
1）convergent validity: is present when there is high correspondence between scores from two different measures of the same construct. （收敛效度)
2）discriminant validity： is present when measures of constructs that are supposed to be independent are found to have a low correspondence. (区别效度）
3）criterion-related validity: is present when the measure of interest is related to another measure judged to be more construct valid。（效标关联效度）
Chapter4： Measurement Applications: Research Questionnaires
1.Content domain（问卷内容范围）:
1）items 2)item wording(一般准则): keep the respondent in the mind； make it simple; be specific； be honest。 3）item sequence（题项顺序） 4)scaling： open-ended response format/closed-ended response formats
2.observations:
（1）Leniency error：（宽容评价误差）present when an observer systematically inflates ratings of a group of objects。
（2)severity error： （严格错误)present when an observer systematically deflates ratings of a group of objects.
(3)central tendency error: （集中趋势的错误）present when an observer clusters responses in the middle of a scale when more variable responses should be recorded.
（4)halo error:（月晕偏差，以偏概全）present when an observer evaluates an object in an undifferentiated manner。
Chapter5: Research design foundations
1.名词解释:
(1)bias relationship（偏误关系): an observed relationship between an independent and a dependent variable that under—or overstates the casual relationship。
(2）spurious relationships： （虚假关系）an observed relationship between an independent and a dependent variable downward.指在两个没有因果关系的事件中，可能基于其他未见的交络因素，或潜在变数，显示出统计学上的相关。
(3)suppressor variables:（抑制变量）原本存在关系，由于某种因素影响,变得没有关系了。
(4)noise relationship:（干扰关系）occurs when a nuisance variable is related to dependent variable but not to the independent variables.
(5)mediators:（中介模型）comes between an independent and a dependent variable in a causal chain。
Fully mediates（完全中介）the independent variable has no direct effect on the dependent variable only through a mediator variable。
Partially mediates（部分中介）the independent variable has a direct effect on the dependent variable and a indirect effect through the mediator.
(6)Moderators（调节模型）:a variable whose values are associated with different relationship between an independent and a dependent variable。
Chapter8： Data analysis foundations
1。 levels of measurement
(1)nominal：没有大小，没有顺序，不可加减，不可乘除.宗教、国家、职业、公司。
（2）ordinal:没有大小，有顺序，不可加减，不可乘除。教育程度、职称。
(3)interval:有大小,有顺序，可加减，不可乘除。温度标、尺码、IQ。
(4)ratio：都可。身高、体重、绝对温度、工资。
2.名词解释
(1）discrete variables: have a countable number of values。 性别分为男性和女性,计算男性和女性的人数。
（2）continuous variables： are infinitely divisible between any two values。身高变化
Chapter9： Analysis applications: Describing scores on a single variable
1.名词解释
（1）skew（偏态）描述正态分布中的对称或不对称的分布关系。
Describes the symmetry or asymmetry of a distribution relative to a normal distribution。当偏态系数为0，分布对称，大于0,又偏态，小于0，左偏态.
(2）kurtosis（峰度）峰度是指次数分布曲线顶峰的尖平程度
Shows the peakedness or flatness of a distribution relative to a normal distribution。
小于0,为低阔峰，大于0为高阔峰。