实证研究——实验总结详解.pptx

总结（1） 实验法 Experimental design is a procedure for devising an experimental setting such that a change in a dependent variable may be attributed solely to the change in an independent variable. 定义：是一种通过系统操作一个或多个自变量以确定此变化是否对其他一个或者多个因素造成影响的方法。 Ch 5 2 自变量 Independent variables are those variables which the researcher has control over and wishes to manipulate. Category vs. Continuous variables (problems with median split) For example: level of ad expenditure; type of ad appeal; price; product features, etc. Ch 5 3 因变量 Dependent variables are those variables that we have little or no direct control over, yet we have a strong interest in. Examples would be customer satisfaction, attitude, purchase intention, actual behavior, choice, etc. Ch 5 4 额外变量 Extraneous variables are those variables that may have some effect on a dependent variable yet are not independent variables. Extraneous variables must be controlled through proper experimental design. Ch 5 5 额外变量 常有的额外变量 实验者效应：由研究者引起的对被试行为产生的无意的作用 双盲程序：与被试接触的主试及被试本身都不知道研究假设 噪音 温度 被试的个体差异 额外变量 控制额外变量的技术 直接消除，如控制噪音 保持额外变量恒定，比如性别对行为的影响，可以全采用男被试或者女被试 测量被试属性，然后按属性匹配的方法 随机化——最常见的控制额外变量的方法 把它们作为自变量或者协变量 实验法 Does the independent variable cause changes in the dependent variable? Ch 5 8 思考困难 Choice deferral or compromise cause 实验法 必须对变量下一个操作定义 实验法 Characteristics manipulation of a variable solves the direction of causality problem randomly assign to conditions solves the 3rd variable problem Ch 5 10 实验法 操纵和选择自变量 自变量的每个值被称为水平（level) 要操纵变量，被试必须被分配到该变量的两个水平上 （最简单实验：实验组vs控制组） 与每个水平相联系的条件叫做该实验的一个处理（treatment) 被试可以接受一个处理（被试者间设计）或所有处理（被试者内设计） 如果实验做得不好, then garbage in, garage out) 实验法也会出现一些问题 The manipulations may not be strong enough to allow for a meaningful test of the theory; A manipulation designed to vary one particular independent variable may also affect other independent variables in the study, (3) Other latent variables (i.e., confounding variables) may be inadvertently affected by the manipulations, Ch 5 14 实验法也会出现一些问题 (4)