ABSTRACT
This chapter focuses on where attempts to draw inferences from experiments can flounder. It distinguishes four types of validity: statistical conclusion validity, internal validity, constructs validity, and external validity. The chapter addresses these questions: how can things go wrong in attempting to draw an inference from an experiment and what are the major types of experimental designs. The simplest experimental design is one involving only a single factor. Designs in which the various groups are defined by the particular level of a single factor they experience are referred to as one-way designs, because the groups differ in one way or along one dimension. In cases of designs with multiple factors, designs differ in which combinations of levels of the different factors are used. In most cases, all possible combinations of levels of the factors occur. After examining threats to each type of validity, the chapter concludes by giving an overview of the designs and analyses.
