MRKT 451 Lecture Notes - Lecture 8: Web Banner, Factorial Experiment, Random Assignment
4 Apr 2018
School
Department
Course
Professor
Document Summary
Three conditions to be met before one can determine causality. Three conditions required to determine causality: concomitant variation, extent to which x and y occur together. If x and y are significantly correlated, then one can say that there is concomitant variation. To meet these three conditions is the only way to conduct experiments. Three major types of experimental designs: quasi experiments, true experiments, advanced designs. If you randomly assign different people to different treatment conditions (ie, the different banner ads you are considering to evaluate), randomization will ensure that all types of age groups are represented in each type of ads you are evaluating. This will average out any effects of age across different treatments and thus you will be able to control (ie, rule out) any effects due to age. See slides for examples of quasi and true experimental design. It is good to know that they exist.
