Chapter 9 ML 4202 Experiment Research approach in which one variable is manipulated and the effect on another variable is observed The researcher changes or manipulates one thing called an experimental treatment independent or exploratory variable to observe the effect on something else referred to as a dependent variable Experiemental research is often referred to as casual not casual research Research designed to determine whether a change in one variable likely caused an observed change in another To demonstrate causation A likely caused B one must be able to show three things 1 Concomitant variation correlation 2 Appropriate time order of occurrence 3 Elimination of other possible causal factors Casual relationships are always inferred and never demonstrated conclusively beyond a shadow of doubt There are three types of evidence used to infer casual relationships listed above Concomitant variation Statistical relationship between two variables A and B must vary together in some predictable fashion There may be a positive or a linear relationship Appropriate Time Order of Occurrence Change in an independent variable occurred before an observed change in the dependent variable To demonstrate that A caused B one must be able to show that A occurred BEFORE B occurred Example To demonstrate that a price change had an effect on sales you must be able to show that the price change occurred before the change in sales was observed Elimination of Other Possible Causal Factors Hard to prove that something else did not cause a change in B The most difficult thing to demonstrate is that the change in B was not caused by the change in A Ex Suppose a company increased its advertising expenditures and observed an increase in the sales of its product Correlation and appropriate time order of occurrence are present But has a likely relationship been demonstrated The answer is no It is possible that the observed change in sales is due to some factor other than the increase in advertising Experimental Settings Laboratory experiments Experiments conducted in a controlled setting Advantages Being able to control extraneous causal factors temperature light humidity and so on and focus on the effect of a change in A on B However sometimes the controlled and possibly sterile environment of the lab may not be a good analog of the marketplace so the findings of the lab experiments sometimes do not hold up when transferred to the marketplace Therefore laboratory experiments are seen as having greater problems with external validity but they are being used at a greater extent today than in the past Field experiments Tests conducted outside the laboratory in an actual environment such as a marketplace They solve the problem of the realism of the environment but open up a whole new set of problems Experimental Validity Validity The degree to which an experiment actually easures what the researcher was trying to measure Internal validity Extent to which competing explanations for the experimental results observed can be ruled out External validity Extent to which causal relationships measured in an experiment can be generalized to outside persons settings and times Experimental Notation X is used to indicate the exposure of an individual or group to an experimental treatment The experimental treatment is the factor whose effects we want to measure and compare Experimental factors may be factors such as prices package designs point of purchase displays advertising approaches or product forms O for observation is used to refer to the process of taking measurements on the test units Test units are individuals groups of individuals or entities that might be the targets of a firm s marketing program O1 X O2 Extraneous Variables History Intervention between the beginning and end of an experiment of outside variables or events that might change the dependent variable Ex Early test os Prego spaghetti sauce by Campbells provides an example of a possible problem with extraneous variables Campbell executives claim that Ragu a competing brand greatly increase its advertising levels and use of cents off deals during their Prego tests Maturation Changes in subjects occurring during the experiment that are not related to the experiment but hat may affect subjects response to the treatment factor Instrument Variation Changes in measurement instruments e g interviewers or observers that might affect measurements Selection bias Systematic differences between the test group and the control group due to a biased selection process Morality Refers to the loss of test units during the course of an experiment which may result in a nonrepresentativeness Testing Effects Effect that is a by product of the research process itself Two forms 1 Main testing effects The possible effects of earlier observations on later 2 observations Ex GMAT score for the second time tends to be better than the first Interactive testing effect The effect of a prior measurement on a subject s response to a later measurement Ex If subjects are asked about their awareness of advertising for various products pre exposure measurement and then exposed to advertising for one or more of these products treatment variable post measurements would likely reflect the joint effect of the pre exposure and the treatment condition Regression to the mean Refers to the tendency of subjects with extreme behavior to move toward the average for that behavior during the course of an experiment Test units may exhibit extreme behavior because of change or they may have been specifically chosen because of their extreme behavior Controlling Extraneous Variables Randomization Random assignment of subjects to treatment conditions to ensure equal representation of subject characteristics Physical control Holding constant the value or level of extraneous variables throughout the course of an experiment Also to match respondents in regard to important personal characteristics age income lifestyle the goal is to make sure therte are no important differences between characteristics of respondents in the test and control groups Design control Use of the experimental design to control extraneous causal factors Statistical control Adjusting for the effects of confounded variables by statistically adjusting the value of the dependent variable for each treatment condition Experimental design Test in which the reseracher has control over and manipulates one or more independent variables Ex