|
|
Experiment Exercises Identify the type of experiment that was conducted, and the threats to external and internal validity you detect. Point out additional information you would need in order to safeguard experiment validity. 1. (Coke has introduced a new can design to the grocery stores, but the test market I describe is hypothetical). Coca Cola has come up with the new can design for its Classic Coke. To determine if this new design has a positive impact on sales, they select 12 supermarkets in Seattle for to test the new can design. They select 6 of these supermarkets at random to receive the new can design. The other 6 supermarkets continue to get the old can design. Sales are monitored in both groups of supermarkets for 6 months. 2. Some retailers are using CD-i technology to train their salespeople to recognize customer Social Styles and to respond accordingly. According to data provided by the company providing the Social Style training, sales revenues per salesperson improve by 30%. Would you need additional information to be convinced of the training program’s effectiveness?
Collecting Survey Data You will have at your disposal a variety of survey methods. They generally fall into three different categories: telephone, personal, and self-administered. There are subcategories within each survey type
Each of these methods has advantages and disadvantages. The method you end up using will depend on: a. Desire for quick response b. Desire for sample control c. Desire for interviewer supervision d. Need to show visual information e. Number of questions in the questionnaire f. Difficult or sensitive questions g. Location of sample h. Cost And, regardless of the method you will use, you will have to deal with the possibility of errors. There are 2 types of errors, random and systematic. There is very little that you can do to prevent random error other than increasing sample size. But, systematic error can be prevented, if you recognize the possibility of problems. Errors a. Population Specification Error b. Frame Error c. Selection Error d. Measurement Error e. Surrogate Information Error f. Interviewer Error g. Measurement Instrument Bias h. Processing Error i. Nonresponse Bias j. Response Bias
Let’s see which group does the best job recognizing errors in the following hypothetical situations. Please note that the people running the SUB are much more savvy than the hypothetical people described in these examples. I am using the SUB and various SPU locations simply because they are familiar to us. Please specify the type of error that was committed. 1. Sales revenues are down at the SUB (or any other business). Patrons are surveyed to determine level of satisfaction. 2. Interviews are conducted during the lunch hour because that’s when the SUB is busiest. 3. Current staff of the SUB conduct the interviews. 4. SUB staff are told to focus on "traditional" SPU students. 5. Once they decide their sample may not include the whole population, they decide to interview students at Hill Hall (since it’s such a big dorm). 6. They arrive at Hill Hall right before the start of Group, and a lot of students don’t have time to complete the survey. 7. In desperation, they leave a bunch of surveys at all the major instructional buildings at SPU with instructions for students to drop the completed surveys in a box set up in the SUB. |
|
|
