Suppose that a producer of single-room air conditioners wishes to test four prototype air conditioning units. The dependent variable is the number of days an air conditioner will function properly before its motor needs major repair. In this case the producer is interested in only one factor, the type of air conditioner, at four different levels. However, the manufacturer suspects that the type of use might affect the time until major repair. Specifically, these air conditioning units are used in three environments:
(1) In residential homes located in northern climates, where they are used only on an occasional basis during the summer months;
(2) In residential homes located in moderate climates, where they are used frequently during the summer months and seldom during other seasons of the year; and
(3) In residential homes in southern climates, where they are used frequently throughout the year except during the cooler winter months. The producer suspects that these different environments may tend to obscure real differences among the four types of air conditioners.
To conduct this experiment, the producer has allocated 20 air conditioners of each type. Provided that the air conditioner producer is interested primarily in how the type of unit affects the time until major repair, how can the company control for the type of environment? Assume that approximately 10% of all single-room air conditioners produced by this company are used in homes located in northern climates, 25% are used in homes located in moderate climates, and 65% are used in homes located in southern climates. Explain, in detail, how the producer should set up this experiment?