Exam 2: Collecting Data Sensibly

In a well-designed experiment, the factors are confounded whenever possible.

The ZZZ chain of motels has a standard method of constructing their rooms to maximize the ease of parking for its customers. The rooms are arranged in adjacent buildings so that each customer can park outside the rented room. The layout for one of the with 48 rooms located along a famous highway is diagrammed below: Route 66 The manager would like to survey customers in 12 of his rooms (one randomly selected customer for each room selected in the sample) to assess their satisfaction with the motel services. The surveys will be placed on the customers' beds before they check in to the motel. In order to make the directions easy to follow, he elects to use systematic sampling. (a) Explain how you would use random numbers to set up the systematic sampling process. (b) Write a short paragraph for the maids that helps them carry out your method in part (a).

Luis being a college student has undertaken some research. The results have shown that people who buy expensive cars are less likely to die from a flu-related illness during the following year than those who do not buy. However, recent research has shown that people who buy expensive cars are also less likely to die from any cause during the following year than those who don't buy. Are there any confounding variables that make it difficult to justify a cause-and-effect conclusion? ​

The two paragraphs below discuss aspects of two studies, each of which exhibit a bias. For each study, decide whether the problem is selection bias, response bias, or nonresponse bias, and in a few sentences explain why you chose your answer. (a) One part of the Nurses' Health Study is concerned with possible causes of skin cancer. Nurses were asked about different behaviors and aspects of their health when they entered the study. Then, the nurses were given the questionnaire again if they were diagnosed with cancer. When the questionnaires were analyzed, the investigators discovered that after the nurses were diagnosed with cancer they tended to report a reduced ability to tan. It is thought that the shift in reporting might be caused by an awareness of their diagnosis. (b) One part of the Demographic and Health Surveys Program is concerned with measures of malnutrition. Investigators measure physical aspects of growing children, and attempt to document the physical characteristics of a population at different ages. Sadly, in some countries many children die early, and thus a bias is introduced in the study when the investigators can not collect the data from the deceased children.

A common practice of teachers is to have students exchange their quizzes and grade each others. In addition to decreasing the teacher's work load, the reduced time between quiz and feedback is thought to be a plus for learning. Your U.S. History teacher, aware of your statistical prowess, has asked you to design an experiment to test this theory. You have decided to use the mid-term exam (not graded by students) as your response measure. Your history teacher has three classes, one early in the morning, one at noon, and one late in the afternoon. Each class contains 30 students. (a) Describe the treatments you will use in your experiment (b) One possible confounding variable is the time of day, since students may be more alert at certain times of the day than at other times. Describe a method would you use to control this variable? (Unfortunately you cannot change the student schedules!) (c) Do you feel the results of your experiment could be generalized to your statistics class? Why or why not?

In competitive sports, video recorders have been used more frequently in recent years. The idea behind the recorder is that coaches can replay training sessions for more effective feedback to the athlete. Some people believe video recording may make the athletes more nervous and actually decrease their performance. You have been asked to design an experiment to address this issue for competitive high school tennis players. You have decided to use the accuracy of tennis serves as your response variable, and the number of successful serves out of 100 as your performance measure. The subjects for your experiment are 60 high school male competitive tennis players of varying ability who have volunteered for the experiment. (a) Describe the treatments in your experiment (b) One possible confounding variable is the experience levels of the players. Explain how you would control this variable? (c) Can the results of this experiment could be generalized to all male tennis players? Why or why not?

Studies made by Kate as a university student have shown that people over age 60 who donate blood twice a year are less likely to die from a heart disease during the following year than those who do not donate. However, recent research has shown that people over age 60 who donate blood twice a year are also less likely to die from any cause during the following year than those who don't donate. Are there any confounding variables that make it difficult to justify a cause-and-effect conclusion?

Selection bias can occur if volunteers only are used in a study.

The following paragraph describes an actual study. After reading the description, determine whether the study is an observational or experimental study. Justify your answer with specific references to the information in the study. "Before the opening of the new International Airport and the termination of the old airport, children near both sites were recruited into aircraft-noise groups (aircraft noise at present or pending) and control groups with no aircraft noise and closely matched for socio-economic status. A total of 326 children (mean age = 10.4 years) took part in three data-collection waves, one before and two after the switch-over of the airports. After the switch to the new airport, long-term memory and reading were found to be impaired in the noise group at the new airport, and improved in the formerly noise-exposed group at the old airport."

Random assignment to treatments will guarantee groups that are exactly alike for experimental purposes.

Clusters are non-overlapping subgroups of a population that have been identified as homogeneous.

A placebo is identical in appearance to the treatment of interest, but contains no active ingredients.

Replicating in an experiment means that the number of subjects is greater than 1.

By definition, a simple random sample of size n is any sample that is selected in a manner to guarantee every individual in the population has an equal chance of selection.

Stratified sampling is a sampling method that in no way involves simple random sampling.

One of the tasks of wildlife biologists is to estimate bird populations. One method for doing this is to walk a "transect," typically a long randomly selected rectangle, and count the number of particular species that are seen. During hunting season the biologists wear brightly colored clothes as a protective measure. A theory, known as the "species-confidence hypothesis," predicts that birds may react to these brightly colored clothes. A robin, for example, when confronted with a biologist wearing an orange color may be less afraid than if the biologist were wearing a brown color. The idea is that the orange color gives the robin some "confidence" that the biologist is another (really big!) bird. This is a difficulty because a robin that is less afraid of the biologist may be easier to see and therefore "count" because the robin lets the biologist get closer. This would result in too high an estimate of the robin population. To test this theory a biologist performed a study in her area, wearing a bright orange vest or a brown vest while working. After training herself to accurately estimate distances she indicated on her notes how close the robins would let her get before they flew away (the "approach distance"), using this as a measure of their "confidence." She randomly selected her starting points in each trip through the transects and also carefully noted the weather conditions, specifically the wind speed. (a) What is the explanatory variable (factor) for this experiment? (b) What is the response variable for this experiment? (c) After completing the study the biologist was examining her results and noticed that on those days she wore the brightly-colored vest the wind tended to be of greater velocity. How does this information affect the interpretation of the results of the experiment?

The likelihood that an animal will successfully avoid a predator's attack should increase if the presence of the predator is detected. Animals are, of course, known to warn others of a predator's presence--the well-known prairie-dog warning call is an example. Some aquatic animals are known to send chemical signals. For example the crayfish and Iowa darter may excrete ammonium from their gills or possibly in their urine during periods of heightened metabolic activity. In an experiment to see if red-legged frogs (Rana aurora) may send or respond to chemical signals, specimens in Oregon were collected as embryos and raised in an aquarium. As tadpoles they were separated into two pairs of aquaria. The "upper tank" in each pair was connected to the "lower tank" to create a (very!) small river. A wooden heron model was placed in one of the upper tanks and moved around for 30 seconds to simulate a predator attack. Both the sender (upper tank) and receiver (lower tank) animals were monitored for added activity (increased movement) indicating antipredator behavior. (a) What is the explanatory variable (factor) for this experiment? (b) What is the response variable for this experiment? (c) During the course of the experiment the investigators were very careful with the wooden heron model not to come in contact with the glass of the aquaria or make noise in any other way. If they had been unsuccessful and their wooden heron made significant amounts of noise, how would that affect the interpretation of the results?

A study is an observational study if the investigator observes the behavior of a response variable when one or more factors are manipulated.

We have distinguished two types of studies: observational and experimental. Briefly explain the essential difference(s) between these two types of study.