HOMEWORK NINE: PART E: EXPERIMENTAL DESIGN Topics 37-42: Experimental Design Isabel Cabrera EDCI-6300.61 Foundations of Research in Education Dr. Alberto Jose Herrera The University of Texas at Brownsville March 24, 2012 1 HOMEWORK NINE: PART E: EXPERIMENTAL DESIGN 2 Exercise on Topic 37: True Experimental Designs 1. What is the purpose of an experiment? The purpose of an experiment is to explore the cause-and-effect relationships. 2. How can researchers ensure that there are no biases in the assignment of participants to groups in an experiment? Researchers are assured that there are no biases in the assignment by assigning participants at random to groups. 3. In a diagram for an experimental design, the symbol “O” stands for what? The “O” stands for observation of measurement. 4. In a diagram for an experimental design, the symbol “X” stands for what? The “X” stands for experimental treatment. 5. What is the name of the potential problem caused by the pretest in the pretest-posttest randomized control group design? This problem is called the pretest sensitization or the reactive effect in testing because participants will find out what material will be covered and to what extent. 6. What is the name of the true experimental design that has no pretests? The experimental design that has no pretests is called the posttest-only randomized control group design. 7. What is the potential drawback to the Solomon randomized four-group design? The only drawback is that a researcher must begin with a reasonably large pool of participants so that when they are divided into four groups, each of the groups will have a sufficient number to yield reliable results. 8. True experimental designs are easy to spot because they are all characterized by what? True experimental designs are easy to spot because they are all characterized by random assignment to treatments. Exercise on Topic 38: Threats to Internal Validity HOMEWORK NINE: PART E: EXPERIMENTAL DESIGN 3 1. What is the name of the threat that indicates that taking a pretest may affect performance on a posttest? Testing is an internal threat that indicates that taking a pretest many affect the performance on a posttest. 2. Suppose an experimental group is being taught letters of the alphabet as a treatment. At about the same time, the students are watching an educational program on television, from which they learn the names of the letters. What is the name of the threat that this problem illustrates? History is a threat when other environmental influences have on the participants between the pretest and the posttest. 3. If observers are more tired and less astute when making posttest observations than when making pretest observations, what threat is operating? Instrumentation is another threat when observers are more tired and less astute when making posttest observations than when making pretest observations. 4. What is the name of the threat posed by nonrandom assignment of participants to experimental and control groups? The researcher is using intact groups when groups are not assigned at random. 5. If infants naturally improve in visual acuity and thus perform better at the end of an experiment than at the beginning, what threat is operating? Maturation is the threat because the participants are maturing during the period between the pretest and the posttest. 6. Under what circumstance will statistical regression operate? Statistical regression is another threat that occurs only if participants are selected on the basis of their extreme low scores. 7. How can researchers overcome all the threats to internal validity? All threats to internal validity can be overcome by using a true experimental design, in which participants are assigned at random to experimental and control conditions. Exercise on Topic 39: Threats to External Validity HOMEWORK NINE: PART E: EXPERIMENTAL DESIGN 4 1. Which type of validity deals with the question of whether a researcher can generalize with confidence to a larger population in a natural setting? External validity is when a researcher can generalize results with confidence to a larger population in a natural setting. 2. Which type of validity deals with whether the treatment is directly responsible for any changes observed in the experimental setting? Internal validity is concerned when treatment is directly responsible for any changes observed. 3. What is the name of the threat that warns researchers to be careful in generalizing the results to a population when an experiment is conducted on a nonrandom sample? This threat is called the selection bias because when participants are not chosen at random the experiment is bias and results cannot be generalized and applied to all populations. 4. Suppose a random sample of workers in a factory is exposed to five different reward systems, with each system being used for one month. What is the name of the threat that reminds researchers that the results may not generalize to the population of workers if the population is to be exposed to only the last reward system tried in the experiment? Multiple-treatment interference is a threat that occurs when a group of participants is given more than one treatment. 5. Suppose an experimental classroom has research observers present at all times. What is the name of the threat that reminds researchers that the results may not generalize to other classrooms without observers? The reactive effects of experimental arrangements is when an experimental setting is different from the natural setting in which the population usually operates, the effects that are observed in the experimental setting may not generalize to the natural setting. 6. If a pretest causes a change in participants’ sensitivity to a treatment, what threat is operating? The reactive effect of testing also known as pretest sensitization is a threat that refers to the possibility that the pretest might influence how the participants respond to the experimental treatment. Exercise on Topic 40: Pre-experimental Designs HOMEWORK NINE: PART E: EXPERIMENTAL DESIGN 5 1. Are pre-experimental designs valuable for identifying cause-and-effect relationships? Pre-experimental designs are of very limited value for investigation cause-and-effect relationships because of their poor internal validity. 2. Suppose a researcher administered a new program to all students in a school. At the end of the school year, the researcher administered a standardized test to the students in the school as well as to students in another school who were serving as a control group. Is the comparison of the average scores for the two groups of students useful for determining the effects of the program? No the comparison would not be useful for determining the effects of the program. The researcher used a static-group comparison design where he/she used two groups of participants that were not assigned at random. The groups would be called intact groups because they were already pre-existing groups. The internal threat would be selection. Furthermore, without a pretest, the researcher does not have a basis for knowing whether the two groups were similar at the beginning of the experiment. 3. What is the name of the pre-experimental design used in Question2? The pre-experimental design used in the scenario would be the static-group comparison design. 4. If a researcher gives a pretest on knowledge of child abuse to a group of social workers, then gives them a series of seminars on child abuse followed y a posttest, what is the name of the pre-experimental design the researcher used? This scenario is using the one-group pretest-posttest design. There is nothing wrong with this design if it is for instructional purposes only. 5. Is the design used in Question 4 useful for determining cause-and-effect relationships? It is of no value of cause-and-effect relationships because it fails to make comparisons. Exercise on Topic 41: Quasi-Experimental Designs 1. What is the name of the design diagrammed immediately below? OXO HOMEWORK NINE: PART E: EXPERIMENTAL DESIGN O 6 O It is called the quasi-experimental design because it has two intact groups not assigned at random. 2. In the design shown in Question 1, what indicates that the participants were not random to the groups? Not assigned by random as indicated by the dashed line. 3. If a researcher uses matching to form the two groups in the design in Question1, would the resulting experiment be superior to a true experiment? It is far from ideal but is superior to the pre-experimental designs, which are not directly interpretable in terms of causal effects. 4. What is a major advantage of the equivalent time-samples design? The major advantage is that the participants are both the control and experimental group and are identical in all ways at the beginning of the experiment. 5. What is a major disadvantage of the equivalent time-samples design? A major disadvantage of this design is the strong possibility of multiple-treatment interference. 6. In Psychology, what is an ABAB design? In Psychology, a single-group design in which treatments are alternated is often called the ABAB design. Exercise on Topic 42: Confounding in Experiments 1. A confound is a source of confusion regarding what? A confound is a source of confusion regarding the explanation for a given difference. 2. What is the name of the effect that refers to the possibility that the control group might become aware of its “inferior” status and respond by trying to outperform the experimental group? HOMEWORK NINE: PART E: EXPERIMENTAL DESIGN 7 The John Henry effect refers to the possibility that the control group might become aware of its “inferior” status and respond by trying to outperform the experimental group. 3. What is the formal name of what is characterized as the “attention effect” in this topic? The Hawthorn effect is also known as the attention effect because there were two intertwined explanations for the experimental results on the industrial psychology study, one being the amount of lighting and the other being the special attention participants received during the study. 4. In addition to an experimental and a traditional control group (with no treatment), what other group can be used to control for the Hawthorne effect? To control the Hawthorn effect, some researchers use three groups: 1) an experimental group 2) a control group that receives attention 3) a control group that receives no special attention. 5. The term “placebo effect” refers to what tendency? The placebo effect refers to the tendency of individuals whom feel they are improving simply because they know they are being treated. 6. In what type of experiment do neither the participants nor the individual dispensing the drug know which is the active drug and which is the placebo? In a double blind experiment, neither the participants nor the individual dispensing the drug know which is the active drug and which is the placebo.