Topic 41: TRUE EXPERIMENTAL DESIGNS
- Briefly describe an experimental problem for which a researcher probably would not be able to assign participants at random to conditions.
A researcher might want to experiment with a new teaching method. The teaching new method that the teacher came up to teach algebra to 6th graders. There are two classes for sixth grade, and they include sixth-grade blue and sixth-grade red. The two classes have an equal number of students. The easiest way for the researcher to experiment would be to start by with sixth-grade students in the blue class as the treatment group. For this experiment, the control group would be the sixth-grade students in the red class. Therefore, this experiment would be nonequivalent group designs. This is because the researcher does not assign students to the two classes’ randomly. This means that there is a very high probability that there exists a considerable difference between the students in sixth-grade blue and the sixth-grade red.
The differences in this experimental design are likely to result from several factors. The parents might have selected the class in which they want their child to belong. Also, the school principal might have assigned the students to the two classes based on various factors like academic achievement and even that student’s personality. Further, there is a variation in the teacher’s teaching styles and also the classroom environment. This would trigger notable variations on the student’s achievement concerning the new method of teaching sixth-grade algebra. Performance differences between the two classes might have been caused by the factors rather than the new teaching method.
Topic 42: THREATS TO INTERNAL VALIDITY
- Suppose a researcher gave a series of wellness workshops over six months and then determined that five of the participants had quit smoking during the six months. The researcher’s interpretation was that the workshop caused a decrease in smoking. Is this interpretation flawed? Explain.
The researcher’s interpretation is flawed because other factors could have led the five research participants to quit smoking. It is more likely that the person had a company of smoking friends with whom they were smoking together. Within the six months, the friends relocated, or something happened, and the friendship was over. This left the person with no friends to smoke with. Such a change in friendship would have triggered the person to quit smoking. Also, some people have regular medical checkups. Within the period of research, the participant might have visited the doctor for an underlying medical condition. As part of the care management plant, the doctor advises the patient to quit smoking to improve his health. As a result, the participant quits smoking within six months of the research.
Maturation can also be used to explain why the participants quit smoking. Within the six months, there is a high chance that the participants came across a book without the knowledge of the researcher. This book could have made them wiser, and they decided to quit. Assuming that the participants were teenagers, they were still growing, and some factors like peer pressure and curiosity disappeared. They could have led them to quit smoking. Moreover, changes in instrumentation could have led to a flawed research conclusion. The variations between the pretest and posttest instrumentation can give varied results. This means that there is a likelihood that the smoking factors that were considered during the pretest are different from the smoking factors considered during the posttest.
TOPIC 43: THREATS TO EXTERNAL VALIDITY
- Briefly describe a hypothetical experiment that has high internal validity but low external validity.
According to Slaxk and Graugalis (2001), external validity refers to how well the study outcomes can be generalized to other settings. This is an example of a hypothetical experiment with high internal validity and low external validity. Shawn is a behavioral therapist who is assigned a task to assess the employee’s level of self-esteem. As part of the research, Shawn is interested in the role played by praise in boosting people’s self-esteem. The hypothesis for the study is; offering praise to the employees will raise their self-esteem. Therefore, complimenting an employee for doing a good job should improve the employees’ self-esteem. Shawn obtained the sample for this research from his colleagues, who are his friends, and the test was done once to all the participants. Most of the participants were white males who had attended the same college with Shawn. Also, the study participants were from privileged backgrounds.
Based on this experiment, internal validity is high because the test was administered once so that no maturation time could affect the study outcomes. Also, the test was standardized and was done on each participant using the same protocols. The research participants were unaware of the construct being tested. However, the external validity of this experiment is very low. First, the majority of the participants are white males. This means that the outcomes for this study cannot be generalized to other males from a different racial background. Also, the sample was obtained from Shawn’s friend’s circle. Thus, the sample was not representative of the general research population. Thus, selection bias leads to low external validity. All the study participants were from privileged families. This means that the study outcomes cannot be generalized to the less privileged families.
TOPIC 44: PREEXPERIMENTAL DESIGN
- Have you ever used a pre-experimental design in any of your professional or everyday activities (e.g., have you ever tried doing something in a new way and observing for its effects without a control group? Briefly describe it. Discuss whether your activity sheds light on a cause and effect relationship.
According to Campell and Stanley (2015), a pre-experimental design is an experiment that adheres to all the steps outlined for experimenting by it lacks a control group. Therefore, the single group is given the treatment, and there is no other group available for the researcher to compare the results. I have severally used the pre-experimental design in my everyday life. I remember when my younger brother was in sixth grade used to struggle with mathematics. Thus, he used to score poorly in Mathematics, and this led to a Grade Point Average (GPA). I decided to help him improve his performance by coaching him in Mathematics for at least thirty minutes after school from Monday to Friday. The main purpose of the coaching session was to evaluate whether coaching had an impact on the student’s achievement in Mathematics. I coached my brother for close to two months before seating for his final exam for that term. To my surprise, my brother had improved in Mathematics, and he managed to score 67 points as compared to the previous score, which was 44 points.
This activity did shed light on the cause and effect relationship. For this experiment, the relationship was between the after school coaching sessions and the student’s Mathematics performance. From the results, there exists a positive relationship between afterschool coaching and Mathematics performance. Thus, after school coaching improved the students’ Mathematics performance. Despite the positive results, it was difficult to tell whether the student could still improve his Mathematics performance without the thirty minutes coaching session.
Topic 45: QUASI EXPERIMENTAL DESIGNS
- Suppose a researcher observed two classes for baseline data on calling-out behavior (i.e., calling out in class when it is inappropriate). In class A, the researcher observed much variation in the amount of calling out from day-to-day. In class B, there was little variation (i.e., there was about the same amount from one day to the next). In your opinion, for which class would it be desirable to have a longer baseline? Why?
Researchers use a baseline to assist in the identification of any notable changes concerning a specific activity. According to Markham and Brown (2007), the aim of a baseline study is to give information that can allow comparison against before and after the completion of a specific activity. Therefore, data is required to measure the changes that already exist and the changes that occur after the implementation of the activity. Based on my opinion, there is a need for a longer baseline for class A as compared to class B. For class A, the researcher noted that there was a continuous change in the number of calling out in class when it is inappropriate. This means that the researcher requires a longer baseline time to assist in finding out whether there a constant variation will be achieved before concluding on the results. A longer baseline will allow the researcher to account for any factors that are likely to cause the day to day changes, that is, many variations in the amount of calling out from day-to-day. However, a small variation would also require a slightly longer baseline. This would allow the researcher to ascertain whether the variation is constant or continuous.
Topic 46: CONFOUNDING IN EXPERIMENTS
- If you were a participant in an experiment, did you try to guess what the researcher’s hypothesis was? If yes, do you think that your guess (es) influenced your behavior in the experiment? Explain.
I remember an experiment that I joined when I was the second year in college. The lecturer grouped the students into two groups of equal numbers and similar characteristics. Generally, the lecturer wanted to experiment whether students who used the library resources to research assignments were better than those who used their smartphones to research for the assignments. First, I had no idea about the lecturer’s intentions. The lecturer provided the assignment question and gave each one of us one day to research the question. Several important psychology books that could help in answering the question were availed to us, and the teacher also allowed some of the students to use the smartphone. In contrast, the third group used both the textbooks and the smartphones. In total, there were three groups for the experiment.
I tried guessing what the teacher was up to, and my guess was, “the teacher wanted to observe the research skills used by the students.” Unfortunately, this was a wrong guess. Throughout the research process, I made sure that I was keen on details and even searched for the meaning of difficult terms. I wrote down my points, and I read widely. This was different compared to how I used to conduct my research previously. After the time-lapse, the lecture marked the assignment and asked students to return to their respective groups. Based on this experience, it is true that my guess influenced my behaviors and responses during the experiment. Thus, I tried to impress the lecturer based on my guess.
Reference
Campbell, D. T., & Stanley, J. C. (2015). Experimental and quasi-experimental designs for research. Ravenio Books. Retrieved from: https://www.researchgate.net/publication/334250281_Experimental_and_quasi-experimental_designs
Markham, H. L., & Brown, N. K. (2007). Baseline Survey Protocol. Retrieved from: https://www.researchgate.net/publication/228477060_Baseline_Survey_Protocol
Slack, M. K., & Draugalis Jr, J. R. (2001). Establishing the internal and external validity of experimental studies. American Journal of Health-System Pharmacy, 58(22), 2173-2181. Retrieved from: https://www.researchgate.net/publication/260267611_Establishing_the_internal_and_external_validity_of_experimental_studies