Kris Parker 12:08pmApr 8 at 12:08pmManage Discussion Entry Good afternoon Class, “Quasi-experimental studies evaluate the association between an intervention and an outcome using experiments in which...

Kris Parker

Good afternoon Class,

“Quasi-experimental studies evaluate the association between an intervention and an outcome using experiments in which the intervention is not randomly assigned (Schweizer& Milstone, 2016). Quasi-experimental studies are commonly used to observe rapid responses to outbreaks or against other groups that require promptness in non-randomized interventions. Quasi-experimental studies have three major types categorized as interrupted time series designs, designs with specific control groups, and designs without specific control groups. This experiment is somewhat flexible and has dimensions in which data could not be misinterpreted.

In the study of the this weeks readings I chose, EL training is questioned to be able to innate higher response to promotions and talent growth within the enterprise. There were employees that had the choice to volunteer for the study and others that did not choose to participate. The volunteer staff would be the true experimental group and the ones that did not volunteer are the controlled group.

In another study that would best be suited by using Quasi-experimental studies, is studying the two fields of healthcare of epidemiology and antimicrobial stewardship that apply interventions at a unit level. Due to outbreaks or the safety of problems of patients would require prompt non-randomized interventions. The Quasi-experiment would examine the association between intervention and an outcome that would intel intervention that is not randomly assigned. Data would be reported in different measures completed over a length of time. The potential treats to validity could be selection bias and as well as maturation bias where the participants may change naturally over time due to the data being collected over time. This would affect the overall results due to them meeting the guidelines at first but then changing during the study. With the choice of having a small intervention over a large, I could choose the small intervention to get a true experimental design limiting the possibility of boredom and participants dropping out. If with the downsizing, this study is best suited for Quasi-experimental study.

Cozby, P. & Bates, S. (2017). Methods in behavioral research (13th ed). New York, NY: McGraw-Hill..

Schweizer, M. L., Braun, B. I., & Milstone, A. M. (2016). Research Methods in Healthcare Epidemiology and Antimicrobial Stewardship-Quasi-Experimental Designs.Infection control and hospital epidemiology,37(10), 1135–1140. https://doi.org/10.1017/ice.2016.117

• 
  

  
  Daniel Hoogestraat
  

  YesterdayApr 7 at 7:05pmManage Discussion Entry
  

  
  Quasi-Experimental Designs
  

  
  

  Conducting research for an academic study is the primary way scholars support and persuade their premises and audiences. However, there are different approaches as to how scientists conduct research. For instance, scientists typically use two designs to conduct experiments: the posttest-only design and a pretest-posttest design. However, some experiments require a quasi-experimental design if the researchers find that they do not have complete control over the experiment's variables (Cozby & Bates, 2017). Quasi-experimental designed experiments also differ from traditionally designed experiments as they cannot employ randomization, nor are they able to reference a control group (Cozby & Bates, 2017). Through the examples provided below, I will demonstrate my understanding by identifying the design's major threats to validity and how a quasi-experimental design can be necessary over an experimental design.

  
  

  
  One-Group Pretest-Posttest Design
  

  
  

  The one-group pretest-posttest design assesses the participants' results before and after the experimenter's manipulation occurs (Cozby & Bates, 2017). Specifically, the researcher will first measure a participant's baselines of specific traits like behavior, emotions, cognition, or whatever the researcher is interested in testing. Then, the researcher will implement something like a treatment, therapy, or regimen and test the subjects again to evaluate the treatment's effectiveness (Cozby & Bates, 2017).

  
  

  
  One-Group Pretest-Posttest Design: Example
  

  
  

  An example of a one-group pretest-posttest design would be testing the cessation of smoking cigarettes. Initially, the researcher would collect a random population of smokers and evaluate their smoking habits (pretest baseline) (Cozby & Bates, 2017, p. 169). After each participant's baseline is established, the researcher would then manipulate or employ the group's smoking regimen. After the manipulation, the dependent variables (smokers) will be analyzed (i.e., how many cigarettes were smoked after the employed regimen) (Cozby & Bates, 2017).

  
  

  
  Potential Threats to Validity
  

  
  

  However, according to our course materials, this quasi-experiment has serious deficiencies. For example, the experiment does not consider several factors. The research does not incorporate experimental "history, maturation, testing, instrument decay, and regression towards the mean" (Cozby & Bates, 2017, p. 231).

  
  
  
  
  1. 
     History– For example, a participant could have died or experienced a confounding event between the pre and posttests. This would make the finding of the participant's smoking habits to be invalid (Cozby & Bates, 2017).
  
  
  2. 
     Maturation– For example, say the experiment's duration between pretest and posttest is five years. During this time, the smokers could have grown wiser or more concerned with their health and quit smoking. Therefore, the researcher's regimen would not be considered, and the experiment would be invalid (Cozby & Bates, 2017).
  
  
  3. 
     Testing– The pretest test itself could be enough for smokers to recognize their unhealthy habits. Therefore, merely taking the pretest could muddy posttest results. Thus, the therapy employed would not be gauged accurately in terms of baseline measurements (Cozby & Bates, 2017).
  
  
  4. 
     Instrument Decay– There is always the possibility of the participants becoming fatigued or bored of the experiment. For example, over time, the participants may not be as motivated to record their smoking habits, thus making the experiment inaccurate (Cozby & Bates, 2017).
  
  
  
  

  
  True Experimental Design
  

  
  

  In order to make this experiment a true experimental design, the researcher would collect a random population of smokers, evaluate their smoking habits (pretest baseline) and divide the smokers into two "equivalent" groups (Cozby & Bates, 2017, p. 169). After each participant's baseline is established, the researcher would then manipulate or employ a smoking regimen on Group X and do nothing to Group Y, the control group. After the manipulation, the dependent variables (smokers) will be analyzed.

  
  

  
  Concluding Remarks
  

  
  

  Assessing whether to employ a pretest-only, pretest-posttest, or a quasi-experimental design is the starting point for most experiments. Implementing either design will help a researcher become more persuasive in supporting claims because it renders data more objectively and scientifically. In short, selecting, observing, and measuring data (pre and post) to determine variances, percentages, and behaviors between groups validate research because it can be easily replicated or generalized for future applications.

  
  

  
  References
  

  
  

  Cozby, P. & Bates, S. (2017).Methods in behavioral research (13th ed). New York, NY: McGraw-Hill.

  
  ReplyReply to Comment
  
  
  


•