Research Methods in Psychology

Experimental Research Methods

Experimental research is basically any study that is steered with a scientific methodology in mind, whereas multiple variables remain constant while the others are measured (Gravetter & Forzano, 2015).  Experimental research gathers a lot of data which helps the decision making process and is the main institution of quantitative research methods. An example of this research method would be conducting a lab test. Any study is considered experimental research as long as it is conducted under scientifically satisfactory conditions. There should always be a cause and effect relationship between the two variables, but sometimes a conclusion cannot always be established.  Experimental research is essentially directed by time and behavior between the variables that are being observed. There are three types of experimental research designs which are based on classification of the variables according to several conditions:  

• Pre-experimental (simplest form)
  • A group(s) retained under observation after all aspects are measured for their cause and effect relationship. This is conducted to recognize whether or not there is a need for further investigation.
  • Three types of pre-experimental methods are:
    • One-shot Case Study: a group studied at a one point in time in which change is assumed.
    • One-group Pretest-posttest: A case that is perceived at two points in time; before and after treatment.
    • Static-group Contrast: A group that had treatment which is compared to a group that hasn’t.

• True experimental (most precise form)
  • Relies on statistical examinations to prove (or not) a hypothesis. Establishes a cause-effect relationship within the group(s). These factors have to be satisfied in order for it to be a true experiment:
    • Control and Experimental Group(s)
      • Research rules do not apply the controlled group as they do to the experimental group
    • Variable (to be manipulated)
    • Arbitrary distribution

• Quasi-experimental (Quasi is defined as resemblance)
  • Similar to experimental research but is distinctive because the group is not random. The manipulation of the independent variable transpires before the dependent variable is even calculated. Random assignments are not required while using this method.

There are many advantages of performing a True Experiment (Brown, 2010):

• Instrumental way of drawing conclusions
• A basic and well-organized type of research which can be realistically used for a variation of disciplines
• Results can be tested and proven
• Results are easily attained due to the controlled environment
• Various experiments can be tailored while preserving the validity of the actual designs

While reading the advantages and benefits of a True Experiment; there are disadvantages and limitations to this method as well (Brown, 2010):

• Can create synthetic situations that do not reflect real-life situations
• Human error can cause validity concerns
• No real control over all of the extraneous variables.
• Health, behaviors, mood, and life experiences may influence test subjects reactions
• It’s the researcher’s duty to adhere to ethical standards for case validity.
• Experimental research can determine the causation, but it usually cannot state “why” the result occurred.
• Occasionally impossible or impractical since the research is essentially effective when carried out in normal or natural environmental settings

Quasi-Experimental

Quasi-experiments are possibly conducted in most field settings where random assignments are almost impossible or it is extremely difficult to maintain. They are frequently used to assess the efficacy of a treatment (i.e. psychotherapy or intervention). The three most common types of quasi-experiments are nonequivalent, pretest-posttest, and interrupted time series group designs.

Nonequivalent Group(s)

This between-subjects’ design is when participants of a group have not been arbitrarily allocated to certain conditions (Campbell & Stanley, 2015). I am currently researching “What is the correlation between the environmental factors of human development and substance abuse in adulthood” and I am in the process of evaluating a new treatment technique of psychotherapy. One way of understanding the nonequivalent design would be to conduct a study with a group of college students in their senior year and a control group comprising of another group of college students in their senior year. From my readings, I would gather that this would classify as a nonequivalent groups design based on the fact that I have not assigned the students to certain treatments arbitrarily; which ultimately means there could be significant differences between the two groups of college seniors. For example, the parents of the higher achieving or more driven students might have recommended they take Mrs. Davidson’s Psychology course. Or the Dean could have assigned the academically troubled seniors to Mr. Jackson’s Psychology course instead, because he has a Ph.D. in Psychology where as Mrs. Davidson does not. The professors’ styles, class environments, and course requirements are very different and could ultimately cause different levels of education and accomplishment or even motivation amongst the students.  The study could result in a difference in the two psychotherapy techniques and  knowledge of treatment plans could have been caused by the difference in their teaching methods; or it can be caused by any other variable as well.

Pretest-Posttest

The dependent variable is measured in one instance before treatment is executed and again after implementation is known to be the pretest-posttest design method (Campbell & Stanley, 2015). In this instance I will do the same research study as listed previously, “…What is the correlation between the environmental factors of human development and substance abuse in adulthood…” and use pretest-posttest in this manner. Assuming I am the researcher who is concerned with the effectiveness of a Students Against Drug Abuse (SADA) education program with the senior college students’ behaviors toward illegal drug usage. In the first week I would measure the behavior of the students that live on campus. The following week is when the implementation of the SADA program would occur and a week after implementation; I would measure the students’ behaviors again. The pretest-posttest strategy is based on each participant/individual being tested initially under the controlled condition and then again under the treatment condition.

Interrupted Time Series

The interrupted time-series (ITS) design is a modification of the pretest-posttest design. A time series is fundamentally a set of measurements that are taken at different intervals over a particular period of time (Bernard & Bernard, 2012). Interrupted time series analysis is debatably a paramount method for dealing with interventions when randomization is not possible or if there is no clinical trial data. This method can be beneficial in providing answers to questions about population level interventions as well as their effects; but on the other hand the design’s implementation can be particularly challenging.

Simply put, an ITS is molded using a regression model, such as logistic or linear, that only includes three time based variables, whose regression measurements are estimated before pre-intervention occurs (Gravetter & Forzano, 2015). The pre-intervention gradient quantifies the movement for the result before the intervention. The level change is then described as an estimation of the change in that level which is credited to the authentic intervention, between the time points and directly before and immediately afterwards as well to account for the pre-intervention movement. The main assumption, with no external systematic factors affecting the trend, is that current pre-intervention trend would ultimately remain unchanged leading into the post-intervention phase.

Advantages and benefits of having a Quasi Experimental research (Brown, 2010):

• Easily feasible since it does not have logistical and time constraints
• Increases peripheral validity since the variables aren’t random with the ability to create reasonable & logical control groups
• Due to the natural research environment the test subjects reactions are genuine
• Statistical analysis reinforces research results
• Randomization and pre-screening is not required which reduces time and valuable resources
• Variations of the research can be tailored while preserving validity

Disadvantages and limitations of conducting a Quasi Experimental research study (Brown, 2010):

• Result limitations due to having non-equivalent testing groups
• Lack of randomization can lead to insignificant statistical analysis
• Having a less controlled environments & variables doesn’t take into account influences and other pre-existing factors
• Human error plays a significant role in the research’s validity
• Ethical standards must be upheld for research’s soundness and validity

 References

• Bernard, H. R., & Bernard, H. R. (2012). Social research methods: Qualitative and quantitative approaches. Sage.
• Brown, L. (2010). Quasi-experimental research. Doing Early Childhood Research: International perspectives on theory and practice, 345.
• Campbell, D. T., & Stanley, J. C. (2015). Experimental and quasi-experimental designs for research. Ravenio Books.
• Gravetter, F. J., & Forzano, L. B. (2015). Research methods for the behavioral sciences (5th ed.).