Table of Contents
Question 1
A true experimental design is wherein researchers exercise absolute control over extraneous or confounding variables. Moreover, in true experiments, researchers can make confident predictions that the observed effect of an independent variable on the dependent one can only be attributed to the manipulation of the former (Campbell & Stanley, 2015). It is because, in the true experiment design, a researcher either controls all the confounding factors or takes into account their effects while ascertaining if the treatment resulted in the change. As a result, the true experiment design is often perceived as the only research design that is capable of measuring the cause-and-effect relationship. Such experiments are characterized by three primary features. The first one is that a true experiment must have at one condition wherein an independent variable differentially applied to groups of participants. This refers to some type of treatment or intervention for purposes of manipulation (Campbell & Stanley, 2015). Manipulation is characterized by the researcher consciously controlling the independent variable order to observe how it affects the dependent variables. Simply stated, the researcher varies an independent variable (intervention) and makes observations on the dependent one. Huber (2011) recommends manipulating and testing only one variable. Although testing more than one variable is possible, such experiments including their statistical analysis are likely to be more difficult and cumbersome. In the aforementioned scenario, the condition being tested and applied differentially to groups of participants is having a mentor and lack thereof after being released from prison, and how this affects recidivism, that is, whether having a mentor causes an offender to resist from taking part in criminal activity. Therefore, having a mentor is the independent variable and recidivism is the dependent variable. The researcher then manipulates the independent variable by creating two conditions – having a mentor and not having one, after which the effect of these two conditions based on the dependent variables are observed on participants.
The second important feature of a true experiment is that the participants in the research study are randomly assigned to the groups (treatment or control), which means that the only differences between the groups can only be because of a chance (Mitchell & Jolley, 2012). Random assignment plays a crucial role in controlling for confounding variables, which makes it possible to make direct inferences on the cause and effect relationship. This is because randomization ensures that the characteristics of the participants in the treatment and control groups are similar; therefore, the effects of the confounding variables are eliminated through the dispersion of the participants’ characteristics variability equally between the groups. Random assignment means that all participants have the same probability of being assigned to a particular group (Huber, 2011). Randomization also helps in the elimination of systemic bias. It also helps in minimizing the threat of internal validity. With respect to the scenario above, participants, which in case comprises of just released inmates, can be randomly assigned to the treatment group (having a mentor) and the control group (not having a mentor). There are various randomization techniques that the researcher can employ such as tossing a coin, drawing slots, using a random table, or computer software for random sequencing.
The third characteristics of a true experiment lies differences measured at the end of the experiment (after the research participants have differentially experienced the independent variable) are compared with the differences that are likely to have occurred by chance (differences of a random magnitude) (Campbell & Stanley, 2015). It means that a true experiment must have a viable control group that does not receive any form of treatment. The presence of a control group helps in ascertaining whether the treatment had an impact. If change is documented in the treatment group after and intervention while the control group did not change or showed less improvement, the researcher can make an inference that the treatment resulted in the improvement (Huber, 2011). However, it is imperative to note that a no-treatment control group is not a requisite of the true experiment; instead, at least two conditions or groups are needed for the true experiment. Depending on the research question, comparison groups may be employed wherein two different treatments are used. In most studies, the control group often either receives the standard treatment or no treatment at all, while the other group gets some alternative treatment (Huber, 2011). With respect to the scenario above, the treatment group will comprise of released inmates having a mentor. Their recidivism rates will be compared with released inmates who do not have a mentor, who will comprise the control group. Because of these three essential characteristics, the true experiment design can help in determining causal relationships. Since a researcher determines and exercises control over the experimental conditions, the findings reported can be attributed to manipulated and known variables rather than uncontrolled and unknown factors.
In answering the research question, the true experiment design that will be used will take the form of post-test only design. This design comprises of two randomly assigned groups – the treatment and the control one. However, neither of these groups is pretested before the intervention has been implemented. Moreover, the intervention (having a mentor) will only be implemented in the treatment group but the post-test will be carried out for both the groups. In the scenario, it is impossible for conduct a pretest on the subjects given the nature of the research.
Question 2
Safeguarding the welfare and the rights of people who volunteer to take part in a research study is a crucial requirement of ethical research. In the scenario at hand, a number of unique ethical standards must be considered when designing the study. The first ethical dilemma unique to the scenario relates to the right to service. Researchers are constantly dealing with the ethical issue of an individual’s right to service. In this respect, an ethical question arises when a desirable treatment offered to the experimental group is not provided to the treatment group. According to Huber (2011), good research practice demands the utilization of a no-treatment control group. However, when the program or treatment is likely to have beneficial impacts, individuals in the no-treatment control groups are likely to feel that they have been denied their rights to access beneficial services. In such a case, an ethical standard is that, when the preliminary findings of the study indicate beneficial outcomes in the treatment, the experiment can be interrupted and the program administered to all participants (Huber, 2011).
This ethical dilemma can also be solved through the random assignment of participants. Campbell & Stanley (2015) posits that an ethical question arises when testing a beneficial policy and assigning the treatment randomly to some individuals and not to others. Nevertheless, there is a general agreement that the random participants’ assignment is an ethical procedure that can be used when deciding how a likely beneficial or harmful treatment is allocated among the participants (Huber, 2011). Another ethical standard that safeguards the rights of participants and is unique in the underlying scenario is no harm to the subjects.
Criminal justice research is capable of resulting in psychological and physical harm, including embarrassment (Huber, 2011). Harm to third parties, researchers and subjects is a possibility in the criminal justice field. For instance, mentors are put in harm’s way when engaging with released inmates. In addition, gathering information from people with prior criminal records presents a likelihood of violence against the researcher and his or her third parties. Psychological harm is also highly probable when participants offer information relating to their criminal experiences (Campbell & Stanley, 2015). Within such context, it is an ethical standard that likely harm may be justified in cases whereby the potential benefits outweigh the likely harm. Hence, a researcher must determine whether the likely benefits offset the potential harms associated with the research.
Another ethical standard pertinent to the scenario is voluntary participation. Criminal justice research is usually intrusive, which is the case of this experimental design in the scenario involving mentoring, which in turn will require participants to share personal information regarding themselves (Huber, 2011). In this respect, participation in the research should be on a voluntary basis. It is important for participants to give their informed consent prior to taking part in the research. Appropriate disclosures are needed to make sure that subjects have a comprehensive understanding of the research. Such disclosures should be done throughout the entire research process.
Question 3
The main role of the IRB is to safeguard the welfare and rights of human participants in the research. The IRB also exists to protect the interests of the participants while at the same time ensuring that the conducted research is ethical. The core ethical principles advocated by the IRB include respect towards persons, beneficence, and justice (Campbell & Stanley, 2015). The respect towards persons requires seeking informed consent from participants. Beneficence assumes minimized risks to participants while justice is concerned with the equitable distribution of the research benefits as well as participants’ selection (Mitchell & Jolley, 2012). The IRB also spells out the researchers’ responsibilities, including making sure that participants are eligible, proper obtaining of the informed consent, and that the research is valid and properly designed.
Another design that can be used to measure the effects of having a mentor on recidivism entails using secondary data design. This will involve tracing records of inmates in prison databases, followed by determining whether or not they were assigned mentors after release. The study will then compare the recidivism rates of inmates who were assigned mentors and those who did not have them.