The Logic of Experimental Design

The Logic of Experimental Design

1. Jackson (2012) even-numbered Chapter Exercises (p. 244).

Question 2: What possible confounds can you identify in this study?

The confounding variable has a meaningful result on the response variable that confounds with the explanatory variable (cite).  Confounding variables as such may impact the independent and dependent variables. Does the new therapy technique for depression help to cure the condition of depression?  This was the target of the study at issue.  The loss of participants during the implementation of the study may have had a detrimental impact on the results of the study.#$^ Maturation can be a threat since the length of time and lack of activities can be viewed as confounding variables.  The duration of the study was quite lengthy.  In addition, participants could have been exposed to new activities of their interest.  This new activities could have contributed to their improvement regardless of the response to the technique.  This study does not exposed depressed and non-depressed groups.   There is a lack of control group in this study, and many variable are confounded.

Question 4: What are internal validity and external validity, and why are they so important to researchers?

Validity can be described as the degree to which a study measures what it clams and intends to measure.  There are two types of validity: internal validity and external validity.  Internal validity refers to the accuracy of a causal effect of a claim (cite).  External validity refers to the ability for the researcher to generalize his or her claim (cite).  Researchers need to bring meaning and a degree of reliability with the data gathered to interpret their claims.  Researchers need to provide data that can allow to establish reliability and validity.  Validity is important in research (cite), as it ensures that results can be used effectively.  If results of a study are not deemed to be valid, then they are meaningless since it would not measure what it is intended to measure and answer the research question.  It is important to remember that just because a study is valid in one instance does not mean that it is valid for measuring something else.

Question 6: What are the similarities and differences between within-subjects and matched-subjects designs?

There are a number of advantages and disadvantages using the two types of subject designs. In a within-subjects design all participants are exposed to all experimental conditions. Whereas in a matched subject design the participants are matched based on the measures that are related to the study.  However some difference exist between the two types of design exist. Within-subjects designs are more efficient than a matched subject design for the following reason: fewer participants needed, instruction given only one to participants, and groups of equivalent at the start of the study.

2. What is the purpose of conducting an experiment? How does an experimental design accomplish its purpose?

A researcher main purpose of conducting an experiment is to establish causal effect relationship by manipulating the variables, controlling the environment, and observing the results.  According to Trochim, & Donnelly, (2008), three necessary criteria must be present to establish cause and effects: a) temporary precedence, b) co-variation of the cause and effect, and 3) no plausible alternative explanations.  An experimental design

3. What are the advantages and disadvantages of an experimental design in a business study?

Experimental research occurs when the researcher control one set of variable and manipulate the others to investigate the outcome or result of an experiment.  Some of the advantages in experimental research is the use of a high level control over the sample data.  Since one specific variable is being tested, the researcher can arrive at clear result about the effects of the chosen variable, by tailoring the experiment with unique attribute and maintain validity. In addition, this type of research can be applied to many different types of fields such as in the pharmaceutical area, the education, and theological areas.

Although a researcher can exhort a high degree of internal validity by controlling the environment and manipulate the group behaviors, the research, the researcher may be limited to generalize the experiment outcome.  In other words, the researcher may not be able to control external validity. A researcher may be faced to limitation in conducting some experiments due to unethical issues.  For example, an experiment may be conducted to give one group a drug that may cure a disease and give to another group a placebo.  This type of treatment and group manipulation present negatives outcomes and raise unethical issues.  Researchers may also use artificial situation with data that can be manipulated to fit the researcher objective

4. What is more important in an experimental study, designing the study in order to make strong internal validity claims or strong external validity claims? Why?

A reliable study should be strong in both internal validity and external validity. Whether internal validity or external validity is more important has been has been a debated topic in the literature.  Some authors have argued that internal validity is crucial to the experiment plausibility.  Others have argued that external validity can never be achievable (cite).  Internal validity and external validity has a negative correlation.  A researcher will employ strong internal validity when the experiment’s objective will be to determine the causal effect of an independent variable over a dependent variable. A researcher will employ a strong external validity to insure that the results obtained from a small sample can be extended to the entire population. 

5. In an experiment, what is a control? What is the purpose of a control group? Of single or multiple comparison groups?

The term control variable refers to variables that are not of primary interest (i.e., neither the exposure nor the outcome of interest) and thus constitute an extraneous or third factor whose influence is to be controlled or eliminated (cite).

Control is achieved when the potential confounder cannot vary between the exposure groups, and thus the observed relationship between the exposure and outcome of interest is independent of the potential confounder.  A control in research design is the degree to which a researcher will manipulate the data to verify or validate his or her claims. A control experiment always has two groups. One group represents the unchanged group, and the other group represents the experimental group.

The subjects in both groups will have the same characteristics and will be similar in numbers, however, the subjects in the control group will not receive the experimental treatment or the variable applied to the experimental group.   For example in a pharmaceutical experiment, the control group will receive the placebo.  The researcher will choose only one independent variables to be measured, and various dependent variables that will be used to test the hypotheses or the claims.  The researcher main objective is to control all factors that may affect the result of the experiment.  In the control group, the independent variable being tested cannot influence the results, and is separated from the experiment.  The main purpose of the control group is used to rule out alternative explanations of the experimental results, so that the researcher can examine the effects the variables can produce. A multiple-comparison examine or compare more than one pair of groups at the same time.  There are many types of multiple-comparisons groups such as the Turkey’s method, the Scheffe’s method and the Bonferroni Method.  The commonly used method is the pairwise method to determine which differences exists between the two groups. Other methods used multiple groups to compare several treatments to a control group (cite).

6. What are confounds? Give an example of a design that has three confounds. Describe three ways to alter the design to address these confounds and explain the advantages and disadvantages of each.

A confound variable is a variable that the researcher failed to control or eliminate, and can pose a threat to internal validity.  An experiment may contain a confound variable when the researcher intentionally or unintentionally manipulated the independent variables and an independent variable co-varies with a variable other than the dependent variable (cite).  In correlation studies, cofounding variables is considered as a third variable problem (cite).  Common sources of confounding are history, maturation, instrumentation, and participant selection (cite).  Confounding variable could be any other variable that will have a hidden effect on the dependent variable weigh loss.  To illustrate the problem faced by confounding variables, I will use a hypothetical experiment design of a new health supplement pill that claim helping individuals to lose weight. A sample of a 100 individuals are selected to participate in this study.  The researchers will take all individuals’ weight at the beginning of the period and at the end of the observed period.  50 participants will take the weight loss pill for a period of a six-month period and the other 50 participants will not take the pill.  If after the experiment those that take the pill loss weigh and half of those in the control group also lose weight.   This study may have other variables that the researchers may have not anticipated.  One confounding variable could be that some participants change their diet.  Another confounding variable could be exsersing, and a third confounder may be participants’ health condition. 

This hypothetical experiment design presents a number of weaknesses.  The researchers did not foresee other factors that could have led to weigh loss.  Although the pill (dependent variable) in the experiment should have been the factor of comparison, the researchers should have consider the effect of compounding variables. This hypothetical design could have been altered to eliminate some biases and preset set some conditions.  I will describe three of the commonly used alternative methods that researchers may use to alter the design to address confounding variables.

Random assignment: Involves the random allocation of individuals to exposure categories. Therefore, the distribution of known and unknown confounding variables will be similar for each of the groups being compared.

Within-subject design: A within-subject design can allow the researcher to use the same subjects under investigation, and compare and analyse the effects of the pill. 

Matched-subjects design:

7. What does “cause” mean and why is it an important concept in research? How are correlation and causation related?

A researcher main objective is to find an explanation to various things.  He or she will use statistical methods such as correlation to establish if causal relationship exists between one or two variables.  A researcher primary concern is to identify if a certain variables (event, or treatment) directly caused changes in the other.  However, there is a difference between causation and correlation.  Devroop, K. (2000) argues that the misinterpretation of causation for correlation can be due to the interpretations of experimental and correctional studies and may present danger of confusing correlation with causality.  A correlation is a relationship between two or more things variables that may or not cause one variable to vary positively or negatively, or may not have no effect.  Correlation between two variables does not necessary mean that one cause the other to change.  For example, if X causes Y to happen, then X is said to be correlated to Y. if X does not cause Y to happen, then nothing happens, correlation still exist but there will be no causation.  There can be correlation between two variables (X and Y) without the existence of any causal effect.

8. You are a researcher interested in addressing the question: does smiling cause mood to rise (i.e., become more positive)? Sketch between-participants, within-participants, and matched-participants designs that address this question and discuss the advantages and disadvantages of each to yielding data that help you answer the question. Describe and discuss each design in 4-5 sentences.

The hypothesis pose in this study is does smiling cause mood change.  To conduct this study, I will attempt to evaluate and discuss the advantages and disadvantages of using between-subject, Within-subjects, and matched-subjects designs.

Between-subject design: In this type of design, participants are divided into two groups, and randomly assigned to each condition (Jackson, 2012). One group, the experimental group, will be instructed to smile throughout the process.  The other group, the control group, will not receive any information about the purpose of this study. Results will be recorded in those two instances.  This type of study is time consuming and testing effects are minimized (Jackson, 2012), since the two groups will allow for comparisons.

Get Help With Your Essay

If you need assistance with writing your essay, our professional essay writing service is here to help!
Find out more about our Essay Writing Service

Within-subjects design: In this type of design, the same subjects are used in all condition (Jackson, 2012).  This design is done in two stages: In first stage, the participants will be instructed to smile or not smile.  In the second stage the participants will be instructed not to smile to see if smiling affect the participant’s mood.  This type of design requires fewer participants and less time compared to between-subject designs.  As Jackson (2012) acknowledges that probability testing effects and demand characteristics are high that 

Matched-subjects design: This type of design, participants are matched between conditions on variable the researcher consider to be relevant to the study (Jackson, 2012). For this study, participants are matched with those who likes to be smiled at. This type of design is costly, and require more participants. The advantage of matched-participants is that it allow for a more accurate assumption between the control group and the experimental one.

References

• Creswell, J.W. (2009) Research design: Qualitative, quantitative, and mixed methods approaches. London and Thousand Oaks: Sage Publications.
• Devroop, K. (2000). Correlation versus Causation: Another Look at a Common Misinterpretation.
• Jackson, S. L. (2012). Research methods and statistics: A critical thinking approach (4th ed.). Belmont, CA: Wadsworth Cengage Learning.
• Leedy, P. D., & Ormrod, J. E. (2010). Practical research: Planning and design. Upper Saddle River, NJ: Merrill.
• Trochim, W. M. K., & Donnelly, J. P. (2008). The research methods knowledge base (3rd ed.). Mason, OH: Cengage Learning.