Why Experiments are done? Explain.
Experiments are done for investigating some scientific issues. While conducting an experiment, the scientists usually forms a hypothesis and through the experiment, this hypothesis is being analyzed, whether it is correct or not. In some other processes, scientist figures out some questions regarding the research topic and based on this the research questions the scientists attempt to design the research in such a way that at the end of the experiment, the scientist can achieve the answers of the research questions (Creswell, 2012). Usually, scientists perform an experiment more than one time for ensuring the accuracy of the data.
The experimental design is an outline, strategy or plan for conducting the research by finding the answers of each research questions. The scientists choose the appropriate research design for the experiment based on the nature of the experiment. In this context, the researcher would analyze the baseline measures based on which the experimental design would be selected. The experimental design helps to provide a guideline to the scientist for carrying out the experiment in a correct flow and thereby ensuring the achievement of expected research outcomes (Jackson, 2015).
The advantages of implementing an experimental design include:
- Provide an insight of the entire research process. In addition, it helps to figure out the right direction of conducting the research in a systematic measure
- In experimental design, the scientist can have control over the variables
- The appropriate research design helps to determine the best population for the experiment
- It gives more transferability in comparison to anecdotal research
- It helps to give the research a definite shape
- Research design helps to avoid bias and reduce experimental errors
The disadvantages of implementing an experimental design include:
- As the research design is implemented by humans, the human errors can happen which might affect the success of the experimental design. Thus, selecting an appropriate experimental design is very important for research success
- There might be personal bias, as it is entirely dependent upon the scientist conducting the research
- The results achieved in the experiments might not be applied on different situations and it might be difficult to be replaced by other
- The results could be artificial as the participants can be influenced by the experimental environment which could enhance the rate of experimental errors and thereby enhancing the chance of getting error prone results from the experiment, that is why the experiments are done more than one time (Trochim, Donnelly & Arora, 2015).
The right balance of two validity claims is important for a successful research study. If a study has high internal validity when all the confounding variables are controlled, thereby ensuring that only independent variable would affect the dependent variable. On the other hand, external validity is high when the real life situation is considered. However, a study cannot have both external and internal validity high, thus it is better to keep both aspects in mid way, having moderate internal and external validity claims (Zikmund et al., 2012).Usually, experiments are conducted by implementing some changes to the experiment or test group of sample. However, to analyze the results of the experiments, the results needed to be compared with a group of sample upon which the experiment has not been done. This group is known as control of the experiment. The purpose of the control group is to compare the changes in the test group after implementation of the experiment upon the test group. The test groups should be compared with the control group for analyzing the changes happened to the test population as well as to analyze it whether the results are satisfactory to the expected results (Ramsey & Schafer, 2012).
Single comparison or control group is required for small population and when only single intervention is implemented on the test group; then only a single control group is required. In contrast, for large population, multiple control groups may be required for enhancing the accuracy of the experiments. Multiple control groups can arise for several reasons, sometimes, treatments were not implemented and in other cases, the treatment were provided but declined. For different group of samples, multiple control groups are required for comparing the test samples before and after implementation of interventions (Andersen et al., 2012).
Confounds are the extraneous variables in the statistical model which correlates with dependent as well as independent variables. These are the variables which are uncontrollable by the scientists conducting the research. It is also known as the third variable which can affect the relation between dependent and independent variables adversely, thereby affecting the accuracy of the results.
For example, a test was done with 200 people including 100 men and 100 women, the results found that lack of exercise led to the weight gain of participants (Pickard, 2012). However, there were other confounding variables which could not be controlled. Here, one confounding factor is age of the participants which can contribute in weight gain. Another confounding variable was participant’s diet, participants were not monitored for their common diet before subjecting in the experiment. The third confounding variable in this experiment was gender, sometimes, women are more likely to gain weight after certain period of life than men, it was also uncontrollable. These variables could affect the relationship within lack of exercise and participant’s weight gain.
One way to reduce the effect of confounding variables is introducing control variables for confounding variables, such as; the inclusion criteria can include participants of only 30-35 years old which can control gender confounding.
Another strategy is, keeping the participants in same diet for 1 month before conducting the experiment. It can control the effect of diet upon the dependent and independent variables.
Another strategy for reducing the effect of confounding variables is counterbalancing while having paired condition. In this context, half of the participants would be measured under one condition and others would be measured under other condition. After certain period the conditions would be interchanged for two groups. It would help to analyze the effect of those conditions upon the groups through a comparison (Lichtman, 2012).
Cause is somewhat which has a specific result in effect. Causation is an event which makes changes over the previous situation upon which the event was happened and it will certainly have an effect. However, it is somewhat different than the event of correlation in research variables. According to the principle of causality, it is a universal concept, in the world, every phenomenon having some specific consequences must have risen as an effect of some cause. It is an important concept in research because it always has an effect which is important in research (Neuman & Robson, 2012). The researcher would attempt to conduct a research based on certain issues. These issues must be raised from some certain causes which have some specific effect. Thus, cause is an important concept in research.
While conducting a research, wide range of confusion arises within causation and correlation. However, these two phenomenons are different. According to Goertz and Mahoney (2012), the correlation is the first step for establishing a cause. When two or more events are happening at the similar time which are associated with each other, but might not have a cause-effect relationship, this is known as correlation. On the other hand, causation must have a cause-effect relationship within two variables. In both correlation and causation, there are relationship establishment within two variables, but in correlation, there must not be a cause-effect relationship unlike causation.
The research question is, ‘does smiling cause mood to rise?’ In this condition, the between-participants design would include the division of participants in two groups. These two groups include one test group and one control group. For the test group, the smiling intervention would be applied; in contrast, the control group would not receive smiling intervention. After certain time, the results would be compared within two groups to understand the effect of intervention upon the participants. Advantage of this process is that two different groups are being compared with two different conditions, thus the event is compared twice, enhancing the accuracy of result. In contrast, fluctuation of conditions can give biased result (Easterby-Smith, Thorpe & Jackson, 2012).
The within-participants design would include only one group of participant. In this design, the participants would be subjected to two different conditions. In one condition, the participants would be subjected to smiling intervention and after a period of interval, they would be treated as control group, no intervention would be implemented, the results would be compared at the end of the study. There is no chance of condition fluctuation which is an advantage, but the comparison is done only once, thus there is a chance of experimental error (Allwood, 2012).
The matched pair design is a randomized block design with two treatment condition. However, here is only one treatment condition for the current research question. Thus, this design would not be applicable here.
Allwood, C. M. (2012). The distinction between qualitative and quantitative research methods is problematic. Quality & Quantity, 46(5), 1417-1429.
Andersen, P. K., Borgan, O., Gill, R. D., & Keiding, N. (2012). Statistical models based on counting processes. Springer Science & Business Media.
Creswell, J. W. (2012). Qualitative inquiry and research design: Choosing among five approaches. Sage.
Easterby-Smith, M., Thorpe, R., & Jackson, P. R. (2012). Management research. Sage.
Goertz, G., & Mahoney, J. (2012). A tale of two cultures: Qualitative and quantitative research in the social sciences. Princeton University Press.
Jackson, S. (2015). Research methods and statistics: A critical thinking approach. Cengage Learning.
Lichtman, M. (2012). Qualitative Research in Education: A User's Guide: A User's Guide. Sage.
Neuman, W. L., & Robson, K. (2012). Basics of social research: Qualitative and quantitative approaches.
Pickard, A. (2012). Research methods in information. Facet publishing.
Ramsey, F., & Schafer, D. (2012). The statistical sleuth: a course in methods of data analysis. Cengage Learning.
Trochim, W., Donnelly, J., & Arora, K. (2015). Research methods: The essential knowledge base. Nelson Education.
Zikmund, W., Babin, B., Carr, J., & Griffin, M. (2012). Business research methods. Cengage Learning.