Quasi Experiment

By Author: Henry Ford
Total Articles: 189
Comment this article

A quasi experiment is a research design which has some but not all the characteristics displayed in a true experiment. In true experiments, the subjects are assigned random treatment conditions and the only difference seen in the groups is due to chance. The most frequently missing element in quasi experiments is that of random assignment of research subjects to control and the experimental conditions. Examples of research design based on quasi experiment are the natural experiments where in this case nature assigns two possible outcomes to the subject and trend analysis. A quasi experiment is almost identical to a true experiment. Some of the common features of quasi experiments are: matching is used rather than randomizing units; time series analysis is applied and unit being analyzed is not people but something else (Angrist, 2003). Kridel (2010) asserts that a fundamental problem in undertaking quasi experiments research is based on the fact that one does not make use of randomization. Randomization is helpful in selection or assignments of samples so as to create comparisons which are useful in inferring that treatment ...
... caused changes have taken place. Instead, non-equivalent groups that are used for comparison in quasi experiments are likely to differ more often in different ways apart from the specific treatment whose effect is being examined.
Herman & Gibbons (1997) have shown the frequently used quasi experiment designs and a brief mention of how they can be challenging if any of them is used for my dissertation:
i. Non equivalent group, posttest only: here, an outcome measure is administered to two groups or to a treatment or program group and a comparison of the same. The major challenge faced is that the two groups might not be necessarily the same before the instructions are given and can therefore interfere with the progress made.
ii. Nonequivalent group, pretest- posttests: it eliminates major limitation of non equivalent group posttest only design. When starting the study, the researcher is supposed to empirically assess the differences of the two groups under study. If the researcher finds out that the group performance is better than the other they can rule out the initial differences and normal development as explanations for the differences. Some of the challenges can develop from the subjects of study in the two groups when they are exposed to treatment conditions, or when the two groups differ on preset measure. If the differences are experienced at the onset of study, differences occurring at the test score conclusion are difficult to interpret.
iii. Time series design: in this case several assessments are obtained from the groups under treatment as well as from the control group. This will take place before and after treatment application. Observations made before or after can be useful in providing information about the items of study. Measurements before and after program are likely to provide a more reliable picture achievement but the time series is useful in providing sensitive information on trends and performance. Challenges faced with time series include, differences experienced because of preset measures and hence difficulties in interpretation of the conclusions made.
iv. Interrupted time series with switching replications: in this case there is refinement of two groups with each serving as either the treatment or the comparison group but done on alternating basis, with the use of multiple replication of treatment removal. Problem is that it requires higher level of control by the researcher over the subjects though it is strong in removal of threats to validity. Another problem is that it is not useful in studies where there has gradual treatment intervention.
v. Regression discontinuity design: the hypothesis is that if there is a chance of treatment, then the regression line’s slope which relates scores before and after treatment will be the same. Even with this there will be a discontinuity in the magnitude on dependent variable after the treatment which is deemed to occur immediately. A challenge occurs when there arises a need for treatment effect which will take form of a teeper regression slope but does not discontinue at point of treatment. This kind of treatment effect is difficult to differentiate from the simple curvilinear relationship.
Qualitative approaches to dissertation
Qualitative dissertation refers to any kind of research which produces findings not arrived at by use of means of statistical approach or other means of quantification (Cobin & Strauss, 1990). It mainly focuses on the qualitative aspects of a research. Unlike quantitative researchers who are interested in determining, predicting and generalizing of findings qualitative researchers mainly focuses on illuminating, understanding and extrapolating to situation that are similar. Qualitative analysis often results in different kind of knowledge than what the quantitative research does. Eisner (1991), says that all knowledge even that obtained from the quantitative research should be referenced in qualities, and that these qualities can be used to show our understanding of our world.
The following are qualitative aspects that are not conducive for a dissertation due to their length of study:
Participant observation- This qualitative approach finds common use among researchers although it is very demanding. This is because it demands the researcher to become part of the context or culture under study. To achieve integration in a study topic context, a lot of time is required as well as the learning of new techniques and languages present in the item of study. Trochim (2006) asserts that the complexity in this method is compounded by the fact that there are no formal tools for collecting data nor is there any official protocol for collection and analysis of results. The ambiguity of structure in this approach leads to generation of a wide range of responses which may take a lot of time to analyze and correlate. Considering that the researcher needs to become part of the context under study, Trochim (2006) argues that this approach requires prolonged period of time, in the range of months to years for its completion.
Unstructured interviewing- This approach is preferable when general information is desired from a particular sample of study. Unlike in structured interviewing, the researcher relies on his personal wit and memory to memorize events and facts since no formal tools are prepared for this approach. Its challenge lies in sustaining consistent objectivity and avoiding ambiguity that may rise as a result of lack of formal research tools. Its timing challenge lies in the unstructured sample that a researcher may need to evaluate. The method also raises the likelihood of too diverse responses considering that respondents are not formally guided in supplying relevant information. As such, a substantial amount of time is needed to collect meaningful data for use in a normal scholarly dissertation. The method can however be appropriate for social research, aimed at collecting vital data for use by organizations and governments.
Mixed method design
A mixed method of research design is one which includes qualitative and quantitative research data, methods and techniques. All these research characters are mixed in one case study. In this research method, there is use of mixed data (texts and numbers) and also application of other or additional means (text and statistics analysis). Mixed research method makes use of inductive and deductive scientific method and has multiple data collecting forms and produces pragmatic and eclectic reports. The reason one can use mixed research design is to take advantage of strengths due to every type of information collection and in order to minimize weak points of each of the aspects. This helps in increasing the accuracy and validity of the gathered information. The qualitative and quantitative aspects of a research are taken into account (Hunt, 2007).
Benefits associated with use of mixed method research design include:
i. Validity of results obtained: this is strengthened by use of more than one method to conduct the study. This approach commonly referred to as triangulation is more often said to be the advantage of using mixed method approach (Frechtling & Sharp, 1993).
ii. Combining the different methods pays off when it comes to instrumentation of all data collection methods and in increasing evaluator’s understanding of findings. One can start with the qualitative aspects of the research like a focus on the group to be discussed which helps in alerting the evaluator the issues to be explored in a survey of the program participants, then followed by quantitative aspects and then qualitative one (Miles & Huberman, 1994).
iii. Use of mixed method can help the evaluator if need be to expand or modify design used for evaluation or the data collection methods. The action can occur when the use of mixed method uncovers discrepancies or inconsistency which will necessitate the evaluator re examine evaluation framework or the procedures applied in data analysis (Frechtling & Sharp, 1993).
iv. Mixed method provides chances of initiation. Here new research questions or challenges results are obtained through the use of one method. It provides new in sights on how the program has been perceived and also valued across the different sites (Green et al., 1989).
v. Complementarily: this helps in clarification and illustration from one method with the help of another method (Green et al., 1989).
vi. Accuracy: mixed method design helps in checking of accuracy as focus is made on a single process using the different approaches. It is done by combining the two positive aspects (the quantitative aspects in conjunction with qualitative ones) and summarizing them so as to produce data that is accurate in a given research.
Challenges when writing a dissertation
According to Kuther (2010), a dissertation is an extended piece of writing usually dividend into chapters. It is a required part of doctoral study that is undertaken after a student has completed his or her course and has passed a comprehensive examination. It is regarded as the last hurdle to be undertaken before one is said to have completed a doctorate or a PhD degree. One of the challenges of writing a dissertation is deciding on topic of study. It should bring a new and creative concept to the respective field of study so as to demonstrate the expertise obtained by a student. In pure sciences and social sciences, a dissertation usually requires that the student conduct an empirical research. Before deciding on the relevant topic of study can be hectic. It should be a topic that interests me most and yet relevant to my area of study.

The author is associated with

The author is senior writer at https://www.allacademicanswers.com