Research Methodology and Statistics Report (Assessment)

Various research approaches

Correlation study

This is the study that uses statistical measures to identify the relationship existing between variables. One of the advantages is that it provides room for simple observation of variables in their original set-up. It also protects the results hence no room for manipulation. Its disadvantage is that it does not have the ability to relate variables (McMillan and Schumacher, 1993).

Gay, Mills, Geoffrey and Airasian (2006) define a correlation study as “a quantitative method of research in which you measure two variables for each individual. Its collected data determines whether, and to what degree, a relationship exists between two or more quantifiable variables” (p. 191).

“Correlation study shows the relationship between two variables but do not give inner details and description on the mechanism of the relationship” (Gay et al., 2006, p. 11). In this study researchers are capable of making predictions on things having familiar correlation unlike in casual comparative study (LaFountain and Bartos, 2002, p.32).

Disadvantage is that it is not easy to remember correlational study since cause is not measured. Researchers are not able to “establish the cause of phenomena despite the existing relationship” (LaFountain and Bartos, 2002, p. 45).

Casual-comparative study

This is the method that reveals casual relationship existing between variables. It relates the subject in question to already available data provided by the management. The availability of management in this design helps in providing sufficient information for the study. However, in some instances in-depth study on independent variables is required so as to develop necessary test differences between groups.

Example is a research study done on the causes of death amongst the newborn babies, the researcher selects the record on actual cases of babies who have died within the first month of life and then record the ‘controls’ who are the babies who survives their first month of life.

The researcher then interviews the nurses to compare the history of these two groups in order to determine the prevailing risk factors that might have caused the deaths as opposed to survival (Holland et al., 1985).

Advantage is that researchers are “able to make predictions” on things that their correlations are known unlike in casual comparative study (LaFountain and Bartos, 2002, p.34).

The variables used in this study cannot be manipulated, hence making this method desirable. This is quite different from other study methods like correlation. The study tries to “highlight the cause of the effect” making implementation easier (Resenthal and Rosnow, 1999, p. 190).

Disadvantage of Correlation is that is difficult to remember since causes are not measured. Researchers are not able to “establish causes of events despite the relationship” (LaFountain and Bartos, 2002, p.34). Also compared to other study types like quasi experimental study, casual-comparative does not provide “an actual or accurate data to the researcher” (Resenthal and Rosnow, 1999, p.222).

Quasi-experimental study

This study examines the results through comparison of subjects that receive program activities and the results of such similar group that do not receive program activities. The results before and after group’s participation are compared. The advantage of this type of experiment is that it has the ability to reveal causes and effects. Its disadvantage is that it cannot establish relationship between the results (Campbell and Stanley, 1963).

Quasi experiment can be very valuable in providing important information such as; detailed information about the population under study, information that identifies the expected changes and results, detailed data on the level of change that occurs over a period of time, it also provides information on the changing outcomes and those that do not change (Bogdan and Biklen, 2007).

Experimental study

This is the method where the variables defining one or more phenomena can be adjusted to suite the definition of other variables. One of the advantages is that it allows for direct manipulation and control of the independent variable. It limits any other explanations and allows direct casual relationship between variables.

One of the limitations is that it requires a laboratory for the experiment to be undertaken; this affects the outcome since the results are artificially generated. There is minimal control of variables and this may sometimes make the experiments difficult to undertake (Campbell and Stanley, 1963).

Sampling approaches

Sampling is the process where a subgroup is picked from a population and then used as the main study phenomena to reveal the characteristics of the whole population. This makes it easy for the researcher to generalize the issues about the population based on the characteristics of the sample group.

Advantage of sampling methods is that it is easier to collect data from small number of participants. Large population can be described using several variables which might prove to be very expensive (Borg, Gall, and Gall, 1993; Gall, Borg and Gall, 1996).

The following are the different types of sampling approaches used;

Random sampling

This is the non-systematic collection of samples from whole population. Every member within the population holds an equal opportunity to be included in the sample, all the process is based on chance. Example when one wants to identify preferred candidate for a post.

A list of citizens who are registered voters within the region is obtained, then the names are written on cards after which the cards are mixed and selected randomly based on chance. Advantage of this method is that it provides opportunity for participation of all subjects. This approach is preferred over others because it provides chance for general results.

For instance if there was a study to collect results from twenty people selected randomly or from 100 people selected using other methods, the small sample is found convenient to work with. Disadvantages include the fact that this method must have a list indicating all population members, it is also more expensive to conduct, time consuming and the data can be easily manipulated.

Stratified sampling

In this approach, random samples are collected from identified groups within the population. Advantage of this approach is that it ensures there is proportional presentation of members from a particular group. The disadvantages is that it’s a bit complicated and requires a lot of input since each strata must be defined in a more careful way.

This approach is used where the population comprises of distinct sub-groups that must be studied. In this case accurate estimate of each sub-group is made by taking sample of adequate size from each stratum. Estimate of the whole population can be done by first knowing the total population of each strata then adding their estimates.

Cluster sampling

This is where samples of successive groups are identified and divided into small units. Cluster sampling results in less accurate estimates of indicators as compared to random sampling. This is because participants found within the clusters may sometimes show similar characteristics.

One advantage of this sampling method is that it is possible to work with local list even without the list of population members; it is also possible to isolate members to avoid unnecessary contradictions. Disadvantage of this method is that sometimes it is difficult to “equate clusters in a level” due to differences in characteristics.

This calls for increase in the size of sample in order to find varied characteristics (Black, 1999, p 118; Wiersma, 1995). An example where it can be used is where population is spread over wide area posing some difficulty in data collection. Cluster sampling is used to concentrate fieldwork in specific clusters.

External and Internal Validity of a Study

Gay, Mills, Geoffrey and Airasian (2006) defines validity as “the degree to which a test measures what it is intended to measure; a test is valid for a particular purpose for a particular group” (p153 and 603). Validity focuses on testing and measuring the intended phenomenon.

There are four types of validity; “Content validity indicates the level to which a test measures a sample of a whole population to establish its characteristics. Criterion validity indicates the capability of predicting test scores hence estimating performance based on the appraised test. While construct validity looks at mental interpretation of test performance” (Creswell, 2008, p172; Gay et al., 2006, p.152).

Internal validity is a case where the research is examined to reveal whether it demonstrated any valid relationship between the variables under study (Patton, 1990). The study is considered internally valid in a case where the effects on dependant variable are due to the differences that exist between independent variables.

Threats to internal validity include; past events, growth, testing, human disadvantages and effects from instruments, statistical regression, selection, mortality and imitation of treatments. On the other hand External validity presents ways on which findings are generalized across entire population or subjects under investigation (Altheide and Johnson, 1994).

Threats to external validity include; “relationship between treatment and attributes, relationship between treatment and setting, conflict between multiple and treatment, pretest and post-test sensitization” (Gay et al., 2008, pp241-247).

External and internal validity are related in that as more controls are implemented which increases the internal validity; the generalization of the experiment reduces. There are several threats that affect validity; however the researcher need to take some necessary measures which include; avoid using pretest results if at all it might have some effects on the posttest results (Katzer, Cook and Crouch, 1998).

Also the length of the treatment period should be shortened when there is prediction that the objects used may not survive the treatment period set, dividing the subjects into half if at all the post-test may differ from the pre-test results. Prediction of non-performance of subjects on an experimental task leads to development of an alternative instruction manual (Brownell, 1992).

To strengthen internal validity researchers must identify all potential barriers before running an experiment. Researchers need to implement control techniques such as balancing samples through random selection. They should further use appropriate experimental designs.

Strengthening of external validity involves experimenting on a phenomenon within labs that are free from real world duplicates. Researchers also need to find out if the environment in the laboratory is the same as that in real world.

Example: A researcher compares the scores of students on discussion question versus those who do not explain their points on the same questions (Yin, 1994). If the findings indicate that those who discuss questions score significantly higher than the other group, it implies that discussion of questions is effective.

If at all the students who elaborate on questions are given more exercises on how to answer the questions than those who do not, then there is involvement of extraneous variable which weakens the quality and validity of the experiment (Cooper and Schindler, 2001; Leedy, 1997).

Example: The percentage change in the incidence of crime in the 106^th Precinct may have been the result of the normal decline in crime experienced during the fall of the year and may have had little or nothing to do with 15% increase patrol officers.

The decrease in crime experienced may be attributed to any of these factors; increased patrol effects, low crime due to season, or both factors. The study is considered to be logically confounded since there was no control on seasonal variation in crime (Friel, 2010).

Statistics

A variable is a subject that is measured in a research study; it can be in form of an object, event or any other measurable category. It is divided into independent and dependent variable (Hittleman and Simon, 2006; Creswell, 2008). An independent variable is a variable whose characteristics are not changed by other variables, it stands alone. Examples include; “age, gender and training course” (Creswell, 2008, p. 306).

While dependent variable is that variable which depends on other factors, it is subject to change depending on how it is influenced by others (Gay and Airasian, 2000, pp 571-590). A good example is “frequency of smoking” (Creswell, 2008, p. 306).

Case study I

In this case the independent variable include teaching, this is because other factors such as new technology, student’s attitude might not change the meaning. Dependent variables include; student’s attitude, technology, Professor, and students. Student’s attitude is subject to change depending on the interest towards the subject and the teacher involved.

Technology is subject to change depending on people’s ideas and its level of effectiveness. Professor is subject to change depending on how he/she adjusts to the new technology. The student’s are also subject to change depending on how involving the new technology is, they might improve on their study performance or not. This is non-experimental research since the independent variables cannot be measured but only observed.

The statistical method to be used in this study is the unpaired t-test since in both groups students and teachers are independent of each other while the teaching process remains continuous (Creswell, 1994; Glass and Hopkins, 1984). Sample size n= 56 representing only one group, sample mean u= 28 STD deviation = 31.113 Variance = 968.01

Case study II

In this case the dependent variable is the reinforcement schedules since the schedules can only be observed and not controlled. While the independent variables are the rein forcers (i.e. food, money and token) since they can easily be manipulated by the experimenter (Hart, 1998).

In this case unpaired t-test will be used since the parameters are continuous and distributed. Comparison is to be made of the continuous distribution parameter in more than two independent variables; ANOVA will be used to generalize the unpaired t-test. This is an experimental research since the independent variables can be manipulated.

Case Study III

In this case independent variable is the student achievement since it can be controlled and adjusted. While the dependent variable is their Scholastic Aptitude Test since it cannot be controlled it is natural. This is non experimental since the researcher cannot manipulate the results (Stake, 1995).

Correlation analysis is suited for the analysis of this case; this is because the study tends to examine the strength of correlation between two variables that is SAT and scores. The paired t-test is recommended for this study since there are continuous parameters is distributed and compared in more than one paired group (Giddens, 1984; Gay, Mills, Geoffrey, Airasian, 2008).

References

Altheide, D. L., & Johnson, J. M., (1994). Criteria for assessing interpretive validity in Qualitative research. In N. K. Denzin & Y. S. Lincoln (Eds.). Handbook of qualitative research (pp. 485-499) Thousand Oaks, CA: Sage. Web.

Black, T. R. (1999). Doing quantitative research in the social sciences: An Integrated approach to research design, measurement, and statistics. Thousand Oaks, CA: SAGE Publications, Inc. (p. 118). Web.

Bogdan, R., & Biklen, K. (2007). Qualitative research for education: an Introduction to theory and methods. 5th ed. Boston, MA: Allyn and Bacon.

Borg, R., Gall, P., & Gall, D. (1993). Applying educational research (3rd ed.). New York: Longman.

Brownell, A., (1992). Reprint of criteria of learning in educational Research. Journal of Educational Psychology (84): 400–404. Web.

Campbell,T., & Stanley, J., (1963). Experimental and quasi experimental designs for research on teaching. In N. L. Gage (Ed.), Handbook of research on teaching (pp. 171–246). Chicago, IL: Rand McNally. Web.

Creswell, J. W. (1994). Research designs: Qualitative and quantitative approaches. Thousand Oaks, CA: Sage.

Creswell, J. W. (2008). Educational research: planning, conducting, and evaluation Quantitative and qualitative research (3^rd ed.). Merill Prentice Hall: Upper Saddle River, NJ.

Cooper, D. R. & Schindler, P. S. (2001). Business research methods (7^th ed.). Boston: McGraw-Hill Irwin.

Friel, M., (2010). Assessing the goodness of a research study. College of criminal justice, Sam Houston state University.

Gall, D., Borg, R., & Gall, P., (1996). Educational research: An introduction. White Plains, NY: Longman.

Gay, L.R., & Airasian, P., (2000). Evaluation of a research report. In Educational research: competencies for analysis and application. (pp. 571-590).

Upper Saddle River, N.J: Prentice-Hall. Web.

Gay, L.R., Mills, Geoffrey. E., Airasian, P. (2008). Educational research: Competencies for analysis and applications. Upper Saddle River, New Jersey: Pearson Merrill Hall.

Giddens, A. (1984). In R. Yin (1993). Applications of case study research. Beverly Hills, CA: Sage Publishing.

Goetz, J. P. & LeCompte, M. D. (1984). Ethnology and qualitative design in educational research. Lexington, MA: D. C. Heath.

Glass, G. V., & Hopkins, K. D. (1984). Statistical methods in education and psychology. Englewood Cliffs, NJ: Prentice Hall.

Gliner, J. A., & Morgan, G. A. (2000). Research methods in applied settings: An integrated approach to design and analysis. Mahwah, NJ: Lawrence Erlbaum Associates.

Hart, C. (1998). Doing a literature review: releasing the social science research imagination. London: Sage Publications.

Hittleman, D., & Simon, A.J. (2006). Interpreting educational research: an introduction for consumers of research (4th ed.). Upper Saddle River, NJ: Merrill.

Holland et al., eds. (1985) Oxford Textbook of Public Health, Volume 3: Investigative methods in public health. Oxford: Oxford University Press.

Katzer, J., Cook, K.H., & Crouch, W.W. (1998). Evaluating information: a guide for users of social science research (4th ed.). Boston, MA: McGraw-Hill.

LaFountain & Bartos, R., (2002). Research and statistics made meaningful in Counselling student affairs. Web.

Leedy, P. D., (1997). Practical research: planning and design (6^th ed). Upper Saddle River, NJ: Prentice-Hall, Inc.

Miller, D.C., (2002). Handbook of research design and social measurement (6th ed). Thousand Oaks, CA.: Sage Publications.

McMillan, J. H. & Schumacher, S. (1993). Research in education: A conceptual understanding. New York: HaprerCollins.

Patton, M. Q. (1990). Qualitative evaluation and research methods (2^nd ed.) Newbury Park, CA: Sage.

Resenthal, R., & Rosnow, R.,(1999). Contrasts and effect sizes in behavioral Research: correlation approach. Web.

Stake, R., (1995). The art of case research. Thousand Oaks, CA: Sage Publications.

Wiersma, W., (1995). Research methods in education: An introduction (6^th ed.). Boston: Allyn and Bacon.

Yin, R. (1994). Case study research: Design and methods (2nd ed.). Beverly Hills, CA: Sage Publishing.