The Use of ANOVA Research Method - Essay Example

ANOVA Research Method Introduction Analysis of variance (ANOVA) statistical technique in research is a method that is appliedin testing statistical significance difference that exists between two or more mean. However, variance in statistics and probability distributive theory is used to measure the differences in characters or qualities a particular set of the group possesses or spread out (Hanneman, Robert 2013, 259). ANOVA statistical method is termed as analysis of variance rather than the analysis of the mean due to the statistical inferences obtained in the research about the mean is obtained by critical and creative statistical analysis of the variance. For example, ANOVA method is applied, in general, statistical difference rather than distinct difference of the research mean (Hox 2010, 35). It is worth noting that, at zero variance, all the values measured are regarded to identical even through variance always maintains a non-negative value. Therefore, as the ANOVA method is used to analyze the difference between several or a single variable, a small variance obtained in the statistical significance difference illustrates that the values or data points obtained are very close to the expected mean. Similarly, High statistical significance difference data points justify that the data values obtained in the research spread out from each other or spread around the mean (Woodward 2014, 210). On a broad aspect, the variance is usually divided into various components by utilizing ANOVA method that is then attributed to various types of variations in different sources of the variations obtained. In simple words, ANOVA method has its origin in experimental studies. However, ANOVA method uses four basic assumption that errors obtained in the statistical data analysis are normally distributed and independent (Salkind, Neil 2010, 34). In addition, it is always assumed that, all variance errors are equal, and the expected statistical values of the errors obtained are equal to zero. In simple terms, the ANOVA method produces a statistical test that determines as to whether the mean obtained from several tested groups are equal or whether they are not equal thereby providing a generalized t-test for more than two evaluated groups. In this regard, ANOVA method is used in providing a comparative test for more variables for the purpose of estimating a reliable statistical significance (Myers, Jerome 2010, 271). Fig 1. Illustrations of a two-way mixed ANOVA ANOVA Research Question/Hypothesis ANOVA method is a very powerful parametric and inferential statistic technique that can find differences or reject a null hypothesis among different variables or groups compared. Therefore, during ANOVA research questions and hypothesis formulation, variance homogeneity assumption, standard distributions of the data scores and equal group sizes assumptions are always considered (Lomax, Richard 2012, 185). The ANOVA statistical technique questions will always intend to answer a null hypothesis that provides a comparative relationship. For example, the ANOVA questions indicate that there is no difference that can be identified or tested between various groups (Myers, Jerome 2013, 17). Therefore, the statistical significance difference in the test outcomes obtained is of interest compared to other parameters obtained in the ANOVA statistical method of research. In formulating the ANOVA hypothesis, mean comparison of two populations that constitutes independent variables samples employs a 2-sample t-test illustrating a pooled variance when the variances used in the population sample is assumed to be equal. However, for more than two variables or populations the hypothetical statistic test is determined by the ratio between the expected mean or sample variance (Lavrakas, Paul 2008, 29). For the essence, under the null hypothesis, the ratio is close to one, but for cases where the ratio is large the null hypothesis is rejected. ANOVA method is used in formulating hypothesis or research questions that relate to a mean difference of two conditions. For example, the independent variable has qualitative characteristics and the dependent variable is qualitative. However, within the groups under study, ANOVA method is used when the same population sample or participants are used both in the independent variable conditions. Similarly, in longitudinal studies, ANOVA method is used when between the groups when the different research participants are used are used in the independent variable conditions (Lininger, Monica 2012, 274). In cross-sectional research questions/hypothesis, ANOVA method is used to provide a casual research hypothesis if there is presence of random recruitment or assignment of participants with no confounders and prior to statistical manipulation of the independent variables in the study. In this regard, the ANOVA method cannot be used in designing attributive research questions and hypotheses even though it can be used in associative studies questions anytime (Lomax, Richard 2012, 326). Therefore, the type of the research hypothesis used is univariable while ANOVA research method is bivariant. Conversely, when formulating associative research questions between one qualitative and quantitative variable, especially when the participants in each of the two distinctive groups is different. In this type of the research questions or hypotheses, the ANOVA method will be used to compare and assess each unique group characteristics assuming that the dependent and the independent variables remain constant (Sreejesh, Sanjay 2014, 27). For example, boys will have higher average mean on exercise participation chat than girls. In this regard, the Research hypotheses used in this kind of research will be directed towards analyzing an associative research rather than attributive research hypothesis. Similarly, in research theory that consists of two groups that are subjected to different treatments conditions, BG ANOVA method is used correctly in determining the mean difference between the two groups under study. From a similar account, the WG ANOVA method is used in formulating research hypotheses involving two groups receiving similar treatment (Martin, William, 2012, 53). The WG ANOVA method used for the studying the research variables will be used to find out the research mean the difference of the same treatment outcomes in the study. Nevertheless, when formulating research questions for two groups of study with different research participants that need the Dependent variables means to be compared with different independent variables. The ANOVA method is used in making comparisons with research. For example, the ANOVA method can be used in comparing same group participants’ characteristics in both conditions (Grissom, Robert 2012, 151). Similarly, when hypothesizing the mean difference between a qualitative and quantitative variable, a between group ANOVA method is used in formulating questions that are used in exposing the research participants to different conditions enabling a comparative study of the mean differences of the variables used in the research. Within the group, ANOVA formulated hypothesis is used in a repeated design measure in providing a comparison between qualitative and quantitative variables for analyzing the mean difference of all participants exposed to all conditions. For example, the ANOVA method can be used when there is a disparity in participants’ sizes. Similarly, group one is greater, equal or less than the other variable making the research questions or hypothesis to be null (Salkind, Neil 2010, 136). Fig 2 Illustration of ANOVA Statistical tests Research Design Issues Associated with ANOVA The ANOVA study design is derived from a statistical experimental design that manipulates variables with the aim of determining the effects attained by measuring the responses. The ANOVA methods employ only a suitable small number of treatment since ANOVA method uses a comparison mean of more than one sample. Therefore, the ANOVA method is splinted in variation between and within sample providing room for the data analysis base on a one-way or two-way ANOVA where residual variation is the variation in the sample size (Wagner, William 2007, 292). In this regard, the ANOVA method consists of three parameters that include replication, randomization and local control of data. Randomization of variables or participants is done to eliminate the biases, even though; there are unexplained variations during randomization and participant selection and observation fitting. ANOVA randomization of data will intend to maintain independence among various variable by the application of valid, significant tests. Treatments are assigned at random completely such that each unit receiving a certain treatment has an equal property in receiving one of the intended treatments. However, assumptions have to be made that the treatments are normally distributed or homogenous. The ANOVA method provides restrictive measures in assigning treatment (Hanneman, Robert 2013, 172). From a different perspective, the factors and levels in ANOVA independent variables are manipulated by the research for response determination. For example, two difference groups with different or similar variables are manipulated and monitored for the response generated. The number of experimental units in the ANOVA research as issues is programmed to fit and satisfy the intended needs or goals of the experiments. As a preparatory ANOVA analysis procedure, the ANOVA experimentation is always s sequential. Prior ANOVA experiments are performed for the final experiment in order to eliminate experimental errors, biases as well as establishing an estimate the treatment effects are likely to be encountered. Within the number of experimental units, the final ANOVA experiments are used to hypothesize how a specific treatment effect has a corresponding significance magnitude. The power analysis in the ANOVA research design is used with the aims of assessing the probability or chances of rejection a research hypothesis successfully. If certain ANOVA design methods are assumed, the research significant level, effect size as well as the sample size will be effectively be determined (Walker 2010, 316). It is worth acknowledging that, the power analysis component of the ANOVA research design increases chances of obtaining an increased confidence interval as well as determining the study sample size that will be used in rejecting the Null hypothesis at a point when the alternative hypothesis is thought to be true. Conversely, the effect size of the ANOVA research design provides a quick summary of the existing strength associated with dependent variables and the predictors. Studied and discipline findings comparisons are facilitated by the standardized effects sizes (Salkind, Neil 2010, 274). Follow-up analysis as an issue in the ANOVA research design is helpful in pointing out all the outliers. Nevertheless, the follow-up analysis the research design in responsible for the errors evidenced in the research results. In addition, the follow-up analysis creates disproportional effects in relation to the statistical conclusion generated in the final research (Hair, Joseph 2008, 217). Model confirmation within the ANOVA research method ensures fulfilment of all the assumption that needs to be closely monitored in the research. For example, the model confirmation is achieved by ensuring that, the results are correct when the use of the ANOVA method on assumption is equal to the standard deviation. The follow-up test demonstrates statistical significant effects within the ANOVA research design that is followed up subsequent tests. In this regard, type I errors are controlled as simple and compound comparisons are used to tests the results trends (Grissom 2005, 68). The one-way ANOVA method will be used in measuring the results between different two or more independent groups that are used to provide reliable information. The ANOVA extended logic consists of the hypothesis testing and the partitioning variance source which are used independently. Additionally, ANOVA regressions provide a more complex model of data collection as well as supporting other statistical tools. However, ANOVA research design applies a repeated design that limit the number of subjects enabling the statistical inferences to be achieved by a limited number of subjects (Dowdy 2011, 56). In this regard, the variance estimations for the subjected treatment effect in the research are significantly reduced by the repeated measure design during the experiment. Similarly, the research efficiency is greatly improved. Fig 3 illustration of analysis of variance ANOVA makes applies the F statistics in conjunction with other analysis of the variance tests to determine the significance of the research. Therefore, depending on the subject factor and the assumptions violation, the ANOVA statistical methods applies standardized Univariate ANOVA test, alternative Univariant test, and the multivariate tests. Collectively, the two methods are helpful in improving the validity and reliability of the hypothetical research question. The statistical effect size obtained during the research applied the multivariate tests when the set assumption rules had been broken (Basso 2009, 214). Based on this assumption, the randomness, and the normally distributed population are interconnected with the normality and spherical. It is worth admitting that, ANOVA method is a helpful method that should be applied only with caution since the method is vulnerable exposed to imputations, missing values and violation of hypothetical and sphericity assumptions. In addition, the ANOVA method experiences type sampling or randomization bias as well as inflated rates (Arthur 2012, 76). Conclusion The ANOVA statistical method is a complex method that can only be understood through a collective bargain. Therefore, the method should be used cautiously for analyzing a statistical model of a given research question /hypotheses. All the ANOVA participants are involved in the testing variable. It is worth acknowledging that, there are two distinctive study groups consisting of different participants in each group. Therefore, the type of study compares independent and dependent variables for manipulation and control of the result outcome. For example, girls will have higher average mean on Eating Habit questionnaire completed to man. Between-group ANOVA, participants in different groups is subjected to different treatment conditions for the research. For example, participants’ cognitive behavioural therapy for anxiety and the Choir singing groups to aid in anxiety. Similarly, within group ANOVA is applied within groups’ design where all the participants under study are subjected to all study treatment conditions. For example, anxiety values /scores are taken be before and after subjecting the participants to therapy. Longitudinal studies apply within group ANOVA while a cross-sectional test of a specific target population is determined by between groups ANOVA. For example, between groups ANOVA can be used in predicting the outcome of a particular treatment results for a repeat of the second treatment for bettering the result. However, ANOVA research method reliable in associative research hypothesis for causal research hypothesis when the selected participants for treatment are assigned randomly. Nevertheless, the ANOVA cannot be used in the presence of confounders or attributive research. Bibliography Arthur, James. 2012. Research methods and methodologies in education. London: Sage publications. Basso, Dario. 2009. Permutation tests for stochastic ordering and ANOVA theory and applications with R. New York: Springer. Dowdy, S. M., Stanley Wearden, and Daniel Chilko. 2011. Statistics for Research. Hoboken: John Wiley & Sons. Grissom, Robert J., and John J. Kim. 2005. Effect sizes for research: a broad practical approach. Mahwah, N.J.: Lawrence Erlbaum Associates. Grissom, Robert J., and John J. Kim. 2012. Effect sizes for research: univariate and multivariate applications. New York: Routledge. Hair, Joseph F. 2008. Essentials of marketing research. Boston: McGraw-Hill/Higher Education. Hanneman, Robert, Augustine J. Kposowa, and Mark Riddle. 2013. Basic statistics for social research. San Francisco, Calif: Jossey-Bass. Hox, J. J. 2010. Multilevel analysis techniques and applications. New York: Routledge.. Lavrakas, Paul J. 2008. Encyclopedia of survey research methods. Thousand Oaks, Calif: SAGE Publications. Lininger, Monica R. 2012. Examining the research designs and analytical techniques in athletic training from 2005-2010. Thesis (Ph.D.)--Western Michigan University, 2012. Lomax, Richard G., and Debbie L. Hahs-Vaughn. 2012. An introduction to statistical concepts. New York: Routledge. Martin, William E., and Krista D. Bridgmon. 2012. Quantitative and Statistical Research Methods From Hypothesis to Results. Hoboken: John Wiley & Sons. Myers, Jerome L., Arnold D. Well, and Robert F. Lorch Jr. 2013. Research Design and Statistical Analysis Third Edition. Hoboken: Taylor and Francis. Myers, Jerome L., Arnold Well, and Robert Frederick Lorch. 2010. Research design and statistical analysis. New York: Routledge. Salkind, Neil J. 2010. Encyclopedia of research design. Thousand Oaks, Calif: Sage. Sreejesh, S., Sanjay Mohapatra, and M. R. Anusree. 2014. Business research methods: an applied orientation. Heidelberg: Springer. Tabachnick, Barbara G., and Linda S. Fidell. 2007. Experimental designs using ANOVA. Belmont, CA: Thomson/Brooks/Cole. Wagner, William E. 2007. Using SPSS for social statistics and research methods. Thousand Oaks, Calif: Pine Forge Press. Walker, Glenn A. 2010. Common statistical methods for clinical research with SAS examples. Cary, N.C.: SAS Institute. Woodward, M. 2014. Epidemiology: study design and data analysis. Read More