Free Essay Sample: Significance Test Vs Multivariate Techniques

Multivariate techniques for data analysis refers to those statistical methods that are used to analyze data arising from two or more variables. Essentially, these techniques model reality whereby every product, decision, or situation is influenced by more than one variable. Significance test on the other hand, is used when determining whether a null hypothesis stands rejected or not, in favor of an alternative hypothesis. Drawing from the above definitions, significance test would better serve me in the study, than multivariate techniques. The aim of the study is to ascertain whether there is a correlation between domestic terrorists and the recruitment methods of street gangs. This aim serves as a null hypothesis, and the probability value that will be obtained from a significance test will reflect the collected evidences strength against the hypothesis. Multivariate techniques analyze the relationship between multiple variables simultaneously (Birkett, 2016). This does not concur with the study as it only involves two variables; domestic terrorists and the recruitment methods of street gangs. Multivariate techniques would have been the most suitable option if the study involved more than one independent or dependent variable, or both. It is only the best choice where it has been hypothesized that a certain aftermath is effected by more than one mechanism. Analysis aside, multivariate techniques require considerably more investment on the setup and design, and they can be a big burden when it comes to my conversion budget (Frane, 2015) i.e. time, informants, and people. Furthermore, since I am not doing optimization with a number of variable, I am confident that significance test is the right option because the speculated changes are large.

Is your time best spent reading someone else’s essay? Get a 100% original essay FROM A CERTIFIED WRITER!

Order now

Considerations Relative to Analysis Techniques

Relative to the techniques of analysis, I considered the objectives of the sampling that was to be done in the study. The objectives of sampling were to; reduce computing and analytical requirements and to examine the representative sub-sets of the collected data. Another consideration that I made is the variability indicators. If separate variability indicators were present, they would have directed the sampling procedure to gather data over a longer duration. The issue of biased estimates was also a major consideration. Biased estimates are always either systematically higher or lower than the actual population value, and this is due to the fact that they are not a true representation of the population sample. Biasness in never easily detected, and as such precision is highly valued. This therefore urged me to apply appropriate stratification. The data collected also has to be as representative as feasibly possible. There also needed to be a solid and well laid out preparation for the study so as to increase the validity of the analysis. There also has to be a valid data set. Another consideration is that there needs to be proper statistical procedures. Finally, there needs to be an accurate interpretation of the gathered results. All valid data analyses have their foundation on how gathered results are interpreted. One has to clearly understand their study and the statistical techniques, together with how to bring the results into the context of the project. The interpreted results also need to be summarized to an extent that anyone, even non-statisticians can understand. On overall, the type of data associated with the study is the biggest consideration that I made relative to the techniques of analysis.

References

Birkett, A. (2016). When To Do Multivariate Tests Instead of A/B/n Tests. ConversioXL.

Frane, A. (2015). Power and Type I Error Control for Univariate Comparisons in Multivariate Two-Group Designs. Multivariate Behavioral Research.