Unit of Analysis in a Research Project

Unit of Analysis
A unit of analysis constitutes an object of study in research undertakings (Cole, 2018). Definition and bounding of the case may prove challenging since a number of variables and points of interest overlap and intersect within cases researches. Development of research questions, propositions for case selection, focus identification, and boundary refining has been recommended for effective establishment of the aforementioned components in the study design. Case bounding proves crucial when it comes to information acquisition and analysis management, focusing, and framing. This entails selectiveness and specificity in the identification of case parameters such as respondents, process and location, in addition to the establishment of a timeframe for investigation of the case. To be more specific, units of analyses (UoIs) are decided by research questions (Merriam, 2009; Stake, 2006; Yin, 2014). Consider, for instance, Francis, Anderson and Stokes’ 1999 research “City Markets as a Unit of Analysis in Audit Research and the Re?Examination of Big 6 Market Shares”; here, city markets were the UoI. The researchers utilized market shares of big 6 companies on the basis of aggregate national-level data for inferring industry expertise and market leadership, besides distinguishing the big 6 accounting companies from each other. 
Incorrect Unit of Analysis
Ecological fallacy is a widely occurring mistake one will get to witness in regards to causality as well as UoI. This happens when a researcher makes claims pertaining to a lower-level UoI on the basis of information from a higher-level UoI. In a large number of instances, this happens when making claims at the individual level, when information has been collected for the group. For instance, one may wish to gain insights into whether or not addiction to electronic gadgets is greater among students of particular campuses as compared to others. Perhaps, the various campuses across the nation have offered the figure for their respective share of gadget addicts, and this information teaches us that gadget addiction occurs more widely in campuses with business programs as compared to those without. It may subsequently be concluded that students enrolled in business programs will more likely become gadget addicts as compared to non-business pupils. But this wouldn’t be the proper conclusion as having only campus addiction rates can only facilitate conclusions on campuses; one cannot possibly find out about individual pupils of these campuses based on those statistics. Perhaps business schools’ sociology majors might be responsible for the high addiction rates there; the point being, only campus-level information cannot help us gain insights into individual enrollee behavior and making such conclusions on students when the information collected pertains to campuses gives rise to ecological fallacy risks (Open Textbooks, 2016).
The second potential error is reductionism, which occurs if one makes claims regarding a higher-level UoI on the basis of information from a lower-level UoI. In such an instance, macro- or group-level claims are made on the basis of individual-level information (Open Textbooks, 2016). This study utilizes pupils from two separate campuses as UoIs.
Sample Size and Statistical Power (SP)
SP analysis aids scholars in choosing the right sample size for facilitating reliable, accurate statistical judgments. Power analysis reflects the likelihood of the research to have a significant impact. This is why it has to be employed in experiment planning. Power analysis proves highly valuable in quality inspections. Here, power analysis reflects the likelihood of an observed process having or lacking a specific quality. Power value is dependent on the following three factors – alpha, sample size, and sample effect. A research scholar needs to find the right levels of the aforementioned factors for attaining satisfactory power. In a majority of instances, power value will be targeted at eighty percent. However, this lacks a scientific basis and is dependent on the case, study purpose and object (Dumi?i? & Žmuk, 2012).
G*Power represents a statistical analysis software package which aids quantitative research scholars in performing analyses of sample size (Faul, Erdfelder, Buchner, & Lang, 2009). SP analysis by utilizing v. 3.1.9 of the G*Power software may help decide on appropriate research sample size. For the current research, a priori SP analysis which assumes medium effect size (f 2 = .15), ? = .05, may identify the minimum size of sample required for achieving a power value of .80 as 68; incorporation of 146 respondents is linked to an increased power value of .99. Thus, 68 to 146 subjects will be sought for the research (Trochin, 2006).
Justification of Unit of Analysis
The research question being “Which students will display maximum likelihood of gadget addiction?”, the individual is the UoI for this study. Students on the campus may be sent surveys via mail. The objective would be student classification based on social class, for ascertaining correlations of class and gadget addiction. We may, for instance, discover that new media majors, high socioeconomic background students, and male students are more inclined to become gadget-addicted as compared to others. One could also pose the question of the similarities and differences between student gadget addictions. Here, gadget addicts could be observed and where, when, how and why they utilize gadgets could be recorded. In both instances (i.e., survey and observation), information will be gleaned from individual pupils. Therefore, the individual represents the unit of observation in both instances. However, the UoIs are different for the two researches. In the former, the goal is describing individual student characteristics followed by making generalizations regarding the populations they belong to; nevertheless, the individual remains the UoI (Open Textbooks, 2016).
References
Cole, N. L. (2018). Units of analysis as related to Sociology. Retrieved from https://www.thoughtco.com/wh-units-of-analysis-matter-4019028
Dumi?i?, K., & Žmuk, B. (2012). Use of power analysis in choosing appropriate sample size for quality inspection. Poslovna Izvrsnost Zagreb, God, VII.
Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analysis using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41(4), 1149–1160. doi:10.3758/BRM.41.4.1149
Francis, J. R., Stokes, D. J., & Anderson, D. (1999). City markets as a unit of analysis in audit research and the re-examination of big 6 market shares. ABACUS, 35(2), 185–206. Retrieved from http://onlinelibrary.wiley.com/journal/10.1111/(ISSN)1467-6281
Merriam, S. B. (2009). Qualitative research: A guide to design and implementation (2nd ed.). San Francisco, CA: Jossey-Bass.
Open Textbooks. (2016). Units of analysis and units of observation. Retrieved from http://www.opentextbooks.org.hk/ditatopic/29184
Stake, R. E. (2006). Multiple case study analysis. New York, NY: Guilford.
Trochin,W. M. K. (2006). Selecting statistics. Retrieved from http://www.socialresearchmethods.net/selstat/ssstart.htm
Yin, R. K. (2014). Case study research: Design and methods. Los Angeles, CA: Sage.