T Tests in Quantitative Research

T-tests in Quantitative Doctoral Business Research
Quantitative research is one of the methodologies that is commonly used in doctoral business research. The use of this approach is attributable to the availability of more data that requires analysis to help generate competitive advantage in the business field. The use of quantitative research entails conducting statistical analysis, which involves the use of different methods such as t-tests and ANOVA. T-test is used in hypothesis testing in quantitative studies to determine whether variations between the averages of two groups is unlikely to have emerged because of a random chance in selection of a sample. In essence, t-tests help to compare whether two groups have varying average values. In light of the role and significance of the assumptions underlying each parametric test, this paper provides a comparison of one-sample, paired-samples, and independent-sample t-tests within the context of quantitative doctoral business research. The comparison is based on a qualitative research proposal.
One-sample, Paired-Samples, and Independent-Samples T-tests
As previously indicated, t-tests are used in quantitative research evaluate whether two groups have varying average values. In this regard, t-tests help to compare two means to evaluate whether they come from the same population. One of the underlying assumptions in t-tests is that both groups have relatively equal variances and are normally distributed. However, when a two-sample t-test is conducted, it is presumed that two groups have relatively equal variances, while the other does not (Lumley et al., 2002).
One-sample t-test is used to compare the average value of one group to a single number or to compare a sample mean to an already identified population mean. The comparison is geared towards determining whether the variation between the two means occurred by chance only or is statistically significant. In quantitative doctoral business research, one-sample t-tests comprises two types of hypotheses i.e. null hypothesis and alternative hypothesis. While the alternative hypothesis assumes the existence of some variations between the actual mean and the comparison value, the null hypothesis presumes that no variation exists. On the contrary, paired-sample t-tests are used to compare two sample means from diverse populations whose members have been paired or matched. Additionally, this t-test is used to compare two sample means from one population on the same variable, but at two different time periods like a pre-test and post-test (Empirical Reasoning Center, 2018). In quantitative doctoral business research, paired-sample t-tests are used when an observation is one group is matched with a correlated observation in another group. Independent-sample t-tests are used to compare two sample means from diverse populations on the same variable. Unlike paired-sample t-tests, independent-sample t-tests do not match members or attributes from the different populations.
Qualitative Research Proposal
An example of a qualitative research proposal that would help in comparison of the above t-tests is the research proposal on international business knowledge transfer and execution within multinational corporations in China (“Example Research Proposal”, n.d.). The research proposal question is how does Chinese multinational corporations (MNCs) implement knowledge transfer to ensure international business success. A research example related to the research proposal question is examining the relationship between knowledge transfer and international business success in multinational companies. Given the complexity in the operations of MNCs, the link between knowledge transfer and global business success should be empirically tested.
The research problem can be addressed through conducting a quantitative study to test the relationship between these two research variables i.e. knowledge transfer (independent variable) and global business success (dependent variable). Quantitative research will entail obtaining comparable data from two different contexts that act as empirical settings for testing hypotheses using t-tests. An example of one-sample t-test for this research is comparing the whether success rate of multinational companies (MNCs) that have adopted knowledge transfer exceeds the global 92.5% success rate of MNCs. An example of a paired-sample t-test is comparing whether MNCs that have adopted knowledge transfer in China and Finland have similar success rate. The average success rate of MNCs in China and those in Finland will be taken and compared to see whether the average is statistically significant. An independent-sample t-test for this research problem is comparing the success rate of MNCs that have adopted knowledge transfer with the success rate of MNCs that have not adopted this concept. This t-test will examine whether the difference between MNCs that have adopted knowledge transfer and MNCs that have not is statistically significant.
Assumptions in Independent-sample T-tests
Independent-sample t-tests are associated with some assumptions whose violation would have implications on the research. First, these t-tests assume that no relationship exists between the observations in every group or between the two groups i.e. independence of assumptions. If this assumption is violated, another statistical test should be used. Secondly, these t-tests assume that there should be no significant outliers (Laerd Statistics, 2018). If this assumption is violated, the validity of the results is reduced. Third, independent-sample t-tests are based on the assumption that the dependent variable is nearly normally distributed within every group. If this assumption is violated, valid results can still be produced.
When the independence of assumption is violated, the researcher should use another statistical test other than independent-sample t-test. In case the assumption of no significant outliers is violated, the researcher should utilize SPSS Statistics to run a test on the data in order to detect possible outliers before proceeding with the test. When the assumption of normality of the dependent variable is violated, the researcher can use several tests such as Shapiro-Wilks test of normality or graphical techniques like Q-Q Plot to address it.
In conclusion, t-tests are statistical tests that are commonly used in quantitative studies, particularly in doctoral business research. There are three major t-tests i.e. one-sample t-tests, paired-sample t-tests, and independent-sample t-tests. These tests differ in how the compare the mean or group averages to determine statistical significance. The use of these t-tests when conducting quantitative research helps to determine statistical significance in group averages or mean, which in turn helps to answer the research problem or research question.

References
Empirical Reasoning Center. (2018). Hypothesis Testing: T-Tests. Retrieved from Barnard College website: https://erc.barnard.edu/spss/t_tests
“Example Research Proposal.” (n.d.). Business School. Retrieved from University of Edinburgh website: https://www.business-school.ed.ac.uk/__data/assets/pdf_file/0020/54821/Example-Research-Proposal.pdf
Laerd Statistics. (2018). Independent T-Tests Using SPSS Statistics. Retrieved September 17, 2018, from https://statistics.laerd.com/spss-tutorials/independent-t-test-using-spss-statistics.php
Lumley, T., Diehr, P., Emerson, S., & Chen, L. (2002). The importance of the normality assumption in large public health data sets. Annual Review of Public Health, 23(1), 151–170. doi:10.1146.annurev.publheath.23.100901.140546