Data is the most important source of insight in decision making. It gives a clear view of the problem at hand while acting as the guide towards a viable solution. For this reason, hypotheses testing becomes a key concept in analyzing and interpreting data. Hypothesis refers to the proposed explanation made based on the available but limited evidence as the beginning point for further investigation. Hypothesis testing is therefore crucial in business settings because it allows professionals to test theories and assumptions before turning them into actions. This paper aims to discuss the general process of hypothesis testing the significance of hypothesis testing in business settings.
The General Process of Hypothesis Testing
The essence of hypothesis testing is to determine whether there is a match between what we think might be the case about a scenario and what exactly the truth is about. The process can be complex, but is generally divided into three phases; hypothesis generation, testing and evaluation (Koehler & Harvey, 2008). Take for example, a health care setting. Hypothesis generation may involve the physician producing at least one hypothesis or assumption about the patient’s pattern of symptoms. The self-generated assumption explains what might be true of the pattern of symptoms. Once the physician has developed a hypothesis, the next stage is to test it. Once the hypothesis is tested, the next stage is to determine whether the test results confirm or disconfirm the results.
Hypothesis development is not, however, confined to health care setting only. It can be applied in almost every aspect of life, more so in the business world. More examples could include determining whether a promotional campaign generated the expected results or not. Organizations engage in hypothesis development in the effort to make sense of data, whether from marketing, sales or hiring processes. Hypothesis development depends on the method of research, that is, whether the research method is qualitative or quantitative (Toledo et al., 2011). With qualitative research, the researcher poses questions while in quantitative method, key is to use questions and hypothesis to focus the purpose of study. The hypothesis gives the researcher the direction to collection and interpretation of data. This process moves forward to the evaluation of the data through deductive and inductive approaches.
Once the hypothesis is developed and deductive or inductive reasoning approach used, the next stage is hypothesis testing. Hypothesis testing is crucial in a variety of business situations. The process entails deducing the consequences that the research expects to be visible if the hypothesis is correct (Mourougan & Sethuraman, 2017). For instance, if a decrease in price during a promotional campaign is expected to increase revenue in the prevailing months, then it is expected that there will be an increase in the number of customers. The next step might involve selecting the research methods that will allow the observation that needs to be done to confirm the occurrence of these consequences. The next stage is to gather necessary data to assess whether the hypothesis is supported. It is possible for this stage to culminate into two possibilities; that nothing happened (null hypothesis) or something happened (alternative hypothesis). Nonetheless, hypothesis evaluation is very critical and may determine the significance of the entire process.
Basic Steps Included in Conducting a Hypothesis Test
State the Hypothesis
The process begins by stating the null hypothesis which is presumed to be true. For example, the marketing team may state the null hypothesis that more promotional campaigns during the off-peak season can increase revenue. The basis is to determine whether it is true that such action by the marketing team can indeed increase revenue. Notably, the essence of testing hypothesis is to promote the null hypothesis into a fact level or disprove it all together. The aim of a scientist or whoever is involved in hypothesis testing is not verification, but the falsification of the initial hypothesis (Banerjee et al., 2009). It is therefore crucial to have the opposite of the assumption, which describes the alternative hypothesis. For example, the marketing team may determine after collecting data that more marketing campaigns in the off-peak season does not increase revenue as postulated, hence confirming the alternate hypothesis and disapproving the null hypothesis.
Set the Criteria for a Decision
The second step in hypothesis testing is to set the criteria for a decision by stating the level of significance for a test. This resembles the criterion used by jurors in criminal trial by decision that depends on whether the evidence presented proves guilt beyond reasonable doubt. In hypothesis testing, the collected data is used to prove the null hypothesis based on the possibility of selecting a representative sample.
Compute the Statistic
To make a decision out of the data, a test statistic is used to determine the likelihood of the sample outcome if null hypothesis is true. By definition, “the test statistic is a measure that allows us to assess whether the differences among the sample means (numerator) are more than would be expected by chance if the null hypothesis is true” (Sullivan, n.d.). A test statistic indicates how far a sample mean is from the population mean. The larger the test statistic value, the farther the distance the sample mean is from the population mean in the null hypothesis.
Make a Decision
The value obtained from the test statistic helps in making a decision about the null hypothesis. The decision assess uses the possibility of getting the sample mean provided that the stated value in the null hypothesis is a fact. If the probability is less than 5% with a true null hypothesis, then the null hypothesis is rejected. If the probability is greater than 5%, the null hypothesis is true.
The Scientific Reasoning behind Hypothesis Testing
The logic behind hypothesis testing is to check whether the data sample is typical or atypical compared to the population and based on the null hypothesis. Data sample refer to a small portion of the population. The population is hypothetical and is assumed to obey the null hypothesis (Emmert-Streib & Dehmer, 2019). With the numerical value of test statistic that is a representative of the data sample and a sampling distribution that represents the population, it is possible to make a comparison of both to evaluate whether the null hypothesis is true or not. From this comparison, we get p-values, which can be used to the similarity or difference of the configuration provided the null hypothesis is true. Then, a decision is made using this p-value. Hypothesis is one of the most significant concepts in statistics because it can give a better explanation of a scenario or tell the possibility of a positive effect from a certain treatment. In short, if an organization wants to prove that the data in its hands is statistically significant, and not a possibility of a chance alone, such a process is crucial.
Scientifically, hypothesis testing is done on a representative sample. Because the population is usually large to be included in the process, the process of obtaining the sample is also critical. The researcher tests the data from the sample to determine if it is sufficient to claim other characteristics of the sample population. The null hypothesis is the working hypothesis and can be accepted or rejected. If rejected, the alternate hypothesis is accepted by default. The null hypothesis is rejected when the p-value is less than the alpha, which is most cases is usually 5%. Interpretation of the statistic can result into type I or type II errors. As described by Emmert-Streib and Dehmer (2019), type I error is also called false positive and occurs when the null hypothesis which is actually true gets rejected. Type II error is also called false negative and results when the researcher fails to reject the null hypothesis that is actually false. Hypothesis testing, thence, plays a crucial role in decision making among others aspects of life because it guides in making inferences from a piece of statistic.
‘If a hypothesis test is carried out using 5% level of significance against a specific alternative with a power of 90% and the null hypothesis is rejected’, the probability that it is actually true is 0.05. In this case, the probability of type I error is 0.05. Usually, the probability of a type I error is the alpha value. For example, of using the significance level of 1%, then 1% of the time we will reject the null hypothesis when it is true. In this case, the power value is ignored because it is the probability of making a correct decision. That is, to reject a false null hypothesis. It is used to indicate the probability of avoiding a type II error.
This paper has provided a primer on hypothesis testing. In business settings, this is a critical concept that carries huge significance in decision making and other activities. Businesses’ reliance on data to make crucial decisions necessitates the need to make the correct inferences from such statistics. It is a process that requires defining the hypothesis, setting the criteria for decision making or significance level, computing the statistic and finally making the decision. As organizations move towards data-driven practices to for competitive purposes, using such a process can be a key asset. Going forward, businesses will require more scientific analysis of statistics, which is an inspiration to small and large organizations to start practicing such methods before it becomes the norm.
Banerjee, A., Chitnis, U., Jadhav, S., Bhawalkar, J., & Chaudhury, S. (2009). undefined. Industrial Psychiatry Journal, 18(2), 127. https://doi.org/10.4103/0972-6748.62274
Emmert-Streib, F., & Dehmer, M. (2019). Understanding statistical hypothesis testing: The logic of statistical inference. Machine Learning and Knowledge Extraction, 1(3), 945-961. https://doi.org/10.3390/make1030054
Koehler, D. J., & Harvey, N. (2008). Blackwell handbook of judgment and decision making. John Wiley & Sons.
Mourougan, S., & Sethuraman, K. (2017). Hypothesis development and testing. IOSR Journal of Business and Management (IOSR-JBM), 9(5), 34-40.
Sullivan, L. (n.d.). Hypothesis Testing – Analysis of Variance (ANOVA). https://sphweb.bumc.bu.edu/otlt/mph-modules/bs/bs704_hypothesistesting-anova/bs704_hypothesistesting-anova_print.html
Toledo, A. H., Flikkema, R., & Toledo-Pereyra, L. H. (2011). Developing the research hypothesis. Journal of Investigative Surgery, 24(5), 191-194.