The effectiveness of a new method can be tested using paired sample t-test. The paired sample t-test was chosen due to using the same sample before and after using the new method.

Kanda Data will provide case examples for exercise materials to make it easier to understand using paired sample t-tests. An example of a detailed case study is a researcher conducting research in a school.

Researchers want to test new learning methods to be applied to students. If it proves effective and provides a significant difference, this new learning method will be proposed to replace the current learning method.

Researchers chose 20 samples at random from a school. Samples were collected in one class. Furthermore, the researchers tested the students’ performance by testing their mathematical abilities.

Next, the researcher applied a new learning method in the classroom. Based on the results of the literature study conducted by the researcher, the new learning method can encourage students to be more active in independent learning.

The period of application of this new learning method is carried out for one semester. At the end of the semester, 20 students who were selected as samples were retested for their mathematical abilities.

The results of mathematics exams for students who use existing learning methods and new learning methods can be seen in the table below:

**How to test the effectiveness of the new learning method**

Based on what I conveyed earlier, the test of the effectiveness of the new learning method compared to the existing learning method will use a paired sample t-test.

Paired sample t-test was chosen considering that the researcher used the same respondents both before and after the implementation of the new learning method.

The paired sample t-test contains assumptions that must be met; namely, the data is normally distributed and uses the same sample for pre-test and post-test.

In this article, Kanda Data does not provide a test method to prove whether the data is normally distributed or not. You can read Kanda Data’s previous article on testing the data’s normality. In this article, I assume that both variables have normally distributed data.

In the paired sample t-test, we test for differences in mean values. Therefore, the data measurement scale should have an interval/ratio data scale. How to distinguish nominal, ordinal, interval, and ratio data measurement scales can read my previous article entitled: “*Nominal, Ordinal, Interval, and Ratio Scales | Types of Data Measurement*“

**Paired Sample t-test in Excel and Interpretation of the Results**

Paired sample t-test can be analyzed easily using excel. On the Excel menu, select “data” then click “Data Analysis.” For those who do not find “Data Analysis, ” you need to activate it first. The trick is to read the article I wrote: “*How to Activate and Load the Data Analysis Toolpak in Excel*.”

After you click “Data Analysis” a Data Analysis window will appear containing options. You select t-Test: Paired Two Sample for Means as shown below:

Next, you input the Existing Learning Method Variable into “Variable 1 Range:” and the New Learning Method variable into “Variable 2 Range:”. If you input data with variable labels at once, you check Labels. Next, you need to set the alpha of 0.05 and selecting the output options will be saved. More clearly can be seen in the image below:

After you click ok, the analysis output will appear. Up to this stage, you have finished analyzing the paired sample t-test. The output of the analysis can be seen in the image below:

Based on the output above, we can see that the value of t stat = -10.227. If using two-tailed, the p-value was 3.659 E-09, so the p value was <0.05. Thus, it can be concluded that the new learning method is significantly different from the existing learning method.

Next, we see that the mean value of the existing learning method is 79.74, which is smaller than the mean value of the new learning method of 81.87. Thus, the new learning method has a higher mean value, and we can say it is quite effective and better than the existing learning method.

Well, that’s our discussion on this occasion. See you in the following article update! Hopefully, it will be helpful for all of you.