T-Test: Excel

Summary

T-tests can be extremely helpful in the world of educational research. There are two types of t-tests, independent t-tests and dependent t-tests. The former is used when comparing a difference between two groups given two continuous and categorical variables. The latter is used when comparing the same group on two separate occasions; it is used to determine whether a change has occurred or there is a difference for the group from the one occasion to the next. For example, I took the same data from the National Center for Education Statistics (NCES) that I used previously (2011 reading scores for boys and girls in the fourth grade) and conducted an independent t-test. The categorical independent variable was gender and the continuous dependent variable is the fourth grade. When conducting a t-test, one will either conclude that there is no significant statistical difference (the p value is less than .05), in other words, the null hypothesis is accepted or one will conclude that there is a statistically significant difference (the p value is greater than .05), in other words, the null hypothesis is rejected. By looking through this site I was better able to understand both types of t-tests.

What I Did

            I had previously exported the data I wanted from the NCES into Excel and accessed the Analysis ToolPak add-in. Below are the steps I took to conduct the t-test analysis.

1.      Chose Data from toolbar

2.      Scrolled down to choose T-test: Two Sample Assuming Equal Variances

3.      Filled in Variable 1 Range by clicking and dragging the cursor to select all scores earned by males (same for Variable 2 Range for females)

4.      Titled the new sheet by filling in New Worksheet Ply section

The t-test analysis opened in a new sheet.

If you don’t have Excel, you can still analyze your data in a t-test by using this free online t-test calculator.

What I Learned

Research Question: Does student’s gender impact 4^th grade reading performance?

Hypotheses:

·         Null Hypothesis- There is no statistical difference between the fourth grade readings scores earned by males and females. Gender does not impact fourth grade reading performance.

·         There is a statistical difference between the fourth grade reading scores earned by males and females. Gender does impact fourth grade reading performance.

Critical P-level (alpha): P=0.05

 

Decision rule: Reject null hypothesis since 1-tail p value > 0.05

 

Summary Statement: Reject null hypothesis, 1.611E-07 (tail p value) <= 0.05

 

Statement of Results: There is a statistically significant difference between the 2011 fourth grade reading scores earned by males and females. Gender does impact fourth grade reading performance.

Research Topic for Students Using a T-Test

            Gathering data and analyzing data for a t-test is a fairly straight forward and simple process. However, I do not believe that I would have elementary or middle school students engage in research using a t-test as the concept behind such statistical analysis is complex for children at the elementary and middle school mathematical achievement levels. However, I believe this type of research could be fun and informative for students at the high school level. For a fun, science experiment using a t-test, I might give students the following research topic.

Research Question: Does preservation method (to be determined by students) impact the life span of a carved pumpkin?

Suggested Design: Each student is given a pumpkin (granted that there is an even number of students). Each pumpkin must be similar in weight, shape, and size. Students will carve their pumpkins, creating two eyes, a nose, and a mouth. Half of the students, randomly selected, will use a particular preservation method (ex. acrylic spray) and the other students will use an alternative method (ex. pumpkin fresh spray). Students will observe their own pumpkin every day. Once the pumpkin matches this description the, profuse mold and rot (more than one student will confirm), the pumpkin will be considered deceased and the student will plug into a classroom Excel document, how many days the pumpkin lasted since its carving. The Excel document will have two columns labeled by preservation method. The students will each practice using the t-test function in Excel and the class will discuss the findings. Throughout the process, each student will maintain notes and complete a lab report.

Dependent, Continuous Variable: Life Span of Carved Pumpkins

 

Independent, Categorical Variable: Preservation Method

 

Hypotheses:

·         Null Hypothesis- There is no statistically significant difference between the numbers of days the pumpkins lasted before being confirmed deceased due to profuse mold and rot. Preservation method does not impact the life span of a carved pumpkin.  

·         There is a statistically significant difference between the numbers of days the pumpkins lasted before being confirmed deceased due to profuse mold and rot. Preservation method does impact the life span of a carved pumpkin.

 

Ideas for preservation methods can be found on this site.

Standards Reflection

Conducting educational research using digital age tools towards a goal of evaluating and reflecting upon teaching practices in order to better support student learning meets standard Five "c" of the ISTE-NETS-T’s standards.

Adapting educational experiences by incorporating student use of digital applications such as Excel meets standard two of the ISTE-NETS-T’s standards and its components.