Home About us Editorial board Ahead of print Current issue Search Archives Submit article Instructions Subscribe Contacts Login 
  • Users Online: 205
  • Home
  • Print this page
  • Email this page


 
 Table of Contents  
BIOSTATISTICS
Year : 2015  |  Volume : 1  |  Issue : 1  |  Page : 27-28

Student's t-test for independent samples


The Research Institute, Temple University/St. Luke's University Health Network School of Medicine, 801 Ostrum Street, Bethlehem, Pennsylvania 18015, USA

Date of Submission16-Sep-2015
Date of Acceptance28-Nov-2015
Date of Web Publication29-Dec-2015

Correspondence Address:
Jill C Stoltzfus
Temple University/St. Luke's University Health Network School of Medicine, 801 Ostrum Street, Bethlehem, Pennsylvania 18015
USA
Login to access the Email id

Source of Support: None, Conflict of Interest: None


Rights and PermissionsRights and Permissions
  Abstract 

When analyzing data for two independent groups (e.g., males vs. females), the Student's t-test is commonly used for normally distributed data measured on a continuous/interval scale (e.g., body mass index). The mathematical formula for the Student's t-test includes the mean between-group difference in the numerator and between-group variability in the denominator.
The following core competencies are addressed in this article: Medical knowledge.

Keywords: Between-group variability, independent groups, Student's t-test


How to cite this article:
Stoltzfus JC. Student's t-test for independent samples. Int J Acad Med 2015;1:27-8

How to cite this URL:
Stoltzfus JC. Student's t-test for independent samples. Int J Acad Med [serial online] 2015 [cited 2019 Jan 17];1:27-8. Available from: http://www.ijam-web.org/text.asp?2015/1/1/27/172708

When analyzing data, it is common practice to compare two independent groups, such as males versus females or smokers versus nonsmokers. In this case, “independent groups” means that the observations from one sample are completely unrelated to the observations from another sample. If the outcome is measured on a continuous/interval scale (i.e., it exists on a continuum that can be added, subtracted, multiplied, or divided) and is normally distributed (i.e., it looks like a “bell-shaped curve”), the Student's t-test is the appropriate statistical method.[1],[2]

The Student's t-test (also known as an unpaired t-test) was developed in 1908 by a chemist named William Sealy Gosset (who went by the pseudonym “Student” when publishing his work) to test the quality of stout at the Guinness Brewery in Dublin, Ireland. Below is the mathematical formula for the Student's t-test:[1],[2]



In this equation, the numerator (X-bar and Y-bar) represents the difference in means between the two independent groups. The denominator represents the between-group variability, with n1 and n2 being the sample sizes of the two groups, and Sp being the pooled standard deviation or the weighted average of the standard deviation of the two groups. The formula for Sp[1],[2] is:



In this equation, n1 and n2 are the sample sizes of the two groups, and S12 and S22 are the standard deviations of the two groups.

Besides requiring continuous/interval-level outcomes that are normally distributed, the Student's t-test was originally designed for data with equal variances between the two groups being compared.[1],[2] This means that the distribution of data in one group should be similar to the other group for example, when comparing body mass indexes (BMIs) in males versus females, the distribution of BMI values should be similar at the low, middle, and high ranges for both males and females. If two groups have unequal variances, which sometimes happen when the two groups' sample sizes are noticeably different or highly imbalanced, a slightly different type of Student's t-test may be used.

Although statistical software packages can quickly compute a Student's t-test, it is easy enough to calculate by hand. Consider the following example: In a group of 50 male athletes, mean BMI is 28.5, and the standard deviation is 10.5. In a group of 50 female athletes, mean BMI is 23.5, and the standard deviation is 12.5. If we apply the Student's t-test formula:

Step 1: Calculate

X-bar (males) – Y-bar (females) = 28.5–23.5 = 5.0

Step 2: Calculate



Step 3: Complete the formula:



The t-test value for our BMI data is 2.16. We then compare this value to the critical t-test threshold that matches our data's parameters, including the level of error we select for finding a significant difference that does not truly exist (usually 0.05, or 5%). (Note that critical t-test threshold tables can be found online).[3] If our value of 2.16 is greater than the critical threshold, we conclude that male athletes' BMIs are significantly greater than female athletes' BMIs. If our t-test value is less than the critical t-test threshold value, we must conclude that BMI does not differ significantly by gender.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

 
  References Top

1.
Socialresearchmethods.net. The t-Test. Available from: http://www.socialresearchmethods.net/kb/stat_t.php. [Last cited on 2015 Sep 15].  Back to cited text no. 1
    
2.
Seltman HJ. The t-Test and Basic Inference Principles. Experimental Design and Analysis; 2015. p. 141-70. Available from: http://www.stat.cmu.edu/~hseltman/309/Book/chapter6.pdf [Last cited on 2015 Sep 15].  Back to cited text no. 2
    
3.
Itl.nist.gov. Critical Values of the Student's t Distribution. Available from: http://www.itl.nist.gov/div898/handbook/eda/section3/eda3672.htm.[Last cited on 2015 Sep 15].  Back to cited text no. 3
    




 

Top
 
 
  Search
 
Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

 
  In this article
Abstract
References

 Article Access Statistics
    Viewed1586    
    Printed54    
    Emailed0    
    PDF Downloaded86    
    Comments [Add]    

Recommend this journal


[TAG2]
[TAG3]
[TAG4]