What is T distribution?

The T distribution sometimes referred to as the Student t-distribution, is a probability distribution that is applied in smaller samples with an unknown population variance to estimate population parameters.

The T distribution is a kind of probability distribution with a bell-shaped pattern resembling the normal distribution but with thicker tails. This distribution will be utilised in place of the normal distribution when the sample size is small. T distributions have broader tails than normal distributions because they are more likely to contain extreme values.

 

 

What does it mean?

Typically, the sampling distribution of a statistic tends to follow a normal distribution for large sample sizes. As a result, calculating a z-score (the statistic linked to normal distribution) is simple as long as the population’s standard deviation is known. This helps statisticians to assess probabilities using the sample mean through the normal distribution. But, when the sample size is small, the population’s standard deviation is frequently unknown. In this case, statisticians utilise the distribution of the t statistic. The t distribution, in essence, enables statisticians to do statistical studies on smaller data sets that would otherwise be impossible to analyse using the normal distribution.

Smaller values of the degrees of freedom parameter of the T distribution result in larger tails, whereas higher values cause the T distribution to resemble a typical standard distribution with “0” as the mean and “1” as the standard deviation. The mean sample, m, and standard sample deviation, d, will differ from M and D due to the sample’s randomness when a sample of n observations is selected from a population with a normally distributed mean, M, and standard deviation, D.

Formula to Calculate Student’s T Distribution

Here’s the T distribution formula 

t = (x̄ – μ) / (s/√n)

Where,

x̄ is the sample mean

μ is the population mean

s is the standard deviation

n is the sample size 

So, we see that the required data are necessary to calculate the T distribution. 

One requires the population mean, which is the population’s average. One needs the sample mean to determine whether the population mean is authentic and whether the sample picked will represent the same statement. To normalise the value, the t distribution formula subtracts the sample mean from the population mean, divides the result by the standard deviation, and multiplies the result by the square root of the sample size.

The value can go awry, and we won’t be able to calculate probability because the student’s t distribution has restrictions on how it can arrive at a number and is therefore only relevant for smaller sample sizes. Additionally, one needs to get that value from the student’s t distribution table to calculate probability after arriving at a score.

Example of T distribution

Let us consider an example of T distribution where the different variables are provided for consideration. 

Here, the population mean is 310, 50 is the standard deviation, 16 is the sample size, and 290 is the sample mean 

We can easily calculate the t-distribution from these values. 

Value of t = (Sample Mean – Population Mean)/( standard deviation /√Sample size)

= (290 – 310) / (50 / √16)

= -1.60

T-distribution’s significance in finance

In finance, when we believe an asset return can be viewed as a random variable, we utilise probability distributions to create images that depict our opinion of an asset return’s sensitivity.

Because it has a little “fatter tail” than the normal distribution, the student’s T distribution is likewise very well-liked. When our sample size is limited, we often use the student’s T. (i.e. less than 30). The left tail in finance stands for the losses. Therefore, we dare underestimate the likelihood of a significant loss if the sample size is limited. We can use the student’s T’s bigger tail in this situation. However, it so happens that the fat tail of this distribution is frequently insufficiently fat.  

What is the t-distribution table?

The proportions associated with z-scores are determined using the t-distribution table. This table is used to calculate the ratio for t-statistics. The critical values of the t distribution are displayed in the t-distribution table. The likelihood of t deriving values from a particular value is displayed in the t-distribution table. The area of the t-curve between the ordinates of the t-distribution, the given value, and infinity is the obtained probability.

Only three values are required to use the t-distribution table:

  • The degrees of freedom of the t-test
  • The number of tails of the t-test, whether one-tailed or two-tailed)
  • The alpha level of the t-test (common choices are 0.01, 0.05, and 0.10)

T distribution vs normal distribution

When a normal population distribution is assumed, normal distributions are employed. The normal distribution and the T distribution are similar, but the T distribution has fatter tails. However, both presuppose a population with a normal distribution. The kurtosis of T distributions is larger than that of normal distributions. Kurtosis is a way to figure out how skewed distribution is. With a T distribution as opposed to a normal distribution, there is a higher chance of finding values quite distant from the mean.

When testing the value of the population mean for any given population, both of the distributions are used. The Z test is used when the standard deviation of the population is known, while the T distribution is used when the parameter of the population is unknown.

Also, the T distribution is usually used when the number of people in the sample is small (less than 30), while the Z distribution is used when the number of people in the sample is large (more than 30).

Conclusion

T distribution is a probability distribution used in various statistical tests for two reasons. One is to test whether observed results are different from what would be expected under the null hypothesis, and the other is to test whether two sets of measurements are drawn from the same parent population, often referred to as paired testing.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *