What is the fundamental difference between the standard deviation used in a z-score versus a t-score?

A z-score uses the known population standard deviation ($\sigma$), whereas a t-score uses the sample standard deviation ($s$) as an estimate. This estimation introduces extra variability, which the t-distribution accounts for with its thicker tails.

When should you use a z-score even if the population is not normally distributed?

You can use a z-score for a non-normal population if the sample size is large ($n \ge 30$). This is justified by the Central Limit Theorem, which states that the sampling distribution of the mean becomes approximately normal as sample size increases.

How does the t-distribution change as the sample size ($n$) increases?

As $n$ increases, the degrees of freedom ($n-1$) also increase, causing the t-distribution to narrow. Eventually, the t-distribution converges to and becomes identical to the standard normal (z) distribution.

What is a common error when calculating a test statistic for a small sample ($n < 30$) with an unknown population variance?

A common error is using a z-score instead of a t-score. Because the sample standard deviation is only an estimate, using the z-distribution would underestimate the probability of extreme values and lead to an overconfident (non-conservative) result.

Why is it incorrect to use a t-score for a very small, highly skewed sample?

The t-test assumes that the underlying population is approximately normal or that the sample size is large enough for the CLT to apply. If the sample is small and the population is skewed, the sampling distribution will not be normal, making the t-score results invalid.

What happens to the margin of error if you mistakenly use a z-critical value instead of a t-critical value for a small sample?

The margin of error will be too small. Since z-critical values are smaller than t-critical values for the same confidence level, the resulting interval will be too narrow and will fail to capture the population mean at the intended confidence level.

Define 'Degrees of Freedom' in the context of a t-score.

Degrees of freedom ($df$) refer to the number of independent pieces of information available in a data set, calculated as $n - 1$ for a single sample. It determines which specific t-distribution curve is used to find critical values or p-values.

What is the formula for the t-statistic for a sample mean?

The formula is $t = \frac{\bar{x} - \mu}{s / \sqrt{n}}$. Here, $\bar{x}$ is the sample mean, $\mu$ is the population mean, $s$ is the sample standard deviation, and $n$ is the sample size.

What does it mean for the t-distribution to be 'conservative'?

Being 'conservative' means the distribution has more area in the tails, requiring a larger test statistic to reach significance. This higher threshold protects against Type I errors that might occur due to the uncertainty of estimating the population standard deviation.

Under what specific condition is no statistical procedure available for inference about a mean?

No standard procedure is available when the population is not normally distributed and the sample size is small ($n < 30$). In this scenario, the Central Limit Theorem cannot be invoked, and the t-distribution's requirement for normality is not met.

Library Podcasts

Courses

Referral & Rewards

t-scores versus z-scores

Summary

The choice between using a z-score and a t-score is a fundamental decision in inferential statistics, primarily determined by the availability of the population standard deviation and the sample size. While z-scores rely on the standard normal distribution, t-scores utilize a family of distributions that account for the additional uncertainty introduced when estimating population parameters from sample data.

1. Definition & Core Concepts

A z-score is a standardized value that indicates how many standard deviations a data point is from the mean, based on the standard normal distribution with a mean of 0 and a standard deviation of 1.

A t-score is used when the population standard deviation is unknown and must be estimated using the sample standard deviation ( $s$ ), resulting in a distribution that is more spread out than the normal curve.

The t-distribution is characterized by 'thicker tails' compared to the normal distribution, which provides a more conservative estimate by accounting for the increased variability inherent in smaller samples.

Comparison of the Normal (z) distribution and the t-distribution, showing the t-distribution's lower peak and thicker tails.

2. Underlying Principles

The Central Limit Theorem (CLT) allows for the use of z-scores even in non-normal populations, provided the sample size is sufficiently large ( $n \ge 30$ ), as the sampling distribution of the mean will approximate normality.

When the population variance ( $\sigma^2$ ) is known, the standard error is precisely defined, justifying the use of the z-distribution regardless of sample size, assuming the underlying population is normal.

The t-distribution relies on degrees of freedom ( $df = n - 1$ ), where the shape of the curve changes as the sample size increases, eventually converging with the standard normal distribution as $n$ approaches infinity.

3. Methods & Techniques

4. Key Distinctions

5. Exam Strategy & Tips

t-scores versus z-scores

Q: What is a common error when calculating a test statistic for a small sample ($n < 30$) with an unknown population variance?

A common error is using a z-score instead of a t-score. Because the sample standard deviation is only an estimate, using the z-distribution would underestimate the probability of extreme values and lead to an overconfident (non-conservative) result.

Q: Why is it incorrect to use a t-score for a very small, highly skewed sample?

The t-test assumes that the underlying population is approximately normal or that the sample size is large enough for the CLT to apply. If the sample is small and the population is skewed, the sampling distribution will not be normal, making the t-score results invalid.

Q: What happens to the margin of error if you mistakenly use a z-critical value instead of a t-critical value for a small sample?

The margin of error will be too small. Since z-critical values are smaller than t-critical values for the same confidence level, the resulting interval will be too narrow and will fail to capture the population mean at the intended confidence level.

Q: Define 'Degrees of Freedom' in the context of a t-score.

Degrees of freedom ($df$) refer to the number of independent pieces of information available in a data set, calculated as $n - 1$ for a single sample. It determines which specific t-distribution curve is used to find critical values or p-values.

Q: What is the formula for the t-statistic for a sample mean?

The formula is $t = \frac{\bar{x} - \mu}{s / \sqrt{n}}$. Here, $\bar{x}$ is the sample mean, $\mu$ is the population mean, $s$ is the sample standard deviation, and $n$ is the sample size.

Q: What does it mean for the t-distribution to be 'conservative'?

Being 'conservative' means the distribution has more area in the tails, requiring a larger test statistic to reach significance. This higher threshold protects against Type I errors that might occur due to the uncertainty of estimating the population standard deviation.

Q: Under what specific condition is no statistical procedure available for inference about a mean?

No standard procedure is available when the population is not normally distributed and the sample size is small ($n < 30$). In this scenario, the Central Limit Theorem cannot be invoked, and the t-distribution's requirement for normality is not met.

Summary

1. Definition & Core Concepts

Comparison of the Normal (z) distribution and the t-distribution, showing the t-distribution's lower peak and thicker tails.

2. Underlying Principles

3. Methods & Techniques

To select the correct score, first determine if the population standard deviation ( $\sigma$ ) is known; if it is, and the population is normal or $n$ is large, a z-score is the standard choice.

If $\sigma$ is unknown, calculate the sample standard deviation ( $s$ ) and use a t-score, ensuring that the population is approximately normal or the sample size is large enough to satisfy the CLT.

In cases where the population is not normally distributed and the sample size is small ( $n < 30$ ), standard parametric procedures like z or t tests cannot be reliably applied without further data transformation.

4. Key Distinctions

The primary distinction lies in the source of the standard deviation: z-scores use the population parameter ( $\sigma$ ), while t-scores use the sample statistic ( $s$ ).

Feature	z-score	t-score
SD Source	Population ( $\sigma$ )	Sample ( $s$ )
Distribution	Standard Normal	t-distribution (by $df$ )
Sample Size	Preferred for $n \ge 30$	Preferred for $n < 30$
Tail Density	Lower (less extreme)	Higher (more conservative)

As sample size increases, the difference between z and t critical values diminishes, making the t-distribution a robust 'default' when the population variance is unknown.

5. Exam Strategy & Tips

Check the Sigma: Always look for whether the problem provides the population standard deviation ( $\sigma$ ) or the sample standard deviation ( $s$ ). This is the most common trigger for choosing between $z$ and $t$ .
Sample Size Threshold: Remember the $n = 30$ rule for the Central Limit Theorem. If $n < 30$ , you MUST verify the population is described as 'normally distributed' or 'approximately symmetric' before proceeding with a t-test.
Conservative Approach: If you are unsure and $\sigma$ is not explicitly given, the t-score is generally the safer, more conservative choice because it accounts for the estimation error of the standard deviation.
Infinity Row: On many t-distribution tables, the very last row (labeled $\infty$ ) provides the critical values for the z-distribution, which can be used as a quick reference.

To select the correct score, first determine if the population standard deviation ( $\sigma$ ) is known; if it is, and the population is normal or $n$ is large, a z-score is the standard choice.

If $\sigma$ is unknown, calculate the sample standard deviation ( $s$ ) and use a t-score, ensuring that the population is approximately normal or the sample size is large enough to satisfy the CLT.

The primary distinction lies in the source of the standard deviation: z-scores use the population parameter ( $\sigma$ ), while t-scores use the sample statistic ( $s$ ).

Feature	z-score	t-score
SD Source	Population ( $\sigma$ )	Sample ( $s$ )
Distribution	Standard Normal	t-distribution (by $df$ )
Sample Size	Preferred for $n \ge 30$	Preferred for $n < 30$
Tail Density	Lower (less extreme)	Higher (more conservative)

As sample size increases, the difference between z and t critical values diminishes, making the t-distribution a robust 'default' when the population variance is unknown.

Check the Sigma: Always look for whether the problem provides the population standard deviation ( $\sigma$ ) or the sample standard deviation ( $s$ ). This is the most common trigger for choosing between $z$ and $t$ .
Sample Size Threshold: Remember the $n = 30$ rule for the Central Limit Theorem. If $n < 30$ , you MUST verify the population is described as 'normally distributed' or 'approximately symmetric' before proceeding with a t-test.
Conservative Approach: If you are unsure and $\sigma$ is not explicitly given, the t-score is generally the safer, more conservative choice because it accounts for the estimation error of the standard deviation.
Infinity Row: On many t-distribution tables, the very last row (labeled $\infty$ ) provides the critical values for the z-distribution, which can be used as a quick reference.