Normal Distribution vs T-Distribution

A normal distribution is a continuous model written $N(\mu, \sigma)$: a density curve, symmetric about its mean $\mu$ (mu), single-peaked, and spread out by its standard deviation $\sigma$ (sigma). Two numbers fix the whole curve. Probability is area underneath it, the total area is exactly 1, and half of that area sits on each side of $\mu$. Because the area over a single point is zero, every normal question is really a question about an interval.

Suppose adult male heights are approximately $N(69, 2.8)$ in inches. To find the share above 72 inches, standardize: $z = \frac{72 - 69}{2.8} \approx 1.07$. The area to the left of $z = 1.07$ is 0.8577, so about 0.142, roughly 14 percent, are taller than 72 inches. The same curve puts about 71.6 percent of men between 66 and 72 inches.

The claim that ruins the most work is "the sample is large, so the data are normal." Sample size does not change the shape of the variable being measured. A right-skewed variable such as household income stays right-skewed however many households you collect. What a large sample buys is that the sampling distribution of $\bar{x}$ (x-bar, the sample mean) is close to normal, and that is a statement about the average of a sample, not about the individual values inside it.

A normal curve also runs on forever in both directions, so a normal model always assigns some probability to values the real variable cannot reach, negative heights included. For heights that leftover is far too small to matter. Where the model genuinely fails is shape: strongly skewed, hard-bounded, or clearly bimodal data should not be pushed through a normal calculation, and no amount of extra data repairs that.

The normal distribution is topic 2.11 in Unit 2, Probability, Random Variables, and Probability Distributions. The empirical rule gives the quick version of its areas and the z-table gives the exact ones.

There is a calculus reason the areas have to come from a table or a calculator rather than from an antiderivative. The bell-shaped function $e^{-x^{2}}$ has no elementary antiderivative at all, so no amount of algebra produces a formula for the area under it between two bounds. CalcLearn works through that function's behaviour in the limit of e to the minus x squared.

The $t$-distribution is not one curve but a family, indexed by the degrees of freedom. It is the distribution of $t = \frac{\bar{x} - \mu}{s / \sqrt{n}}$ (x-bar minus mu, over s divided by root n) when the data come from a normal population. Swapping the fixed $\sigma$ (sigma) for the sample standard deviation $s$, which itself changes from sample to sample, is what puts the extra weight in the tails. Every member is symmetric about 0, and the family closes on the standard normal as the degrees of freedom grow.

The numbers make that convergence concrete. For a 95 percent interval the t-table gives $t^* = 2.145$ at 14 degrees of freedom, $2.042$ at 30, $1.984$ at 100 and $1.962$ at 1000, against $z^* = 1.960$ for the normal. The gap is 8.6 percent of the critical value at 14 degrees of freedom and about 0.1 percent at 1000.

"The sample is small so use $t$, and large so use $z$" is the wrong rule, and it is the one most students carry in. The trigger is whether $\sigma$ is known, not how big $n$ is. With 500 observations and a standard deviation estimated from them, the correct model is $t$ on 499 degrees of freedom, which happens to sit very close to the normal. Knowing $\sigma$ with $n = 8$ would put you back on $z$.

The heavier tails change verdicts, not just widths. A statistic of 2.00 read on $t$ with 14 degrees of freedom has a two-sided p-value of 0.0653, against the 0.0455 the normal returns for the same 2.00, so at $\alpha = 0.05$ one model rejects and the other does not.

Those tails cover the uncertainty in $s$ and nothing else. They do not repair a skewed population or a stray outlier, which is why a $t$ procedure still asks you to look at the shape of the sample first. The $t$-distribution enters the course at topic 4.2 of Unit 4, Inference for Quantitative Data: Means.