Maths

Common mistakes in the normal distribution

6 mistakes learners make with the normal distribution, each one named and explained.

All common mistakes

Standardise first, then read the tail the question actually asked for.

Divided by n rather than the root of n

What it looks like

The standard error is given as sigma over n.

Why it happens

Dividing by the sample size is the shape of the formula and the root is the small detail.

A worked example

Question. Standard error

A common answer. sigma / n

The answer. sigma / root(n)

Why. The variance divides by n, and the standard deviation is its square root.

Put the two side by side

Variance of the mean: sigma squared over n.
Standard deviation of the mean: sigma over root n.

Take the root of the variance formula. The n comes out as root n.

The population standard deviation used unchanged

What it looks like

Sigma is given as the standard deviation of the sample mean.

Why it happens

Sigma is the standard deviation in the question, so it is the one to hand.

A worked example

Question. Standard error of the sample mean

A common answer. sigma

The answer. sigma over the root of n

Why. Averaging several observations makes the average less variable than any single one.

Put the two side by side

Sigma: how much one observation varies.
Sigma over root n: how much the mean of n varies.

A mean of many is always steadier than one value. The standard error has to be smaller.

The value never standardised

What it looks like

A probability is read off without converting the value to a z-score first.

Why it happens

The numbers in the question look usable and standardising is an extra step.

A worked example

Question. Mean 50, standard deviation 8, P(X < 58)

A common answer. 0.500

The answer. 0.841

Why. 58 is one standard deviation above the mean, which is z = 1.

Put the two side by side

Without standardising: no useful number.
z = (58 - 50) / 8 = 1, which gives 0.841.

Every normal question starts by turning the value into a z. Nothing works before that.

The variance given where a standard deviation was asked for

What it looks like

Sigma squared over n is offered as the standard error.

Why it happens

The variance formula is the one derived first and it is the more memorable expression.

A worked example

Question. Standard deviation of the sample mean

A common answer. sigma^2 / n

The answer. sigma / root(n)

Why. Sigma squared over n is the variance; the standard deviation is its root.

Put the two side by side

Variance: squared units.
Standard deviation: the original units.

Check the units. A standard deviation always matches the data it describes.

The wrong tail of the distribution read

What it looks like

P(X < 58) is answered with the upper tail probability, or a value is found on the wrong side.

Why it happens

Tables and calculators give one tail by default, and which one is easy to lose track of.

A worked example

Question. Mean 50, standard deviation 8, find P(X < 58)

A common answer. 0.159

The answer. 0.841

Why. 58 is above the mean, so more than half the distribution lies below it.

Put the two side by side

Above the mean: the probability below it exceeds 0.5.
Below the mean: it is less than 0.5.

Sketch the curve and shade the region. The shaded fraction tells you if the answer is sensible.

The z-value added without multiplying by the standard deviation

What it looks like

k is found as the mean plus the z-value, with the standard deviation never applied.

Why it happens

The z-value looks like a number of units, because in standardised terms it is.

A worked example

Question. Mean 70, standard deviation 5, z = 1.2816

A common answer. 70 + 1.2816 = 71.3

The answer. 70 + 1.2816 x 5 = 76.4

Why. A z-value counts standard deviations, so it has to be multiplied by one.

Put the two side by side

z is measured in standard deviations.
x is measured in the original units.

Rearranging z = (x - mean) / sd always leaves the standard deviation as a multiplier.

WAJD spots these patterns in your child's answers and names the one behind their wrong answers, instead of just marking them wrong.

More maths topics