Writing parameters in AP Statistics

Almost every inference problem in AP Statistics starts by defining a parameter, and it is one of the most common places to lose easy points. The good news is that the fix is a habit, not a new idea. Once you are clear on what a parameter is and how to write one, hypotheses and confidence intervals stop feeling like guesswork.

A parameter is not a statistic

A parameter is a fixed number that describes a whole population. A statistic is a number you calculate from a sample. Inference is always about the parameter: you use the statistic you can see to draw a conclusion about the parameter you cannot.

The symbols matter, because using the wrong one changes the meaning:

• A population proportion is \(p\); the sample proportion is \(\hat{p}\).
• A population mean is \(\mu\); the sample mean is \(\bar{x}\).

So hypotheses and parameter definitions are written with \(p\) and \(\mu\), never with \(\hat{p}\) or \(\bar{x}\). A hypothesis like \(H_0: \hat{p} = 0.5\) is automatically wrong, because \(\hat{p}\) is something you measured, not something you are making a claim about.

A template that always works

Define every parameter the same way:

Let [symbol] = the true [mean or proportion] of [the variable] for [the population].

Naming the population and the variable in words is what earns the point. A bare symbol is not enough.

One proportion

A school claims 60% of students bike to school. You suspect it is lower and survey a random sample.

\[p = \text{the true proportion of all students at the school who bike to school}\] \[H_0: p = 0.60 \qquad H_a: p < 0.60\]

One mean

A coffee shop advertises a mean wait time of 4 minutes. You time a random sample of visits.

\[\mu = \text{the true mean wait time, in minutes, for all visits to the shop}\] \[H_0: \mu = 4 \qquad H_a: \mu \neq 4\]

Two proportions

You compare the proportion of first-year and second-year students who use the tutoring center, sampling each group separately. Use a difference of proportions, and define which group is which:

\[p_1 - p_2, \text{ where } p_1 = \text{true proportion of first-years who use the center, and } p_2 = \text{the same for second-years}\] \[H_0: p_1 - p_2 = 0 \qquad H_a: p_1 - p_2 \neq 0\]

Writing \(H_0: p_1 - p_2 = 0\) says the two proportions are equal, which is usually the claim of “no difference.”

Two means

You compare the mean study time of athletes and non-athletes from two independent samples:

\[\mu_1 - \mu_2, \text{ where } \mu_1 = \text{true mean study time for athletes and } \mu_2 = \text{true mean study time for non-athletes}\] \[H_0: \mu_1 - \mu_2 = 0 \qquad H_a: \mu_1 - \mu_2 \neq 0\]

Paired data: the one everyone mixes up

If the data are paired (each subject measured twice, or matched pairs), you do not have two independent groups. You work with the population of differences, and there is just one parameter:

\[\mu_d = \text{the true mean difference (after minus before) for the population}\] \[H_0: \mu_d = 0 \qquad H_a: \mu_d > 0\]

For example, if you record each student’s score before and after a study workshop, every student gives you one difference, so you use \(\mu_d\), not \(\mu_1 - \mu_2\). The tell is the design: two measurements on the same subjects (or matched pairs) means paired; two separate groups means a difference of means.

For a confidence interval, define the parameter too

The same definitions apply. A confidence interval estimates a parameter, so you still write, for instance, “\(\mu\) = the true mean wait time for all visits,” and then interpret the interval as capturing that \(\mu\), never as capturing \(\bar{x}\) or “the sample.”

The mistakes to avoid

• Writing hypotheses about \(\hat{p}\) or \(\bar{x}\) instead of \(p\) or \(\mu\).
• Giving a symbol with no context: define the population and the variable in words.
• Using \(\mu\) for a proportion, or \(p\) for a mean.
• Treating paired data as two independent groups (using \(\mu_1 - \mu_2\) when you should use \(\mu_d\)).