1.
Hypothesis Test for the Mean
Hypothesis testing is a method in statistics that helps us decide if what we
see in our sample data is just by chance or if it really reflects something
true about the population. When we test the population mean, we are
asking: “Is the average value in my sample close enough to the
population mean I expect, or is it different?” For example, if a company
claims the average life of their battery is 10 hours, we can collect a
sample of batteries, test them, and then check if the sample mean
supports that claim or not.
2. Null and Alternative Hypotheses
Every test starts with two statements. The null hypothesis (H_0) is like
the default assumption — it says nothing has changed, or the mean is
equal to a certain value. The alternative hypothesis (H_1) is the
opposite — it says the mean is different, or there is some effect. For
example, if we want to test whether students’ average marks are 40, we
set H_0:\mu =40. If we think the average is not 40, then H_1:\mu \neq 40.
These two hypotheses guide the whole test.
3. The p-Value
The p-value is a very important number in hypothesis testing. It tells us
the probability of getting results as extreme as our sample, assuming the
null hypothesis is true. If the p-value is very small, it means our sample
result is unlikely under the null, so we may reject it. If the p-value is large,
it means the sample result is consistent with the null, so we keep it. In
simple words: small p-value → strong evidence against the null; large p-
value → weak evidence against the null.
4. When σ is Known
Sometimes, we already know the population standard deviation (\sigma ).
In that case, we use the z-statistic. The formula is:
z=\frac{\bar {x}-\mu _0}{\sigma /\sqrt{n}}
Here, \bar {x} is the sample mean, \mu _0 is the hypothesized mean, and
n is the sample size. Once we calculate z, we look at the normal
distribution to find the p-value. This method works well when \sigma is
known and the sample size is large.
5. Sample Variance, Standard Deviation, and Standard Error
In real life, we usually don’t know the population standard deviation. So
we estimate it using the sample. The sample variance measures how
spread out the data is. The sample standard deviation is just the
square root of the variance, and it tells us the average distance of data
points from the mean. The standard error is the sample standard
deviation divided by the square root of the sample size. It shows how
much the sample mean is expected to vary from sample to sample.
Smaller standard error means our sample mean is more reliable.
6. When σ is Unknown
If we don’t know \sigma , we use the t-distribution instead of the normal
distribution. The formula looks the same, but we replace \sigma with the
sample standard deviation s. The t-distribution is similar to the normal
curve but has fatter tails, which means it accounts for more uncertainty.
This is especially important when the sample size is small, because small
samples can vary more.
7. The t-Statistic
The t-statistic is calculated as:
t=\frac{\bar {x}-\mu _0}{s/\sqrt{n}}
We then compare this value with critical values from the t-table, which
depend on the degrees of freedom (usually n-1). If the t-statistic is larger
than the critical value, we reject the null hypothesis. The t-test is very
common in practice because we rarely know the true population standard
deviation.
8. Significance Level
Before we start the test, we decide on a significance level, usually
written as \alpha . A common choice is 0.05, which means we are willing
to accept a 5% chance of wrongly rejecting the null hypothesis. If the p-
value is less than \alpha , we reject the null. If it is greater, we keep the
null. This rule makes our decision clear and consistent. It also controls the
risk of making a Type I error, which is rejecting the null when it is
actually true.
9. One-Sided Alternatives
Sometimes, we are only interested in testing in one direction. For
example, a company may want to prove that their average battery life is
greater than 10 hours, not just different. In that case, we set up a one-
sided test: H_0:\mu \leq 10 versus H_1:\mu >10. One-sided tests focus
only on one tail of the distribution, which makes them more powerful for
detecting differences in that specific direction.