Test on Single
Sample
Hypothesis Testing with Unknown Variance
What is Hypothesis
Testing?
Hypothesis Testing is a statistical method used to make decisions or
draw conclusions about a population based on sample data. Rather
than measuring every individual in a population (which is often impossible
or impractical), we collect a sample and use it to test our beliefs about the
population.
Null Hypothesis
(H₀)
"No difference" or no effect. This represents the status quo or
baseline claim we're testing against.
Example: μ = 46 kWh/year
Alternative Hypothesis
(H₁)
"There is a difference" or an effect exists. This is what we're trying to
find evidence for.
Example: μ ≠ 46 or μ < 46
When to Use the One-
Sample t-Test?
The one-sample t-test is specifically designed for situations where you're comparing a sample mean to a known population value. It's the tool of choice when
population variance is unknown—which is almost always the case in real-world applications.
Use t-Test When: Don't Use for:
• Analyzing one sample only • Comparing two independent groups (use two-sample t-test instead)
• Comparing to a known value (μ₀) • Known population variance with large samples (use z-test)
• Population variance (σ²) is UNKNOWN → use sample standard • Paired samples without accounting for the pairing
deviation s
• Sample is small (n < 30) or data is approximately normal
The t-Statistic
Formula
Numerator Denominator Degrees of
Freedom
The difference between your sample mean and The standard error, which accounts for both the df = n − 1. This parameter determines which t-
the claimed population value—this measures how sample variability (s) and sample size (n). Larger distribution to use when looking up your critical
far your data deviates from the null hypothesis. samples reduce this error. value in the t-table.
Setting Up Your
Hypotheses
The way you structure your null and alternative hypotheses depends on your research question. There are three common
configurations, each leading to different decision rules and critical regions.
Test Type Null Hypothesis (H₀) Alternative Hypothesis (H₁)
Two-tailed μ = μ₀ μ ≠ μ₀
Left-tailed μ ≥ μ₀ μ < μ₀
Right-tailed μ ≤ μ₀ μ > μ₀
Decision Rule (using α
= 0.05)
Two-tailed Test One-tailed Test
Reject H₀ if |t| > tα/2,df. The rejection region is split Left-tailed: Reject if t < −tα,df. Right-tailed: Reject if t >
equally between both tails of the distribution. tα,df. All rejection area concentrates in one tail.
Worked Example:
Applying the t-Test
This example demonstrates the complete process of
conducting a one-sample t-test: calculating the test
statistic, determining degrees of freedom, finding the
critical value from the t-table, and making a statistical
decision. Following these steps systematically ensures
accurate hypothesis testing and valid conclusions.
Why Use t-Test Instead of z-
Test?
This is a critical distinction in statistics. The choice between z-test and t-test depends on what information you have about the
population and your sample size. Understanding these differences ensures you apply the correct test to your data.
Feature z-Test t-Test
Population Variance Known (σ) Unknown (use s)
Sample Size Large (n ≥ 30) Small (n < 30)
Distribution Shape Normal t-distribution (fatter tails)
Relationship Standard normal As n ↑ → t ≈ z
Key Insight: The t-distribution has fatter tails than the normal distribution, especially with small sample sizes. This
accounts for the extra uncertainty when estimating the population standard deviation from sample data. As your sample
grows, the t-distribution converges to the normal distribution.