“The 95% confidence interval for \(\theta\) is \(29.5 \leq \theta \leq 51.3\).”
“The interval from 29.5 to 51.3 was calculated using a procedure that, in 95% of repeated samples, would produce an interval containing the true \(\theta\).”
But writing that long sentence every time is impractical. So statisticians use the shorter notation as a convenient shorthand, even though it’s technically imprecise from a strict frequentist perspective.
When you write \(29.5 \leq \theta \leq 51.3\):
| Group | Notation They Use | What They Believe |
|---|---|---|
| Strict frequentists | Write \(29.5 \leq \theta \leq 51.3\) as shorthand | But would say: “We cannot assign probability to this statement” if pressed |
| Applied researchers | Write \(29.5 \leq \theta \leq 51.3\) | Often misinterpret it as “There’s a 95% probability \(\theta\) is in here” (Bayesian thinking) |
| Bayesians | Write \(29.5 \leq \theta \leq 51.3\) | Mean it literally: \(P(29.5 \leq \theta \leq 51.3 \mid \text{data}) = 0.95\) |
Some textbooks try to be more precise with the confidence interval notation:
Some authors write:
“The 95% confidence interval is 29.5 to 51.3”
Without the inequality symbols, to avoid implying a probability statement about \(\theta\).
You’re not wrong to see people writing \(29.5 \leq \theta \leq 51.3\) — it’s standard practice. But:
This is one of the most common and persistent misunderstandings in all of statistics. The fact that you noticed this discrepancy means you’re thinking more carefully about statistical philosophy than most practitioners!
If you see \(29.5 \leq \theta \leq
51.3\) in a frequentist paper, mentally
translate to:
“The interval (29.5, 51.3) was produced by a 95% confidence
procedure”
If you see it in a Bayesian paper, you can
literally interpret it as:
“There’s a 95% probability \(\theta\) is between 29.5 and 51.3, given
the data and prior”
The notation looks identical, but the philosophical interpretation is completely different!