19-CIs-Intro.Rmd

# (PART) Analysis: confidence intervals {-}



# Introducing confidence intervals {#CIIntro}

Chap. \@ref(SamplingVariation) introduced the idea of sampling variation.
In Chaps. \@ref(CIOneProportion) to \@ref(OddsRatiosCI), this idea is used to form *confidence intervals*.

Confidence intervals help answer [estimation-type RQ](#TypesOfRQs), where the precision of a *statistic* is of interest.
In this Part, answering estimation-type RQs is discussed for:

* Descriptive RQs: 
  * One proportion: Chap. \@ref(CIOneProportion), where the response variable is qualitative.
  * One mean: Chap. \@ref(OneMeanConfInterval), where the response variable is quantitative.
  * Mean difference: Chap.  \@ref(PairedCI), for *paired* quantitative data.
* Relational or interventional RQs with a *comparison*:
  * Comparing means in two groups: Chap. \@ref(CITwoMeans).
  * Comparing odds in two groups: Chap. \@ref(OddsRatiosCI).

Answering estimation-type RQs for relational or interventional RQs with a *connection* is explored later (Chaps. \@ref(Correlation) and \@ref(Regression)).

The precision of statistics influences [decision-type RQ](#TypesOfRQs) too: when [statistics](#StatisticsAndParameters) precisely estimate [parameters](#StatisticsAndParameters), making decisions about parameters is easier.

The previous chapters, where tools for describing sampling variation were introduced, are used in this part to understand the precision of statistics.


```{r fig.cap="", fig.align="center", fig.width=3, out.width="35%"}
SixSteps(5, "Comparison RQs: CIs")
```