**[MV] alpha** -- Compute interitem correlations (covariances) and Cronbach's
alpha

__Syntax__

**alpha** *varlist* [*if*] [*in*] [**,** *options*]

*options* Description
-------------------------------------------------------------------------
Options
__a__**sis** take sign of each item as is
__c__**asewise** delete cases with missing values
__d__**etail** list individual interitem correlations and
covariances
__g__**enerate(***newvar***)** save the generated scale in *newvar*
__i__**tem** display item-test and item-rest correlations
__l__**abel** include variable labels in output table
__m__**in(***#***)** must have at least *#* observations for inclusion
__r__**everse(***varlist***)** reverse signs of these variables
__s__**td** standardize items in the scale to mean 0, variance
1
-------------------------------------------------------------------------
**by** is allowed; see **[D] by**.

__Menu__

**Statistics > Multivariate analysis > Cronbach's alpha**

__Description__

**alpha** computes the interitem correlations or covariances for all pairs of
variables in *varlist* and Cronbach's alpha statistic for the scale formed
from them. At least two variables must be specified with **alpha**.

__Options__

+---------+
----+ Options +----------------------------------------------------------

**asis** specifies that the sense (sign) of each item be taken as presented
in the data. The default is to determine the sense empirically and
reverse the scorings for any that enter negatively.

**casewise** specifies that cases with missing values be deleted listwise.
The default is pairwise computation of covariances and correlations.

**detail** lists the individual interitem correlations and covariances.

**generate(***newvar***)** specifies that the scale constructed from *varlist* be
saved in *newvar*. Unless **asis** is specified, the sense of items
entering negatively is automatically reversed. If **std** is also
specified, the scale is constructed by using standardized (mean 0,
variance 1) values of the individual items. Unlike most Stata
commands, **generate()** does not use casewise deletion. A score is
created for every observation for which there is a response to at
least one item (one variable in *varlist* is not missing). The
summative score is divided by the number of items over which the sum
is calculated.

**item** specifies that item-test and item-rest correlations and the effects
of removing an item from the scale be displayed. **item** is valid only
when more than two variables are specified in *varlist*.

**label** requests that the detailed output table be displayed in a compact
format that enables the inclusion of variable labels.

**min(***#***)** specifies that only cases with at least *#* observations be included
in the computations. **casewise** is a shorthand for **min(***k***)**, where *k* is
the number of variables in *varlist*.

**reverse(***varlist***)** specifies that the signs (directions) of the variables
(items) in *varlist* be reversed. Any variables specified in **reverse()**
that are not also included in **alpha**'s *varlist* are ignored.

**std** specifies that the items in the scale be standardized (mean 0,
variance 1) before summing.

__Examples__

Setup
**. webuse automiss**

Obtain average interitem covariance and Cronbach's alpha
**. alpha price headroom rep78 trunk weight length turn displ**

Obtain item-test and item-rest correlations and individual interitem
correlations
**. alpha price headroom rep78 trunk weight length turn displ, std item**
**detail**

__Stored results__

**alpha** stores the following in **r()**:

Scalars
**r(alpha)** scale reliability coefficient
**r(k)** number of items in the scale
**r(cov)** average interitem covariance
**r(rho)** average interitem correlation if **std** is
specified

Matrices
**r(Alpha)** scale reliability coefficient
**r(ItemTestCorr)** item-test correlation
**r(ItemRestCorr)** item-rest correlation
**r(MeanInterItemCov)** average interitem covariance
**r(MeanInterItemCorr)** average interitem correlation if **std** is
specified

If the option **item** is specified, results are stored as row matrices for
the **k** subscales when one variable is removed.