# pwcorr_a

**Repository Path**: arlionn/pwcorr_a

## Basic Information

- **Project Name**: pwcorr_a
- **Description**: stata command for pairwise correlation matrix
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 63
- **Forks**: 16
- **Created**: 2017-10-18
- **Last Updated**: 2026-03-13

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

## pwcorr_a

For details, see 
- 王若溪, 连玉君, 2020, [Stata结果输出：pwcorr_a输出相关系数矩阵](https://www.lianxh.cn/details/348.html), 连享会 No.348.


The command `pwcorr_a` computes and displays the pairwise correlation coefficients for the variables listed in `varlist`. If `varlist` is not specified, it calculates the correlations for all variables in the dataset.

The key distinction between `pwcorr_a` and Stata's official `pwcorr` command lies in their handling of significance levels. Specifically, `pwcorr_a` can simultaneously display `***` (indicating significance at the 1% level), `**` (significance at the 5% level), and `*` (significance at the 10% level). In contrast, the `pwcorr` command is limited to displaying only one significance level at a time.

## Install

The installation files for `pwcorr_a` are stored on the [lianxh.cn](https://www.lianxh.cn) server and can be installed using the `cnssc` command. Therefore, you need to first install the `cnssc` command by running `ssc install cnssc`. The complete installation commands are as follows:

```stata
ssc   install cnssc, replace
cnssc install pwcorr_a, replace
```

Then you can type `help pwcorr_a` to view the help ducoment, or check the version of `pwcorr_a`:

```Stata 
. which pwcorr_a

D:/stata/plus\p\pwcorr_a.ado
*! version 3.0.9  26Jan2025
```



## The difference between `pwcorr` and `pwcorr_a` ?

The defult syntax of `pwcorr_a` is same as Stata's official command `pwcorr`:

```stata
.   sysuse auto, clear
(1978 Automobile Data)

. pwcorr price wei len mpg

             |    price   weight   length      mpg
-------------+------------------------------------
       price |   1.0000 
      weight |   0.5386   1.0000 
      length |   0.4318   0.9460   1.0000 
         mpg |  -0.4686  -0.8072  -0.7958   1.0000 

. pwcorr_a price wei len mpg

             |  price   weight   length     mpg   
-------------+------------------------------------
   price     |    1.000   
   weight    |    0.539***   1.000   
   length    |    0.432***   0.946***   1.000   
    mpg      |   -0.469***  -0.807***  -0.796***   1.000  
```

## Customizing Significance Level Indicators

By default, `pwcorr_a` uses `***`, `**`, and `*` to denote significance levels at the $1\%$, $5\%$, and $10\%$ thresholds, respectively. For example:

```stata
. sysuse "nlsw88.dta", clear
. local xx "wage hours ttl_exp age tenure"
. pwcorr_a `xx' if 3.race

         |     wage    hours  ttl_exp      age   tenure
---------+---------------------------------------------
    wage |   1.000   
   hours |   0.415**   1.000   
 ttl_exp |   0.349*    0.467**   1.000   
     age |  -0.461**  -0.115     0.070     1.000   
  tenure |   0.236     0.219     0.571***  0.203     1.000   
```

To customize the significance level indicators, you can use the **star#()** option. For instance, if you want to use `***`, `**`, and `*` to represent the $0.1\%$, $1\%$, and $5\%$ significance levels, respectively. Here is an example:

```stata
. sysuse "nlsw88.dta", clear
. local xx "wage hours ttl_exp age tenure"
. pwcorr_a $xx if 3.race, ///
     star1(0.001) star5(0.01) star10(0.05)

         |     wage    hours  ttl_exp      age   tenure
---------+---------------------------------------------
    wage |   1.000   
   hours |   0.415*    1.000   
 ttl_exp |   0.349     0.467*    1.000   
     age |  -0.461*   -0.115     0.070     1.000   
  tenure |   0.236     0.219     0.571**   0.203     1.000   
```



## Export to Word document

You can use the `logout` command to export the correlation matrix generated by `pwcorr_a` to an Excel or Word document. The command is as follows:

```stata
*-----Table 2: Correlation Matrix-------
sysuse auto, clear
local v "price wei len mpg" // Specify variables
local s "Table2_corr"       // file name
logout, save("`s'") excel replace:  ///
        pwcorr_a `v', format(%6.2f) //star(0.05)
``` 

This code snippet demonstrates how to save the correlation matrix for the specified variables (`price`, `weight`, `length`, and `mpg`) into an Excel file named `Table2_corr`. The `format(%6.2f)` option ensures the correlation coefficients are displayed with two decimal places. The `star(0.05)` option (commented out here) can be used to highlight significance levels if needed.

![20250126185414](https://fig-lianxh.oss-cn-shenzhen.aliyuncs.com/20250126185414.png)