Question about calculating beta post analysis

[Hutchins Yeo]

The formula to calculate beta is: b*(s_x)/s_y)

After running the following code

sysuse nlsw88, clear
sum union wage
reg union wage hours, beta

We see that for x = wage:

b = .0017124
sd_x = 5.755523
sd_y = 0.4304825

. di (.0150858)*(5.755523/.4304825)
.20169616

However, this is clearly wrong. -reg, beta- says the actual standardized coefficient is .1460648. What am I missing? I think the issue is differences in SD calculation for binary? For instance, these don't match:

. di sqrt(.2454739*(1-.2454739)/1878)
.00993098

. sum union

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
       union |      1,878    .2454739    .4304825          0          1

[William Lisowski]

Let's look at the actual output, rather than your report of the output.

. sysuse nlsw88, clear
(NLSW, 1988 extract)

. sum age wage

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
         age |      2,246    39.15316    3.060002         34         46
        wage |      2,246    7.766949    5.755523   1.004952   40.74659

. reg age wage hours, beta

      Source |       SS           df       MS      Number of obs   =     2,242
-------------+----------------------------------   F(2, 2239)      =      2.00
       Model |  37.5163127         2  18.7581563   Prob > F        =    0.1350
    Residual |  20954.3142     2,239  9.35878258   R-squared       =    0.0018
-------------+----------------------------------   Adj R-squared   =    0.0009
       Total |  20991.8305     2,241  9.36717113   Root MSE        =    3.0592

------------------------------------------------------------------------------
         age |      Coef.   Std. Err.      t    P>|t|                     Beta
-------------+----------------------------------------------------------------
        wage |   -.017124    .011369    -1.51   0.132                -.0322135
       hours |  -.0066185   .0062286    -1.06   0.288                -.0227258
       _cons |     39.532   .2433198   162.47   0.000                        .
------------------------------------------------------------------------------

So we see that your "x" and "y" correspond to wage (sd = 5.755523) and age (sd = 3.060002), reversed from what you transcribed in post #1.

I think perhaps you intended for wage, not age, to be the independent variable, and didn't notice that you'd reversed them in your regression command.

[Hutchins Yeo]

Let's look at the actual output, rather than your report of the output.

. sysuse nlsw88, clear
(NLSW, 1988 extract)

. sum age wage

Variable | Obs Mean Std. Dev. Min Max
-------------+---------------------------------------------------------
age | 2,246 39.15316 3.060002 34 46
wage | 2,246 7.766949 5.755523 1.004952 40.74659

. reg age wage hours, beta

Source | SS df MS Number of obs = 2,242
-------------+---------------------------------- F(2, 2239) = 2.00
Model | 37.5163127 2 18.7581563 Prob > F = 0.1350
Residual | 20954.3142 2,239 9.35878258 R-squared = 0.0018
-------------+---------------------------------- Adj R-squared = 0.0009
Total | 20991.8305 2,241 9.36717113 Root MSE = 3.0592

------------------------------------------------------------------------------
age | Coef. Std. Err. t P>|t| Beta
-------------+----------------------------------------------------------------
wage | -.017124 .011369 -1.51 0.132 -.0322135
hours | -.0066185 .0062286 -1.06 0.288 -.0227258
_cons | 39.532 .2433198 162.47 0.000 .
------------------------------------------------------------------------------

So we see that your "x" and "y" correspond to wage (sd = 5.755523) and age (sd = 3.060002), reversed from what you transcribed in post #1.

I think perhaps you intended for wage, not age, to be the independent variable, and didn't notice that you'd reversed them in your regression command.

Actually, I used union instead of age. I originally had age accidentally.