How to use Star for Significance Level

[Imran]

Dear,


I am using stata for panel data estimation. I want to use star for the significance level and i am using the below method. But in this method the significance level did not show when i estimate the equation/Correlation. Please tell me the method/ way, when i estimate the equation/Correlation the significance level show automatically.
Statistics > Postestimation > Manage estimation results > Table of estimation results

[Joe Canner]

When you get to the menu (Statistics > Postestimation > Manage estimation results > Table of estimation results) click the check box at the bottom ("Denote significance of coefficients with stars"). You can also choose which p-values indicate significance. By default one star is p<0.05, two stars is p<0.01 and three stars is p<0.001. Note also that you can accomplish the same thing without the menu, by typing the command:

estimates table, star(.05 .01 .001)

[Roger Newson]

You should be able to do this using the -parmest- package, which you can download from SSC. The -parmest- package has a -list()- option, causing coefficients to be listed, and a -stars()- option, causing stars to be displayed to indicate significance levels.

For instance, the user might use the -auto- data to fit a regression of -mpg- with respect to levels of -rep78- and -foreign- by, typing

sysuse auto, clear
regress mpg i.rep78 i.foreign

The user might then display the confidence intervals and P-values with stars, indicating significance at the leels 0.05, 0.01 and 0.001, by typing

parmest, list(,) stars(0.05 0.01 0.001)

which lists the parameters, with their P-values and confidence limits, with stars after the P-values.

Alternativey, the user might use the -format()- option of -parmest- to display the parameters and their confidence limits to a sensible number of decimal places. As in:

parmest, list(,) stars(0.05 0.01 0.001) format(estimate min* max* %8.2f)

The -parmest- package creates an output dataset (or resultsset), with 1 observation per estimated parameter and data on the parameter names, estimates, confidence intervals and P-values, and other parameter attributes specified by the user. This resultsset can be listed and then discarded, as in these examples. Or it can be saved to a disk file, using the -saving()- option. Or it can be saved to the memory, overwriting the existing dataset, using the -norestore- option. The parameters may be listed, tabulated or plotted.

I hope this helps.

Best wishes

Roger

[Imran]

When you get to the menu (Statistics > Postestimation > Manage estimation results > Table of estimation results) click the check box at the bottom ("Denote significance of coefficients with stars"). You can also choose which p-values indicate significance. By default one star is p<0.05, two stars is p<0.01 and three stars is p<0.001. Note also that you can accomplish the same thing without the menu, by typing the command:

estimates table, star(.05 .01 .001)

Thanx Dear I am also adopt this method but in this method the star did not show when i estimate the equation.
Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------y
| .0005323 .0001273 4.18 0.000 .0002822 .0007825
| .0000656 .0013123 0.05 0.960 -.0025127 .002644
| -.0056204 .0051406 -1.09 0.275 -.0157205 .0044796
| .0125138 .0088619 1.41 0.159 -.0048977 .0299252

if i use the code which u send me, then the result will be show as

estimates table, star(.05 .01 .001)

------------------------------
Variable | active
-------------+----------------
? | .00053231***
? | .00006562
? | -.00562043
? | .01251375
------------------------------
legend: * p<.05; ** p<.01; *** p<.001

But i want , when i estimate the equation the significance level will be show on the result......... I am very thankful for this help and wait for ur reply

[Joe Canner]

The significance stars are not available in estimates table if you also include the standard error, t-statistic, and/or p-value. So, you will have to use something like parmest, as suggested by Roger Newson above:

parmest, list(,) stars(0.05 0.01 0.001) format(estimate min* max* %8.2f)

[Nick Cox]

But it does show. The first coefficient is flagged *** and the others aren't flagged. That's consistent with the P-values given alongside.

[Imran]

You should be able to do this using the -parmest- package, which you can download from SSC. The -parmest- package has a -list()- option, causing coefficients to be listed, and a -stars()- option, causing stars to be displayed to indicate significance levels.

For instance, the user might use the -auto- data to fit a regression of -mpg- with respect to levels of -rep78- and -foreign- by, typing

sysuse auto, clear
regress mpg i.rep78 i.foreign

The user might then display the confidence intervals and P-values with stars, indicating significance at the leels 0.05, 0.01 and 0.001, by typing

parmest, list(,) stars(0.05 0.01 0.001)

which lists the parameters, with their P-values and confidence limits, with stars after the P-values.

Alternativey, the user might use the -format()- option of -parmest- to display the parameters and their confidence limits to a sensible number of decimal places. As in:

parmest, list(,) stars(0.05 0.01 0.001) format(estimate min* max* %8.2f)

The -parmest- package creates an output dataset (or resultsset), with 1 observation per estimated parameter and data on the parameter names, estimates, confidence intervals and P-values, and other parameter attributes specified by the user. This resultsset can be listed and then discarded, as in these examples. Or it can be saved to a disk file, using the -saving()- option. Or it can be saved to the memory, overwriting the existing dataset, using the -norestore- option. The parameters may be listed, tabulated or plotted.

I hope this helps.

Best wishes

Roger

Sorry dear, I did not understand what are u saying

[Nick Cox]

Roger gave you code you can try for a program you can download with a help file you can read.

[Roger Newson]

Imran may not know about SSC. The SSC archive is described under

help ssc

and you can view the entry for the -parmest- package by typing, in Stata,

ssc describe parmest

and install it by typing

ssc install parmest

and then type

help parmest

to see the online help for -parmest-. This online help contains a few examples, plus some hypertext references at the bottom of the online help, where you can use your browser to find more about -parmest-.

I hope this helps.

Best wishes

Roger

[Imran]

The significance stars are not available in estimates table if you also include the standard error, t-statistic, and/or p-value. So, you will have to use something like parmest, as suggested by Roger Newson above:

parmest, list(,) stars(0.05 0.01 0.001) format(estimate min* max* %8.2f)

Thanx for this and Please verify the code again. When i enter this code in Stata the result will be shown as

parmest, list(,) stars (0.05 0.01 0.001) format(estimate min* max* %8.2f)
unrecognized command: parmest
r(199);

[Joe Canner]

Please re-read Roger Newson's post #9. You need to install parmest first since it is not part of Stata:

ssc install parmest

[Imran]

Imran may not know about SSC. The SSC archive is described under

help ssc

and you can view the entry for the -parmest- package by typing, in Stata,

ssc describe parmest

and install it by typing

ssc install parmest

and then type

help parmest

to see the online help for -parmest-. This online help contains a few examples, plus some hypertext references at the bottom of the online help, where you can use your browser to find more about -parmest-.

I hope this helps.

Best wishes

Roger

Thanx dear for Guidence. I follow below pattern. I write the below command (ssc install outreg) on stata then i write next command( ssc install parmest) the result shown as and please guide me next what i do.

ssc install outreg
checking outreg consistency and verifying not already installed...
all files already exist and are up to date.

. ssc install parmest
checking parmest consistency and verifying not already installed...

the following files already exist and are different:
c:\ado\plus\p\parmest.ado

no files installed or copied
(no action taken)

[Joe Canner]

Try:

ssc install parmest, replace

[Hamada Sukkar]

I love you guys, I was browsing to get some answers to a problem I was facing, I just got my answers.
I really appreciate your efforts

[Hamada Sukkar]

ssc install parmest, replace
parmest, list(,) stars (0.05 0.01 0.001) format(estimate min* max* %8.2f)

solved my problem