@Mauricio Below link takes you to a tuitorial on how to do Welch's Test in Stata.

https://www.statology.org/welchs-t-test-stata/#:~:text=The%20most%20common%20way%20to,two-sample%20t-test.&text=If%20you%20suspect%20that%20the,two-sample%20t-test.

@Mauricio I really don't recommend using ANOVA without first using some corrective measures. I suggest perhaps a variance stabilizer. See Box Cox method which should be be in Stata. Stay safe, David Booth

Welch talks about doing this for multiple means: Welch, B. L. (1951). "On the Comparison of Several Mean Values: An Alternative Approach". Biometrika. 38 : 330-336. \url{doi:10.2307/2332579}.

Imran Anwar Unfortunately, it is not possible to perform Welch’s ANOVA. So I performed it in R. Thanks.

https://www.statology.org/welchs-anova-in-r/

It is the default for oneway.test in R, but it is possible in STATA, just requires writing it.

Daniel Wright What should I type in STATA? Please help

Mauricio López-Espejo , I don't use STATA. Anyone who uses STATA seriously will use its underlying language. Though as noted above doing it is R may be easiest. If for some reason it is is necessary not just to read the data into R, see https://stackoverflow.com/questions/29631612/calling-r-from-stata.

I'm about a month late to this party, but see the attached DO file for code that computes Welch's F. For the sample data I used, the results match those from the ONEWAY command in SPSS. HTH.

PS- Despite the .do extension, the file is just a text file.

Just for the record, this is the R code for Bruce Weaver 's example data. Same results. The Welch citation can be found here: https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/oneway.test

# Bruce Weaver's example data

Data = read.table(header=TRUE, stringsAsFactors=TRUE, text="

group y

1 3

1 2

1 1

1 1

1 4

2 5

2 2

2 4

2 2

2 3

3 7

3 4

3 5

3 3

3 6

")

oneway.test(y ~ group, data=Data)

Thanks for confirming the results, Sal. For completeness, here is (a minimal version of) the ONEWAY command for SPSS:

ONEWAY y BY group /STATISTICS=WELCH.

Well, then, here's the code for SAS. (Same results, n'at):

/* Bruce Weaver's example data */

data Data;

input Group $ Y;

datalines;

1 3

1 2

1 1

1 1

1 4

2 5

2 2

2 4

2 2

2 3

3 7

3 4

3 5

3 3

3 6

;

proc glm;

class Group;

model Y = Group;

means Group / welch;

run;

Additional note: For a free, GUI-based option for Welch's anova, Jamovi (Windows, MacOS, Linux, ChromeOS) conducts Welch's anova by default for one-way anova.