[Niels Henrik Bruun]

Stephen Jenkins Thank you Have grown quite fond of it myself. Note that it also find ado files for extended functions.

[Niels Henrik Bruun]

Thanks to Kit Baum there is a new update.
Underlying code is changed.
Minor changes.

[Niels Henrik Bruun]

Thanks to Kit Baum there is a new update.

Styling is modified.
Option toxl for export to Excel is added to -matprint-, -sumat- and -regmat-
-regmat- is added two options for refining the output.

[Niels Henrik Bruun]

Thanks to Kit Baum is there a new update of matrixtools at SSC.
There was a bug in the toxl parser

[Niels Henrik Bruun]

Thanks to Kit Baum is there a new update of matrixtools at SSC.
Added command -onewai-, which is an immediate oneway analysis of variance command similar to -oneway- and -loneway-.
However, here all results are gathered in matrices.
Input is either a matrix of n's, means and standard deviations or n's, means and standard deviations given separately.

Eg. compare three means with same n and two standard deviations:

. onewai, n(20) means(122 113 130) sds(6 7)
  
Summary table of data:
------------------------------------------------------------
        N     Mean  Std Dev  Std Error  [95% Conf  Interval]
------------------------------------------------------------
1      20  122.000    6.000      1.342    119.192    124.808
2      20  113.000    7.000      1.565    109.724    116.276
3      20  130.000    7.000      1.565    126.724    133.276
Total  60  121.667    9.602      1.240    119.186    124.147
------------------------------------------------------------
  
Analysis of Variance:
------------------------------------------------------------
                            SS  df       MS      F  Prob > F
------------------------------------------------------------
Between groups         2893.33   2  1446.67  32.39      0.00
Within groups (Error)  2546.00  57    44.67                 
Total                  5439.33  59    92.19                 
------------------------------------------------------------
  
Bartlett's test of equal variances:
---------------------------------------
                  Chi2  df  Prob > chi2
---------------------------------------
Bartlett's test  0.567   2         0.75
---------------------------------------
  
ICC and asymptotic confidence interval:
---------------------------------------------------------------
                      Estimate  Std Error  [95% Conf  Interval]
---------------------------------------------------------------
ICC                      0.611      0.250      0.122      1.100
Treatment std dev        8.485                                 
Error within std dev     6.683                                 
---------------------------------------------------------------

. return list

matrices:
              r(anova) :  3 x 5
          r(bartletts) :  1 x 3
              r(table) :  4 x 6
                r(icc) :  3 x 4
              r(total) :  1 x 6

It is also integrated with eg command -sumat- (also in matrixtools).
Look at:

webuse apple
sumat weight, statistics(n mean sd) rowby(treatment)
onewai , nmsmatrix(r(sumat))

Enjoy

[Martyn Sherriff]

I get the following error:
. webuse apple
(Apple trees)

. sumat weight, statistics(n mean sd) rowby(treatment)
nhb_mt_labelmatrix::new(): 3021 class compiled at different times [3]
nhb_mt_labelmatrix(): - function returned error
lmtable(): - function returned error
<istmt>: - function returned error
r(2);

Is there anything else that I have to update?

Thanks.
Martyn

[Niels Henrik Bruun]

Dear Martyn
Thank you very much for letting me know.
However, I can't reproduce the error by me.

Please try

ssc install matrixtools, all replace

and see if this works.
Do you have -basetable- installed as well? If so -basetable- should be installed before -matrixtools-
Which version of Stata are you using?

Please let me know the result of running the code.

[Niels Henrik Bruun]

It seems like the new command -onewai- has slipped out of the ssc list.
I'll be back when the update is complete

[Martyn Sherriff]

I am using Stata 16.1 IC and basetable is:

*! Package basetable v 0.2.5
*! Support: Niels Henrik Bruun, [email protected]
*!version 0.2.5a 2021-01-28 &gt; Bug fixes
*!version 0.2.5 2020-12-28 &gt; Converted to ltoxl_v1[34].mata and added columnwidth for toxl
*!version 0.2.5 2020-12-07 &gt; varlists instead of varnames
*!version 0.2.5 2020-12-07 &gt; First variable _none
*!version 0.2.5 2020-12-07 &gt; Overall if kept

However

ssc install matrixtools, all replace

seems to have solved the problem. I ran

ado update matrixtools, update

before.

Thanks for

[Niels Henrik Bruun]

Thanks to speedy action from Kit the update is complete. So forget about the comment #23

[Niels Henrik Bruun]

Martyn Sherriff Nice to hear. Hope -onewai- becomes useful for you

[Niels Henrik Bruun]

Thanks to Kit Baum there is a new update of matrixtools.
Minor corrections and a new -egen- command marktop, which collapse a categorical variable into fewer categories eg top 5 and/or selected values. The remaining values are combined into one category with default label other.

[Niels Henrik Bruun]

Thanks to kKit Baum, there is a new update of -matrixtools.
Changes are:

• -regmat-
  • A version -stregmat- is made for st set data
  • If there is no convergence, the estimates are set to missing
  • option nonzero is added. When set, zero values are not shown
• -matprint-
  • option nonzero is added. When set, zero values are not shown
• -sumat-
  • option nonzero is added. When set, zero values are not shown
  • bug when reporting deciles by idi is fixed
• -subselect-
  • is rewritten in Stata code instead of Mata code to be more efficient
  • more options are added
• -confmat- is added
  • It estimates sensitivity, specificity, prevalence, accuracy, NPV, and PPV with confidence intervals. Just like, e.g., -diagt- do
  • there are options for the calculation of the confidence intervals (see -cii proportions-)
  • it is possible to make pairwise comparisons (see -cs-) of sensitivity, specificity, prevalence, accuracy, NPV, and PPV

Enjoy

[Niels Henrik Bruun]

Thanks to Kit Baum, there is a new update of -matrixtools-.
There was a bug in -confmat-. And due to my mistake the new files, ie -confmat- and -stregmat-, weren't accessible.
But now, everything should be fine.
Enjoy

[Niels Henrik Bruun]

Thanks to Kit Baum is there a new update of matrixtools at SSC.
Some minor bug fixes. And adding option todocx which makes it possible to send one table into one wordfile.