[paulvonhippel]

Faster random slope models. Currently -mixed- and -meglm- are unacceptably slow, and often do not converge at all, when fitting random slope models to moderately large datasets.

[Bruce Weaver]

This is a repeat of a request I posted for v18.

As discussed in this thread, there is no direct and easy way to use some of the alternative covariance structures (ACSs) Lesa Hoffman illustrates in Chapter 4 of her 2015 book on Longitudinal Analysis. I would like to see more ACS options for -mixed- similar to the options available in both SAS and SPSS.

• For SAS options, see Table 65.17 here
• For SPSS options, see this page

Cheers,
Bruce

[John Kane]

1. A built in suite to create animated graphs. For example, you save a set of .png/.gph/.pdf files and then use a Stata suite that allows you to combine them all into one animated .gif, with some controls over the speed of the .gif (e.g, see the "ezgif.com" website to get an idea of what controls could be used.)
2. More capability for text analysis.
3. Built-in package for making maps
4. A scheme that includes a gradient of colors within one graph (Excel graphs have these--e.g., it might be darker toward the edges of a graph and lighter in the center).

[Abdullah Algarni]

Adopt community -diagt- command as a built-in command, with further improvement (e.g., have interactive graphic control window)

[Bruce Weaver]

Re #139, I like -diagt- (SSC), but I wish that it had test result as the row variable (T+ in row 1, T- in row 2) and true disease status as the column variable (D+ in column 1, D- in column 2). I know that some resources on diagnostic tests lay out the table as the author of -diagt- has done it, but in the resources I have consulted, at least, it is far more common to lay out the table with test result as the row variable and true disease status as the column variable.

If -diagt- is adopted, perhaps table layout could be an option? ;-)

[Abdullah Algarni]

Re #140, totally agree 👍

[Lars Egner]

The generate option of the alpha function is not in line with standard ways to handle reversed items in a index. Currently, it "reverses" the scale by adding a "-" in front of the value. The standard way (at least in my own field), is to "turn" the item so that the highest score becomes the lowest score. For example on a Likert scale, 7 becomes 1, 6 becomes 2, ect.
The results of the two approaches are different when trying to create an index of the responses. In my opinion the flawed logic of the alpha, generate() becomes especially clear when some items have missing responses. In my limited experience their correlation hovers around .8.
For example, on 3 variables on a 1-7 Likert scale where V3 is reversed :

V1	V2	V3	alpha , generate()	Normal (IMO) approach
1	1	7	-1.7	1
3	3	5	0.3	3
3	3	.	3	3

I would personally want the logic of the rightmost column to be the default way of creating indexes with the alpha function. At least having the option of applying the right column would be very appreciated.

[daniel klein]

Re: #142

Arguably, alpha (Cronbach's alpha that is) is used with continuous variables. Therefore, multiplying with (-1) is not a (logical) flaw at all. Quite the opposite. Say, we reverse 1, 2, 42. The distance between two observations with the two lowest values jumps from 1 to 41; this I would call flawed.

Also, social scientists may well have a different understanding of the "standard way" of handling reversed items than, say, geologists.

[Lars Egner]

Re: #143

Fair points. I underestimated how flawed the suggested method would be with the type of datasets I don't usually work with. I change my wish to include an option in the alpha command to reverse the scores instead of inversing them.

[Bert Lloyd]

Improvements to graph text, e.g.

Aligning subscripts and superscripts in graphs' text elements as requested by John Mullahy

allowing "hats" and other decorations over characters as requested by Michael Anbar

Ideally there would be a simple way to enter "latex mode", e.g.
title("{latex: $\hat{\beta}_{0}^{i}$}")
or
title("{latex: $\hat{\beta}_{0}^{i}$}")

[Chris Stefancik]

Whenever I work on a new project, I start by copying and pasting a "header" or "preamble" to each new do file. I searched to see if there was functionality to automatically append something like this whenever I start a new file but came up short on the search. I think this would be a useful enhancement. For example, here the default header I load (nothing special but seems to be the most common first-steps for me):

********************************************************************************
* Author:    
* Date created:    
* Filename:        
* Description:
* Prerequisite steps:    
*
********************************************************************************
* Change Log:
*
*    08/04/2022    Created File
*
********************************************************************************
* Initialization
********************************************************************************
        discard
        clear
        clear mata
        clear all
        estimates clear
        snapshot erase _all

********************************************************************************
* Logs
********************************************************************************        
        capture cd "\Stata Logs\"

        local datetime_string : display %tc_CCYYNNDD_HHMMSS clock(c(current_date) + c(current_time), "DMYhms")
        local datetime_string = subinstr("`datetime_string'", " ", "", .)
        
        capture log close
        log using "__`datetime_string'.txt", replace text
********************************************************************************

[Leonardo Guizzetti]

This could be something you wrap into an ado file, but I find it easier in my workflow to either copy a blank header file as needed, or use a text editor with a macro capability to auto-paste in the text.

[Scott Adams]

Allow graphs to be exported without drawing in the graph window (which is slow for large datasets), whether after creating the graph with nodraw option, or adding a direct export() option to twoway and other graph commands (which could be used along with nodraw option).

[Bert Lloyd]

Allow graphs to be exported without drawing in the graph window (which is slow for large datasets), whether after creating the graph with nodraw option, or adding a direct export() option to twoway and other graph commands (which could be used along with nodraw option).

Scott Adams does the following not achieve what you want?

set graphics off
twoway scatter y x
graph export mygraph.ext, replace // .ext could be .pdf, .png, etc

[wbuchanan]

Scott Adams would there be any undesirable side-effects of directly writing a raster image to disk without rendering first? Vector formats make sense to me since they effectively contain the information on how to render the image, but raster images (e.g., png, jpg, gif, tiff) work a bit differently.