How to use If condition to restrict t test on only two groups with a categorical var that has 3 groups

[Shamsudini Amidu]

Hi Family,
I have 1 categorical variable that has 3 grps( blue,yellow, green), I wanted check if the variance between only these two groups( blue and green) differs. So how do i use if condition to constraint the categorical variable to work with only the 2 groups out of the 3.

Given that;
the value label for blue=1 yellow=2 and green=3
wait= Weight of individuals

Below is the command I used but stata still generate an error.


sdtest wait , by( grps ) if grps==1 & grps==3

option if not allowedr(198);

I need your help?

Thank you.
Amidu

[Joseph Coveney]

Try

sdtest wait if inlist(grps, 1, 3), by(grps)

[Shamsudini Amidu]

Hi Joseph,
Thank you soo much.
It works so well.

[Nick Cox]

 if grps==1 | grps==3

would be the if qualifier you need if you didn't use inlist().

[Shamsudini Amidu]

Ok thanks

[Shamsudini Amidu]

Dear Sir/Madam,
I am currently working on a data set in a roster format containing timestamp, and I am facing an issue with the conversion of the timestamps to the required format. I have created two new variables, but unfortunately, I keep getting a "type mismatch" error while using the collapse command to sum the time differences according to the unique ID.
To convert the timestamps, I used the following commands:

• gen start_time_stata = clock(START_TIME_SECS, "YMDhms"): to convert the start timestamp
• gen end_time_stata = clock(STOP_TIME_SECS, "YMDhms"): to convert the stop timestamp
• gen time_diff = tdelta(end_time_stata, start_time_stata, "seconds"): to get the difference

I would greatly appreciate your help in resolving this issue.
I have attached the dataset for your reference.

* Example generated by -dataex-. To install: ssc install dataex
clear
input double(MRESPONSE_ID START_TIME_SECS STOP_TIME_SECS)
1 1681182712 1681182762
1 1681183070 1681183080
1 1681183410 1681183423
1 1681183467 1681183472
2 1681184993 1681185138
2 1681185312 1681185464
2 1681185708 1681185926
end



Thank you very much in advance.


Sham

[Joseph Coveney]

1 It would have been better to have started a new topic.

2. clock() takes a string variable as its first argument. Those variables are not string.

3. Use double-precision when working with date-time variables. (Actually, use double-precision for all floating-point variables and you won't have to worry about such things.)

4. I didn't find the function tdelta() in Stata's help file for date-time functions.

5. Try something along the lines of the following:

.ÿ
.ÿversionÿ17.0

.ÿ
.ÿclearÿ*

.ÿ
.ÿquietlyÿinputÿdouble(MRESPONSE_IDÿSTART_TIME_SECSÿSTOP_TIME_SECS)

.ÿ
.ÿ*
.ÿ*ÿBeginÿhere
.ÿ*
.ÿformatÿ*TI*ÿ%tcHH:MM:SS.sss

.ÿrenameÿ*,ÿlower

.ÿgenerateÿdoubleÿtime_diffÿ=ÿclockdiff_frac(start_time_secs,ÿstop_time_secs,ÿ"second")

.ÿformatÿtime_diffÿ%05.3f

.ÿ
.ÿlist,ÿnoobsÿabbreviate(20)ÿsepby(mresponse_id)

ÿÿ+-------------------------------------------------------------+
ÿÿ|ÿmresponse_idÿÿÿstart_time_secsÿÿÿstop_time_secsÿÿÿtime_diffÿ|
ÿÿ|-------------------------------------------------------------|
ÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ10:59:42.712ÿÿÿÿÿ10:59:42.762ÿÿÿÿÿÿÿ0.050ÿ|
ÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ10:59:43.070ÿÿÿÿÿ10:59:43.080ÿÿÿÿÿÿÿ0.010ÿ|
ÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ10:59:43.410ÿÿÿÿÿ10:59:43.423ÿÿÿÿÿÿÿ0.013ÿ|
ÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ10:59:43.467ÿÿÿÿÿ10:59:43.472ÿÿÿÿÿÿÿ0.005ÿ|
ÿÿ|-------------------------------------------------------------|
ÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿ2ÿÿÿÿÿÿ10:59:44.993ÿÿÿÿÿ10:59:45.138ÿÿÿÿÿÿÿ0.145ÿ|
ÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿ2ÿÿÿÿÿÿ10:59:45.312ÿÿÿÿÿ10:59:45.464ÿÿÿÿÿÿÿ0.152ÿ|
ÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿ2ÿÿÿÿÿÿ10:59:45.708ÿÿÿÿÿ10:59:45.926ÿÿÿÿÿÿÿ0.218ÿ|
ÿÿ+-------------------------------------------------------------+

.ÿ
.ÿexit

endÿofÿdo-file


.

[Shamsudini Amidu]

Dear Joseph and Team,,

I would like to express my gratitude for your prompt advice. I must inform you that I do not possess the Version 17 of Stata, but instead I am utilizing Version 14. Although I have thoroughly reviewed your guide, I am unable to comprehend the instructions, possibly due to my utilization of the outdated version.

I am grateful for your suggestion, however, I believe it may be a distinct topic and therefore, I plan to repost my inquiry. Please accept my apologies for any inconvenience this may have caused.

Thank you once again for your assistance.