Today is 19 Dec 2021
This is the version of 20. December 2021This introductory course was financed by the Axe Santé / Health Technology Innovation Center of the HES-SO Valais-Wallis.
Just a few words on how the webpages of this course were made. I used dyndoc see for example this youtube video here.
Here is another nice website for a tutorial on [how to use dyndoc.](https://www.techtips.surveydesign.com.au/post/how-to-create-an-html-webpage-in-stata-using-markdown)
In a nutshell: I created a txt file where i can mix comments, text and stata code together. I used some twists: For the comments, I had problems with the "[comment]: <> (this is a comment )" approach, as some comment lines were still shown in the html document.
Therefore I used the solution proposed here.
I added some jquery to add a simple table of content see here and css see here
In the end, you put all in one folder, set the working directory of Stata to this folder and then type dyndoc name_of_the_textfile.txt in the command line of Stata. If you want to replace the created html file, add ,replace to the command.
Of course, more time would buy me the opportunity to make this better looking…
This is not a course on statistics, just an introduction to Stata. We use Stata 17 for this course - but most things will work in older versions too. By the way, its Stata and not - as a lot of people write in their publications - STATA. So behave and don’t shout see here for the proof There is a face to face course with this content, however, the course could also be done alone by just following these pages. As almost always, there will be some errors and I would be happy for a comment.
Stata has an excellent collection of books, check out their bookstore
There are different flavours of stata and before you buy a licence, you should check this website to decide on which version you need
By the way, Stata has its own YouTube channel.
Stata has several windows that you will use. You can move this windows and rearrange them differently, if you like see a short video here. By the way, if you rearranged the windows, you might also want to change the colours: edit–>preferences–>genereal preferences–>Results and Viewer to e.g. classic see video here
Some settings can easily be found by clicking the edit–>preferences menues; however, would be nice to do this programmatically. type query in the command line (and hit enter) You’ll see a list of settings you can change.
We will not talk that much about the point and click menu. During the first lessons, we will mainly use the command windows. Of course, everatime you want something to be stored and be reproducable, you would want to create a do-file.
Please see this video for a short illustration of the three methods
Most commands you will use are built in already in Stata. From time to time, you might want to do something that is either not exactly possible or you do not like the way Stata does it, then you can download user written commands, so called ado-files. If you know the name of an ado-file, you can download it from the command line with ssc install name_of_ado . Or you could type findit name_of_ado.
If you would like to watch a video that show how you could produce user written commands, you find a video here.
Minor updates, i.e. the ones not requiring a new licence, can be installed by typing update all in the commant line; or you type update query. If you want to update ado files, you can write ado update in the command line. Stata will show a list with updatable ado-files. You can then update them with ado update, update
If you know the name of a command for which you might want some help, just type help name_of_command, so for example: help table In the helpfile, you will also see a link to the PDF documentation - which is just great with many examples. Most or all examples can be run with datasets that are provided by Stata. You can load the datasets with sysuse (for example sysuse auto). Some examples use webuse to download a dataset fromm a repository. If you type help resources in the commmand line, a list of very useful resources pops up. Stata has a very active user community and online you also find very useful help.
To play around with some commands, we will first load the data: copy the following into your command line:
You can have a look at the data by typing browse - or just br in the command line. If you would type edit or ed in the command line, then you could even edit the data - what we normally don’t want. Most of the data have some notes that comes with. So type notes into the command line and see what happens. We could also add a comment to the dataset. Try this out and write the following in the command line:
*notes: We downloaded this dataset with the following command use http://people.ucalgary.ca/~patten/Datasets/deidentified_dataset.dta into Stata. The paper can be found here: https://link.springer.com/article/10.1186/s12875-018-0862-y *
Then type again notes into the command line. Our new note will then appear. You can also add notes to a variable (what the authors already did) by writing *notes variablename: Age in years*. You can add more than one Note to each variable. We can use the command *describe* to get an overview of the data. You might also want to [look here for other useful commands to explore a dataset](https://stats.idre.ucla.edu/stata/seminars/notes/stata-class-notesexploring-data/) For example the command *codebook* gives a codebook - in the example below just for two variables. The command labelbook will present all the labels stored with the dataset. In the example below just for the labels named sex and SF369C
. use http://people.ucalgary.ca/~patten/Datasets/deidentified_dataset.dta
. describe
Contains data from http://people.ucalgary.ca/~patten/Datasets/deidentified_datas
> et.dta
Observations: 516
Variables: 133 13 Oct 2018 14:07
--------------------------------------------------------------------------------
Variable Storage Display Value
name type format label Variable label
--------------------------------------------------------------------------------
idno int %8.0g
PRACTICE float %9.0g
collectionpoint byte %8.0g COLLECTI Data collection point
ageR float %9.0g Age Group
CSRI2 byte %8.0g sex What is your gender?
CSRI3 byte %8.0g CSRI3 What is your marital status?
CSRI4 byte %8.0g CSRI4 What is the highest degree or
level of school you have
completed?
CSRI5 byte %8.0g CSRI5 What is your total household
income?
CSRI6 byte %8.0g CSRI6 What is your mother tongue?
CSRI7 byte %8.0g CSRI7 Country of Birth?
CSRI8 byte %8.0g CSRI8 Who do you usually live with?
CSRI9 byte %8.0g CSRI9 Current employment status
TIME float %9.0g
PHQ91 byte %8.0g PHQ91 Little interest or pleasure in
doing things
PHQ92 byte %8.0g PHQ92 Feeling down, depressed, or
hopeless
PHQ93 byte %8.0g PHQ93 Trouble falling or staying asleep,
or sleeping too much
PHQ94 byte %8.0g PHQ94 Feeling tired or having little
energy
PHQ95 byte %8.0g PHQ95 Poor appetite or overeating
PHQ96 byte %8.0g PHQ96 * Feeling bad about yourself or that
you are a failure or have let
yourself or you
PHQ97 byte %8.0g PHQ97 * Trouble concentrating on things,
such as reading the newspaper or
watching telev
PHQ98 byte %8.0g PHQ98 * Moving or speaking so slowly that
other poeple could have noticed?
Or the opposi
PHQ99 byte %8.0g PHQ99 Thoughts that you would be better
off dead or of hurting yourself
in some way
PHQ_TOT byte %8.0g
LEAPS1 byte %8.0g LEAPS1 What kind of paid work do you do
LEAPS2 byte %8.0g LEAPS2 hours scheduled or expected to
work
LEAPS3 byte %8.0g LEAPS3 hours of work missed because of
the way you were feeling
LEAPS4A byte %8.0g LEAPS4A low energy or motivation
LEAPS4B byte %8.0g LEAPS4B Poor concentration or memory
LEAPS4C byte %8.0g LEAPS4C Anxiety or irritability
LEAPS4D byte %8.0g LEAPS4D Getting less work done
LEAPS4E byte %8.0g LEAPS4E Doing poor quality work
LEAPS4F byte %8.0g LEAPS4F Making more mistakes
LEAPS4G byte %8.0g LEAPS4G Having trouble getting along with
people, or avioding them
LEAPS_M double %12.0g
LEAPS_tot double %12.0g
SDS1 byte %8.0g SDS1 The symptoms have disrupted your
work/school work:
SDS2 byte %8.0g SDS2 The symptoms have disrupted your
social life/leisure activities:
SDS3 byte %8.0g SDS3 The symptoms have disrupted your
family life/home
responsibilities:
SDS_M double %12.0g
SDSTOT_fixed float %9.0g
SDS4 byte %8.0g SDS4 days last week lost
SDS5 byte %8.0g SDS5 days last week unproductive
SF361 byte %8.0g SF361 In general, would you say your
health is
SF362 byte %8.0g SF362 Compared to one year ago, how
would you rate your health in
general now?
SF363A byte %8.0g SF363A * The following questions are about
activities you might do during a
typical day.
SF363B byte %8.0g SF363B * Moderate activities, such as
moving a table, pushing a vacuum
cleaner, bowling,
SF363C byte %8.0g SF363C Lifting or carrying groceries
SF363D byte %8.0g SF363D Climbing several flights of stairs
SF363E byte %8.0g SF363E Climbing one flight of stairs
SF363F byte %8.0g SF363F Bending, kneeling, or stopping
SF363G byte %8.0g SF363G Walking more than a mile
SF363H byte %8.0g SF363H Walking several hundred yards
SF363I byte %8.0g SF363I Walking one hundred yards
SF363J byte %8.0g SF363J Bathing or dressing yourself
SF364A byte %8.0g SF364A * During the past 4 weeks, how much
of the time have you had any of
the following
SF364B byte %8.0g SF364B Accomplished less than you would
like
SF364C byte %8.0g SF364C Were limited in the kind of work
or other activities
SF364D byte %8.0g SF364D * Had difficulty performing the work
or other activities (for
example, it took ext
SF365A byte %8.0g SF365A * During the past 4 weeks, how much
of the time have you had any of
the following
SF365B byte %8.0g SF365B Accomplished less than you would
like
SF365C byte %8.0g SF365C Did work or other activities less
carefully than usual
SF366 byte %8.0g SF366 * During the past 4 weeks, to what
extent has your physical health
or emotional pr
SF367 byte %8.0g SF367 How much bodily pain have you had
during the past 4 weeks?
SF368 byte %8.0g SF368 * During the past 4 weeks, how much
did pain interfere with your
normal work (incl
SF369A byte %8.0g SF369A * These questions are about how you
feel and how things have been
with you during
SF369B byte %8.0g SF369B Have you been very nervous?
SF369C byte %8.0g SF369C Have you felt so down in the dumps
that nothing could cheer you up?
SF369D byte %8.0g SF369D Have you felt calm and peaceful?
SF369E byte %8.0g SF369E Did you have a lot of energy?
SF369F byte %8.0g SF369F Have you felt downhearted and
depressed?
SF369G byte %8.0g SF369G Did you feel worn out?
SF369H byte %8.0g SF369H Have you been happy?
SF369I byte %8.0g SF369I Did you feel tired?
SF3610 byte %8.0g SF3610 * During the past 4 weeks, how much
of the time has your physical
health or emotio
SF3611A byte %8.0g SF3611A * How true or false is each of the
following statements for you? I
seem to get sic
SF3611B byte %8.0g SF3611B I am as healthy as anybody I know
SF3611C byte %8.0g SF3611C I expect my health to get worse
SF3611D byte %8.0g SF3611D My health is excellent
SF36_PF float %9.0g
SF36_RP float %9.0g
SF36_RE float %9.0g
SF36_VT float %9.0g
SF36_MH float %9.0g
SF36_SF float %9.0g
SF36_BP float %9.0g
SF36_GH float %9.0g
SF36_TOT float %9.0g
CSIa byte %23.0g CSI1 The services I get here are a big
help to me
CSIb byte %23.0g LABK People here really seem to care
about me
CSIc byte %23.0g LABK I would come back here if I need
help again
CSI4 byte %23.0g LABK I feel that no one here really
listens to me
CSI4_R float %9.0g
CSIe byte %23.0g LABK People here treat me like a
person, not like a number
CSIf byte %23.0g LABK I have learned a lot here about
how to deal with my problems
CSI7 byte %23.0g LABK People here want to do things
their way, instead of helping me
find my way
CSI7_R float %9.0g
CSIh byte %23.0g LABK I would recommend this place to
people I care about
CSIi byte %23.0g LABK People here really know what they
are doing
CSIj byte %23.0g LABK I get the kind of help here that I
really need
CSIk byte %23.0g LABK People here accept me for who I am
CSIl byte %23.0g LABK I feel much better now than when I
first came here
CSIm byte %23.0g LABK I thought no one could help me
until I came here
CSIn byte %23.0g LABK The help I get here is really
worth what it costs
CSIo byte %23.0g LABK People here put my needs ahead of
their needs
CSI16 byte %23.0g LABK People here put me down when I
disagree with them
CSI16_R float %9.0g
CSI1q byte %23.0g LABK The biggest help I get here is
learning how to help myself
CSI18 byte %23.0g LABK People here are just trying to get
rid of me
CSI18_R float %9.0g
CSIs byte %23.0g LABK People who know me say this place
has made a positive change in me
CSIt byte %23.0g LABK People here have shown me how to
get help from other places
CSIu byte %23.0g LABK People here seem to understand how
I feel
CSI22 byte %23.0g LABK People here are only concerned
about getting paid
CSI22_R float %9.0g
CSIw byte %23.0g LABK I feel I can really talk to people
here
CSIx byte %23.0g LABK The help I get here is better than
I expected
CSIy byte %8.0g CSI25 I look forward to the sessions I
have with people here
csimean float %9.0g
csi_tot float %9.0g
sci_SF_mean float %9.0g
csi_SF_tot float %9.0g
T1 float %9.0g One month indicator
T2 float %9.0g Two month indicator
T3 float %9.0g Three month indicator
T6 float %9.0g Six month indicator
grp float %12.0g group_labels
Study group
int1 float %9.0g Month one by treatment interaction
int2 float %9.0g Month two by treatment interaction
int3 float %9.0g Month three by treatment
interaction
int6 float %9.0g Month four by treatment
interaction
Tb float %9.0g Baseline PHQ9 total score
employment float %9.0g
antidep float %9.0g YesNo antidepressant use (any time
point)
* indicated variables have notes
--------------------------------------------------------------------------------
Sorted by: TIME
. codebook ageR LEAPS1
--------------------------------------------------------------------------------
ageR Age Group
--------------------------------------------------------------------------------
Type: Numeric (float)
Range: [1,6] Units: 1
Unique values: 6 Missing .: 152/516
Tabulation: Freq. Value
66 1
98 2
78 3
59 4
37 5
26 6
152 .
--------------------------------------------------------------------------------
LEAPS1 What kind of paid work do you do
--------------------------------------------------------------------------------
Type: Numeric (byte)
Label: LEAPS1
Range: [1,27] Units: 1
Unique values: 19 Missing .: 270/516
Examples: 10 Human Resources
21 Other
.
.
. labelbook sex SF369C
--------------------------------------------------------------------------------
Value label SF369C
--------------------------------------------------------------------------------
Values Labels
Range: [-99,6] String length: [11,20]
N: 7 Unique at full length: yes
Gaps: yes Unique at length 12: yes
Missing .*: 0 Null string: no
Leading/trailing blanks: no
Numeric -> numeric: no
Definition
-99 Not Applicable
1 All of the time
2 Most of the time
3 Some of the time
4 A little of the time
5 None of the time
6 No response
Variables: SF369C
--------------------------------------------------------------------------------
Value label sex
--------------------------------------------------------------------------------
Values Labels
Range: [0,1] String length: [4,6]
N: 2 Unique at full length: yes
Gaps: no Unique at length 12: yes
Missing .*: 0 Null string: no
Leading/trailing blanks: no
Numeric -> numeric: no
Definition
0 female
1 male
Variables: CSRI2
We can also add a label to a dataset, which can be seen when we use describe. Be aware, that this will overwrite existing data labels.
. label data "This is our exercise data set"
. label data "This is our data set to play around"
We will talk later on data management more, but here just how to save a dataset: the command save will save it to the current working directory, which you can find out by typing pwd. So, normally, at the beginning of a working session - most often in a do-file - , you would set the working directory with, for example:
F:\Dropbox\000_homepages\pt-wissen\stata.
You can also do this by clicking on FILE–>change working directory… You could also add a path in front of the dataset name, either an absolute path or a relative path.Absolute path: "C:\Windows\calc.exe"
Relative path: "../02_data/file_to_load.dta" (the ../ means that you tell Stata to go one folder up and then enter the 02_data folder, and there load the file_to_load.dta file. )
. pwd
F:\Dropbox\000_homepages\pt-wissen\stata
. save "this_is_our_exercise_dataset.dta", replace
file this_is_our_exercise_dataset.dta saved
The basic anatomy of a command line is: [prefix:] command [varlist] [if] [in] [weight] [, options]
The parts that are put in [] above are optional, you don’t need them always.
Here an example: The prefix bysort: can be used in front of a command to produce a stratified execution of the command. *bysort ageR: sum LEAPS_tot *
. bysort ageR: sum LEAPS_tot
--------------------------------------------------------------------------------
-> ageR = 1
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS_tot | 40 7.9125 7.556528 0 28
--------------------------------------------------------------------------------
-> ageR = 2
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS_tot | 75 5.813333 4.625702 0 21
--------------------------------------------------------------------------------
-> ageR = 3
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS_tot | 60 6.166667 5.90078 0 27
--------------------------------------------------------------------------------
-> ageR = 4
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS_tot | 40 7.725 5.444298 0 22
--------------------------------------------------------------------------------
-> ageR = 5
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS_tot | 8 5.625 4.438066 0 12
--------------------------------------------------------------------------------
-> ageR = 6
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS_tot | 0
--------------------------------------------------------------------------------
-> ageR = .
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS_tot | 0
A similar results - even a bit nicer - could be reached by using the option by instead of the prefix bysort
. tabstat LEAPS_tot, by(ageR) stats(mean, sd, min max, p25, p50, p75, n)
Summary for variables: LEAPS_tot
Group variable: ageR (Age Group)
ageR | Mean SD Min Max p25 p50 p75
---------+----------------------------------------------------------------------
1 | 7.9125 7.556528 0 28 2.5 5 13
2 | 5.813333 4.625702 0 21 2 5 8
3 | 6.166667 5.90078 0 27 2 5 9.5
4 | 7.725 5.444298 0 22 4 6.5 10
5 | 5.625 4.438066 0 12 1 7 8.5
6 | . . . . . . .
---------+----------------------------------------------------------------------
Total | 6.621076 5.751314 0 28 2 5 9
--------------------------------------------------------------------------------
ageR | N
---------+----------
1 | 40
2 | 75
3 | 60
4 | 40
5 | 8
6 | 0
---------+----------
Total | 223
--------------------
Let’s look at the [if] part. Have a special look at the second example, where we combine - with or - two conditions.
. summarize LEAPS_tot if ageR==1
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS_tot | 40 7.9125 7.556528 0 28
. summarize LEAPS_tot if ageR==1 | ageR==2
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS_tot | 115 6.543478 5.867905 0 28
We can also show how the in functions, have a look at the following example (of course it depends on how the data are sorted):
You can sort in ascending order with the command sort. If you want to sort in descending order, you use gsort - (gsort+ = ascending, gsort - = descending ).
. sort idno
. list LEAPS_tot in 1/10
+----------+
| LEAPS_~t |
|----------|
1. | 13 |
2. | . |
3. | 5 |
4. | 7 |
5. | . |
|----------|
6. | . |
7. | . |
8. | . |
9. | 4 |
10. | 12 |
+----------+
or like this:
. sort idno
. count
516
. list LEAPS_tot in 15
+----------+
| LEAPS_~t |
|----------|
15. | . |
+----------+
A negative number indicates minus from the end , the lowercase letters l and f mean last observation (l) and first observation (f)
or like this:
. sort idno
. count
516
. list LEAPS_tot in -10/l
+----------+
| LEAPS_~t |
|----------|
507. | 9 |
508. | 2 |
509. | 6 |
510. | 5 |
511. | 6 |
|----------|
512. | 3 |
513. | 7 |
514. | 8 |
515. | 1 |
516. | 0 |
+----------+
Stata is case sensitive. Variable names can’t start with numbers. A variable can be a maximum of 32 characters. If there is no ambuigity, you don’t need to write the whole variable name, however, if there is ambuigity, Stata will throw an error. You can also put a * as a wildcard for zero, one or several characters.
Look at the example, the variable is called ageR, but its enough to write age. However, because there are several variables that start with LEAPS, the second command would not run (that’s why I added the prefix *capture: * in front of it, so that the programme continues to run despite the error (if not, this html would not be created). The error message (that is now suppresed by the capture: command) would be: LEAP ambiguous abbreviation However, the third line works.
. sum age
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
ageR | 364 2.947802 1.486401 1 6
. capture: sum LEAP // This will not run, error: LEAP ambiguous abbreviation
. sum LEAP*
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
LEAPS1 | 246 14.93902 7.840784 1 27
LEAPS2 | 229 7.262009 2.578827 1 11
LEAPS3 | 226 1.5 1.567021 1 9
LEAPS4A | 223 1.367713 1.056613 0 4
LEAPS4B | 222 1.085586 1.005356 0 4
-------------+---------------------------------------------------------
LEAPS4C | 223 1.278027 1.108449 0 4
LEAPS4D | 223 .8654709 1.039673 0 4
LEAPS4E | 222 .5630631 .8734835 0 4
LEAPS4F | 223 .632287 .8643269 0 4
LEAPS4G | 223 .8206278 1.144562 0 4
-------------+---------------------------------------------------------
LEAPS_M | 223 .945868 .8216163 0 4
LEAPS_tot | 223 6.621076 5.751314 0 28
There are three kind of labels in Stata: the data label, which we saw already above, the value labels, at which we will look later, and the variable labels. In our example dataset, some variables already have labels. You seee the labels either in the Variables window, or you can type describe in the command line.
. describe
Contains data from this_is_our_exercise_dataset.dta
Observations: 516 This is our data set to play
around
Variables: 133 19 Dec 2021 09:46
--------------------------------------------------------------------------------
Variable Storage Display Value
name type format label Variable label
--------------------------------------------------------------------------------
idno int %8.0g
PRACTICE float %9.0g
collectionpoint byte %8.0g COLLECTI Data collection point
ageR float %9.0g Age Group
CSRI2 byte %8.0g sex What is your gender?
CSRI3 byte %8.0g CSRI3 What is your marital status?
CSRI4 byte %8.0g CSRI4 What is the highest degree or
level of school you have
completed?
CSRI5 byte %8.0g CSRI5 What is your total household
income?
CSRI6 byte %8.0g CSRI6 What is your mother tongue?
CSRI7 byte %8.0g CSRI7 Country of Birth?
CSRI8 byte %8.0g CSRI8 Who do you usually live with?
CSRI9 byte %8.0g CSRI9 Current employment status
TIME float %9.0g
PHQ91 byte %8.0g PHQ91 Little interest or pleasure in
doing things
PHQ92 byte %8.0g PHQ92 Feeling down, depressed, or
hopeless
PHQ93 byte %8.0g PHQ93 Trouble falling or staying asleep,
or sleeping too much
PHQ94 byte %8.0g PHQ94 Feeling tired or having little
energy
PHQ95 byte %8.0g PHQ95 Poor appetite or overeating
PHQ96 byte %8.0g PHQ96 * Feeling bad about yourself or that
you are a failure or have let
yourself or you
PHQ97 byte %8.0g PHQ97 * Trouble concentrating on things,
such as reading the newspaper or
watching telev
PHQ98 byte %8.0g PHQ98 * Moving or speaking so slowly that
other poeple could have noticed?
Or the opposi
PHQ99 byte %8.0g PHQ99 Thoughts that you would be better
off dead or of hurting yourself
in some way
PHQ_TOT byte %8.0g
LEAPS1 byte %8.0g LEAPS1 What kind of paid work do you do
LEAPS2 byte %8.0g LEAPS2 hours scheduled or expected to
work
LEAPS3 byte %8.0g LEAPS3 hours of work missed because of
the way you were feeling
LEAPS4A byte %8.0g LEAPS4A low energy or motivation
LEAPS4B byte %8.0g LEAPS4B Poor concentration or memory
LEAPS4C byte %8.0g LEAPS4C Anxiety or irritability
LEAPS4D byte %8.0g LEAPS4D Getting less work done
LEAPS4E byte %8.0g LEAPS4E Doing poor quality work
LEAPS4F byte %8.0g LEAPS4F Making more mistakes
LEAPS4G byte %8.0g LEAPS4G Having trouble getting along with
people, or avioding them
LEAPS_M double %12.0g
LEAPS_tot double %12.0g
SDS1 byte %8.0g SDS1 The symptoms have disrupted your
work/school work:
SDS2 byte %8.0g SDS2 The symptoms have disrupted your
social life/leisure activities:
SDS3 byte %8.0g SDS3 The symptoms have disrupted your
family life/home
responsibilities:
SDS_M double %12.0g
SDSTOT_fixed float %9.0g
SDS4 byte %8.0g SDS4 days last week lost
SDS5 byte %8.0g SDS5 days last week unproductive
SF361 byte %8.0g SF361 In general, would you say your
health is
SF362 byte %8.0g SF362 Compared to one year ago, how
would you rate your health in
general now?
SF363A byte %8.0g SF363A * The following questions are about
activities you might do during a
typical day.
SF363B byte %8.0g SF363B * Moderate activities, such as
moving a table, pushing a vacuum
cleaner, bowling,
SF363C byte %8.0g SF363C Lifting or carrying groceries
SF363D byte %8.0g SF363D Climbing several flights of stairs
SF363E byte %8.0g SF363E Climbing one flight of stairs
SF363F byte %8.0g SF363F Bending, kneeling, or stopping
SF363G byte %8.0g SF363G Walking more than a mile
SF363H byte %8.0g SF363H Walking several hundred yards
SF363I byte %8.0g SF363I Walking one hundred yards
SF363J byte %8.0g SF363J Bathing or dressing yourself
SF364A byte %8.0g SF364A * During the past 4 weeks, how much
of the time have you had any of
the following
SF364B byte %8.0g SF364B Accomplished less than you would
like
SF364C byte %8.0g SF364C Were limited in the kind of work
or other activities
SF364D byte %8.0g SF364D * Had difficulty performing the work
or other activities (for
example, it took ext
SF365A byte %8.0g SF365A * During the past 4 weeks, how much
of the time have you had any of
the following
SF365B byte %8.0g SF365B Accomplished less than you would
like
SF365C byte %8.0g SF365C Did work or other activities less
carefully than usual
SF366 byte %8.0g SF366 * During the past 4 weeks, to what
extent has your physical health
or emotional pr
SF367 byte %8.0g SF367 How much bodily pain have you had
during the past 4 weeks?
SF368 byte %8.0g SF368 * During the past 4 weeks, how much
did pain interfere with your
normal work (incl
SF369A byte %8.0g SF369A * These questions are about how you
feel and how things have been
with you during
SF369B byte %8.0g SF369B Have you been very nervous?
SF369C byte %8.0g SF369C Have you felt so down in the dumps
that nothing could cheer you up?
SF369D byte %8.0g SF369D Have you felt calm and peaceful?
SF369E byte %8.0g SF369E Did you have a lot of energy?
SF369F byte %8.0g SF369F Have you felt downhearted and
depressed?
SF369G byte %8.0g SF369G Did you feel worn out?
SF369H byte %8.0g SF369H Have you been happy?
SF369I byte %8.0g SF369I Did you feel tired?
SF3610 byte %8.0g SF3610 * During the past 4 weeks, how much
of the time has your physical
health or emotio
SF3611A byte %8.0g SF3611A * How true or false is each of the
following statements for you? I
seem to get sic
SF3611B byte %8.0g SF3611B I am as healthy as anybody I know
SF3611C byte %8.0g SF3611C I expect my health to get worse
SF3611D byte %8.0g SF3611D My health is excellent
SF36_PF float %9.0g
SF36_RP float %9.0g
SF36_RE float %9.0g
SF36_VT float %9.0g
SF36_MH float %9.0g
SF36_SF float %9.0g
SF36_BP float %9.0g
SF36_GH float %9.0g
SF36_TOT float %9.0g
CSIa byte %23.0g CSI1 The services I get here are a big
help to me
CSIb byte %23.0g LABK People here really seem to care
about me
CSIc byte %23.0g LABK I would come back here if I need
help again
CSI4 byte %23.0g LABK I feel that no one here really
listens to me
CSI4_R float %9.0g
CSIe byte %23.0g LABK People here treat me like a
person, not like a number
CSIf byte %23.0g LABK I have learned a lot here about
how to deal with my problems
CSI7 byte %23.0g LABK People here want to do things
their way, instead of helping me
find my way
CSI7_R float %9.0g
CSIh byte %23.0g LABK I would recommend this place to
people I care about
CSIi byte %23.0g LABK People here really know what they
are doing
CSIj byte %23.0g LABK I get the kind of help here that I
really need
CSIk byte %23.0g LABK People here accept me for who I am
CSIl byte %23.0g LABK I feel much better now than when I
first came here
CSIm byte %23.0g LABK I thought no one could help me
until I came here
CSIn byte %23.0g LABK The help I get here is really
worth what it costs
CSIo byte %23.0g LABK People here put my needs ahead of
their needs
CSI16 byte %23.0g LABK People here put me down when I
disagree with them
CSI16_R float %9.0g
CSI1q byte %23.0g LABK The biggest help I get here is
learning how to help myself
CSI18 byte %23.0g LABK People here are just trying to get
rid of me
CSI18_R float %9.0g
CSIs byte %23.0g LABK People who know me say this place
has made a positive change in me
CSIt byte %23.0g LABK People here have shown me how to
get help from other places
CSIu byte %23.0g LABK People here seem to understand how
I feel
CSI22 byte %23.0g LABK People here are only concerned
about getting paid
CSI22_R float %9.0g
CSIw byte %23.0g LABK I feel I can really talk to people
here
CSIx byte %23.0g LABK The help I get here is better than
I expected
CSIy byte %8.0g CSI25 I look forward to the sessions I
have with people here
csimean float %9.0g
csi_tot float %9.0g
sci_SF_mean float %9.0g
csi_SF_tot float %9.0g
T1 float %9.0g One month indicator
T2 float %9.0g Two month indicator
T3 float %9.0g Three month indicator
T6 float %9.0g Six month indicator
grp float %12.0g group_labels
Study group
int1 float %9.0g Month one by treatment interaction
int2 float %9.0g Month two by treatment interaction
int3 float %9.0g Month three by treatment
interaction
int6 float %9.0g Month four by treatment
interaction
Tb float %9.0g Baseline PHQ9 total score
employment float %9.0g
antidep float %9.0g YesNo antidepressant use (any time
point)
* indicated variables have notes
--------------------------------------------------------------------------------
Sorted by: idno
Maybe we take some minutes to skim through the help file for the labels. So let’s type help labels in the command line. This will open the viewer window with the help for labels. This is a shorter version of the help for labels. If you want a longer version, click on the (View complete PDF manual entry) at the second line of the help file.
. help labels
In the help file we saw that it is quite easy to label a variable. So let’s label the variable idno, which does not have a label yet.
. label variable idno "Identification Number of Participants"
. describe idno
Variable Storage Display Value
name type format label Variable label
--------------------------------------------------------------------------------
idno int %8.0g Identification Number of
Participants
Can we change a variable label? Let’s try it: (works fine).
. label variable idno "Identification Code of Participants"
. describe idno
Variable Storage Display Value
name type format label Variable label
--------------------------------------------------------------------------------
idno int %8.0g Identification Code of
Participants
Maybe its time to look at different data types. The best way to learn this is with the PDF manual. We can easily access this chapter via the help file of data types. There, we will click on the second line (View complete PDF manual entry), and in the PDF wi will cklick on See [U] 12 Data for details.
We can also visit this same document online by clicking here Of course I will not repeat here what we can read in the help chapter. So see here again after reading the chapter 12 on Data.
. help data types
We should also read the chapters 24 * working with strings* and 26 working with categorical data and factor variables of the manual. We save the chapter 25 working with dates and times for later.
Now, let’s do an exercise. We will create a new variable called sex. We will first create this as a string variable with the content women and men and other. This way of storing the information on sex is not efficient, we would better store it as a categorical numerical variable. Therefore, we create from the string variable a categorical numerical variable.
. tab CSRI2
What is |
your |
gender? | Freq. Percent Cum.
------------+-----------------------------------
female | 264 72.53 72.53
male | 100 27.47 100.00
------------+-----------------------------------
Total | 364 100.00
. tab CSRI2, nol
What is |
your |
gender? | Freq. Percent Cum.
------------+-----------------------------------
0 | 264 72.53 72.53
1 | 100 27.47 100.00
------------+-----------------------------------
Total | 364 100.00
. gen sex_string=""
(516 missing values generated)
. replace sex_string="Men" if CSRI2==1
variable sex_string was str1 now str3
(100 real changes made)
. replace sex_string="Women" if CSRI2==0
variable sex_string was str3 now str5
(264 real changes made)
. tab sex_string CSRI2
| What is your gender?
sex_string | female male | Total
-----------+----------------------+----------
Men | 0 100 | 100
Women | 264 0 | 264
-----------+----------------------+----------
Total | 264 100 | 364
As I wrote above, to store data like this is not the best way, so most often we would want to create a numeric variable and add value labels to each different value.
. encode sex_string, gen(sex_num)
. tab sex_string sex_num
| sex_num
sex_string | Men Women | Total
-----------+----------------------+----------
Men | 100 0 | 100
Women | 0 264 | 264
-----------+----------------------+----------
Total | 100 264 | 364
. tab sex_num
sex_num | Freq. Percent Cum.
------------+-----------------------------------
Men | 100 27.47 27.47
Women | 264 72.53 100.00
------------+-----------------------------------
Total | 364 100.00
. tab sex_num, nol
sex_num | Freq. Percent Cum.
------------+-----------------------------------
1 | 100 27.47 27.47
2 | 264 72.53 100.00
------------+-----------------------------------
Total | 364 100.00
. recode sex_num (1=0) (2=1)
(364 changes made to sex_num)
. tab sex_num
sex_num | Freq. Percent Cum.
------------+-----------------------------------
0 | 100 27.47 27.47
Men | 264 72.53 100.00
------------+-----------------------------------
Total | 364 100.00
. labelbook sex_num
--------------------------------------------------------------------------------
Value label sex_num
--------------------------------------------------------------------------------
Values Labels
Range: [1,2] String length: [3,5]
N: 2 Unique at full length: yes
Gaps: no Unique at length 12: yes
Missing .*: 0 Null string: no
Leading/trailing blanks: no
Numeric -> numeric: no
Definition
1 Men
2 Women
Variables: sex_num
. label define sex_num 1"Women" 0"Men", modify
. label value sex_num sex_num
. tab sex_num CSRI2
| What is your gender?
sex_num | female male | Total
-----------+----------------------+----------
Men | 0 100 | 100
Women | 264 0 | 264
-----------+----------------------+----------
Total | 264 100 | 364
.
In our data, we have only two values for sex. But even there is no other value, we could add a label for potential other values, such as “other”. We could overwrite the existing label (which is named sex_num), or we could create a new label - let’s call it sex_WomenMenOther.
. label define sex_WomenMenOther 0"Men" 1"Women" 2"Other"
. label value sex_num sex_WomenMenOther
. tab sex_num, mi
sex_num | Freq. Percent Cum.
------------+-----------------------------------
Men | 100 19.38 19.38
Women | 264 51.16 70.54
. | 152 29.46 100.00
------------+-----------------------------------
Total | 516 100.00
Please read also the PDF manual entry for encode and the help file for recode.
We use the new table function that are available since the version 17. After the table command, you put two parentheses, the first for the rows and the second for the columns. Look at this example (the second variant is witouth the totals column):
. table (ageR)(CSRI2)
-------------------------------------
| What is your gender?
| female male Total
----------+--------------------------
Age Group |
1 | 53 13 66
2 | 72 26 98
3 | 55 23 78
4 | 36 23 59
5 | 27 10 37
6 | 21 5 26
Total | 264 100 364
-------------------------------------
. table (ageR)(CSRI2), nototals
-----------------------------------
| What is your gender?
| female male
----------+------------------------
Age Group |
1 | 53 13
2 | 72 26
3 | 55 23
4 | 36 23
5 | 27 10
6 | 21 5
-----------------------------------
We can also create nested tables.
. table (ageR, CSRI2) (collectionpoint)
-------------------------------------------------------------------------
| Data collection point
| 1 Month 2 Month 3 Month 6 Month Total
-------------------------+-----------------------------------------------
Age Group |
1 |
What is your gender? |
female | 16 12 12 13 53
male | 4 3 3 3 13
Total | 20 15 15 16 66
2 |
What is your gender? |
female | 18 17 16 21 72
male | 6 5 7 8 26
Total | 24 22 23 29 98
3 |
What is your gender? |
female | 16 12 13 14 55
male | 6 7 6 4 23
Total | 22 19 19 18 78
4 |
What is your gender? |
female | 10 9 9 8 36
male | 7 5 5 6 23
Total | 17 14 14 14 59
5 |
What is your gender? |
female | 8 7 6 6 27
male | 3 3 2 2 10
Total | 11 10 8 8 37
6 |
What is your gender? |
female | 5 4 6 6 21
male | 1 1 1 2 5
Total | 6 5 7 8 26
Total |
What is your gender? |
female | 73 61 62 68 264
male | 27 24 24 25 100
Total | 100 85 86 93 364
-------------------------------------------------------------------------
Let’s add the percentages, too
. table (ageR, CSRI2) (collectionpoint),statistic(frequency) statistic(percent)
--------------------------------------------------------------------------
| Data collection point
| 1 Month 2 Month 3 Month 6 Month Total
-------------------------+------------------------------------------------
Age Group |
1 |
What is your gender? |
female |
Frequency | 16 12 12 13 53
Percent | 4.40 3.30 3.30 3.57 14.56
male |
Frequency | 4 3 3 3 13
Percent | 1.10 0.82 0.82 0.82 3.57
Total |
Frequency | 20 15 15 16 66
Percent | 5.49 4.12 4.12 4.40 18.13
2 |
What is your gender? |
female |
Frequency | 18 17 16 21 72
Percent | 4.95 4.67 4.40 5.77 19.78
male |
Frequency | 6 5 7 8 26
Percent | 1.65 1.37 1.92 2.20 7.14
Total |
Frequency | 24 22 23 29 98
Percent | 6.59 6.04 6.32 7.97 26.92
3 |
What is your gender? |
female |
Frequency | 16 12 13 14 55
Percent | 4.40 3.30 3.57 3.85 15.11
male |
Frequency | 6 7 6 4 23
Percent | 1.65 1.92 1.65 1.10 6.32
Total |
Frequency | 22 19 19 18 78
Percent | 6.04 5.22 5.22 4.95 21.43
4 |
What is your gender? |
female |
Frequency | 10 9 9 8 36
Percent | 2.75 2.47 2.47 2.20 9.89
male |
Frequency | 7 5 5 6 23
Percent | 1.92 1.37 1.37 1.65 6.32
Total |
Frequency | 17 14 14 14 59
Percent | 4.67 3.85 3.85 3.85 16.21
5 |
What is your gender? |
female |
Frequency | 8 7 6 6 27
Percent | 2.20 1.92 1.65 1.65 7.42
male |
Frequency | 3 3 2 2 10
Percent | 0.82 0.82 0.55 0.55 2.75
Total |
Frequency | 11 10 8 8 37
Percent | 3.02 2.75 2.20 2.20 10.16
6 |
What is your gender? |
female |
Frequency | 5 4 6 6 21
Percent | 1.37 1.10 1.65 1.65 5.77
male |
Frequency | 1 1 1 2 5
Percent | 0.27 0.27 0.27 0.55 1.37
Total |
Frequency | 6 5 7 8 26
Percent | 1.65 1.37 1.92 2.20 7.14
Total |
What is your gender? |
female |
Frequency | 73 61 62 68 264
Percent | 20.05 16.76 17.03 18.68 72.53
male |
Frequency | 27 24 24 25 100
Percent | 7.42 6.59 6.59 6.87 27.47
Total |
Frequency | 100 85 86 93 364
Percent | 27.47 23.35 23.63 25.55 100.00
--------------------------------------------------------------------------
An excellent video on how to build tables in Stata can be found here, there is also a [second part for two-way tables]https://youtu.be/u_Efw1oWxWk).
You might also find these webpages interesting:
Stata 17, part2, the new collect command for tables
If we create a numeric variables, the standard type of this variable is float. You could change this by useing *set type double * - check the standad settings with query. Maybe you want to read this webpage
. gen numeric_float_variable=rnormal(50, 12)
. describe numeric_float_variable
Variable Storage Display Value
name type format label Variable label
--------------------------------------------------------------------------------
numeric_float~e float %9.0g
. sum numeric_float_variable
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
numeric_fl~e | 516 50.38684 11.60436 11.99885 92.54129
Sometimes, the type of the variable can cause some troubles. In the following example, a comparison of a float number is problematic:
. generate x = 1.1
. count if x==1.1
0
We would expect that the command count if would find observations, because we created a lot of them. In the next code-sections you see two solutions:
<<dd_-do>>
count if x==float(1.1)
<</dd_do>>
Here we specified that the comparison is made with “float” precision. In the next solution, we create the variable as double (which has more precision).
. gen double y=1.1
. describe y
Variable Storage Display Value
name type format label Variable label
--------------------------------------------------------------------------------
y double %10.0g
. count if y==1.1
516
You find the explanation here.
When you will calculate with dates, it will be important to use the double format see here.
Back to replacing values of variables. Often you generate an empty variable and then you replace it based on conditions.
. tab CSRI2
What is |
your |
gender? | Freq. Percent Cum.
------------+-----------------------------------
female | 264 72.53 72.53
male | 100 27.47 100.00
------------+-----------------------------------
Total | 364 100.00
. tab CSRI2, nol
What is |
your |
gender? | Freq. Percent Cum.
------------+-----------------------------------
0 | 264 72.53 72.53
1 | 100 27.47 100.00
------------+-----------------------------------
Total | 364 100.00
. capture: drop sex
. gen employed_women=.
(516 missing values generated)
. replace employed_women = 1 if CSRI2 == 0 & employment == 1
(119 real changes made)
. replace employed_women = 0 if CSRI2 == 1 & employment == 0
(47 real changes made)
. label define empl_wom 0 "Unemployed (Men or Women)" 1"Employed Women"
. label values employed_women empl_wom
. tab employed_women
employed_women | Freq. Percent Cum.
--------------------------+-----------------------------------
Unemployed (Men or Women) | 47 28.31 28.31
Employed Women | 119 71.69 100.00
--------------------------+-----------------------------------
Total | 166 100.00
We will often use extended gen functions, as implemented in egen. For example if we want to calculate the number of missing values in a series of variables.
. egen missing_LEAPS = rowmiss(LEAPS1 - LEAPS4G)
. tab missing_LEAPS
missing_LEA |
PS | Freq. Percent Cum.
------------+-----------------------------------
0 | 217 42.05 42.05
1 | 4 0.78 42.83
2 | 2 0.39 43.22
7 | 6 1.16 44.38
8 | 2 0.39 44.77
9 | 16 3.10 47.87
10 | 269 52.13 100.00
------------+-----------------------------------
Total | 516 100.00
There are many other egen functions, so check out the help file help egen.
For those who worked with older Stata versions, Stata added new features, see here. We will now save the current dataset and load an exemple with dates and times.
. save exercise.dta, replace
file exercise.dta saved
. // import delimited dates_and_times.csv, clear // if you have saved it locally
>
. use http://pt-wissen.ch/stata/dates_and_times.dta, clear
The first date is saved as a string variable, so we first need to transform it:
. gen tagmonatjahr=date( Tag_Monat_Jahr, "DMY")
Open the data in the browser and have a look at what we have created. These are the number of days since January first 1960. Now we want to transform this into something more readable.
. format tagmonatjahr %td
Now we can read it.
For the variable with the time dmy_hms, we need to create a double type variable:
. gen double tagmonatjahr_stundeminutesekunde=clock(dmy_hms, "DMYhms")
. format tagmonatjahr_stundeminutesekunde %tc
. list tagmonatjahr tagmonatjahr_stundeminutesekunde
+--------------------------------+
| tagmona~r tagmonatjahr_stu~e |
|--------------------------------|
1. | 28jan2020 28jan2020 17:55:05 |
2. | 03feb2019 03feb2019 09:30:10 |
3. | 01dec1987 01dec1987 20:30:10 |
4. | 08sep1821 08sep1821 10:45:20 |
+--------------------------------+
. gen date_of_birth = date(dob, "DMY")
. format date_of_birth %td
. gen date_of_death = date(dod, "DMY")
(2 missing values generated)
. format date_of_death %td
. list dob dod date_of_birth date_of_death
+-------------------------------------------------+
| dob dod date_~rth date_~ath |
|-------------------------------------------------|
1. | 18.06.1942 18jun1942 . |
2. | 09.10.1940 08.12.1980 09oct1940 08dec1980 |
3. | 25.02.1943 29.11.2001 25feb1943 29nov2001 |
4. | 7.7.1940 07jul1940 . |
+-------------------------------------------------+
Calculating the time between two dates is easy :
. gen age_years_at_death_v1 = (date_of_death - date_of_birth)/365.25
(2 missing values generated)
Since Stata 17 there is a new way to do this (version 4 in the code below).
. gen age_years_at_death_v2=age(date_of_birth, date_of_death)
(2 missing values generated)
. gen age_years_at_death_v3=age_frac(date_of_birth, date_of_death)
(2 missing values generated)
. gen age_years_at_death_v4=datediff(date_of_birth, date_of_death, "year")
(2 missing values generated)
Let’s do some exercises: calculate the dates for the other variables, and then calculate the time between the white album and the rooftop concert, as well as the duration of the root top concert.
First, we read in some data
<dd_do>>
input id str10 Tag_Monat_Jahr str10 Monat_Tag_Jahr str16 dmy_hms str10 dob str10 dod str10 date_event str10 date_release_white_album str16 start_rooftop_concert str165 end_rooftop_concert
1 28.1.2020 1.28.2020 "28:01:2020:17:55:05" 18.06.1942 NA 3.01.1962 22.11.1968 "30.01.1969.12:30" "30.01.1969.13:12"
2 03.2.2019 2.3.2019 "03.2.2019:09:30:10" 09.10.1940 08.12.1980 3.01.1962 22.11.1968 "30.01.1969.12:30" "30.01.1969.13:12"
3 01.12.1987 12.01.1987 "01/12/1987/20:30:10" 25.02.1943 29.11.2001 3.01.1962 22.11.1968 "30.01.1969.12:30" "30.01.1969.13:12"
4 8.9.1821 9.8.1821 "08:09:1821:10:45:20" 7.7.1940 NA 3.01.1962 22.11.1968 "30.01.1969.12:30" "30.01.1969.13:12"
end
<</dd_do>
The data are in string format, so we need to transform it to Stata date (if only date and not time in the string) or datetime (if date and time is present in the string variable). First, the string with only date. The first line transforms it to a number which contains the days since 01jan1960
<<dd_do>
gen tagmonatjahr=date( Tag_Monat_Jahr, "DMY")
list tagmonatjahr
<</dd_do>
The second line makes this readable for us.
<<dd_do>
format tagmonatjahr %td
list tagmonatjahr
<</dd_do>
If the string variables contains time, we need to generate a double variable.
<<dd_do>
gen double tagmonatjahr_stundeminutesekunde=clock(dmy_hms, "DMYhms")
format tagmonatjahr_stundeminutesekunde %tc
<</dd_do>
The same for the remaining variables:
<<dd_do>
gen date_of_birth = date(dob, "DMY")
format date_of_birth %td
gen date_of_death = date(dod, "DMY")
format date_of_death %td
list dob dod date_of_birth date_of_death
<</dd_do>
There are different methods to calculate the years between two timepoints, here to calculate the age at death. The first method does not take into account the differences in lengths of specials years, it just assumes that every year is 365.25 days.
<<dd_do>
gen age_years_at_death_v1 = (date_of_death - date_of_birth)/365.25
<</dd_do>
The next method calculates the years - rounded to an integer:
<<dd_do>
gen age_years_at_death_v2=age(date_of_birth, date_of_death)
<</dd_do>
The next method calculates years with fractions of a year added:
<<dd_do>
gen age_years_at_death_v3=age_frac(date_of_birth, date_of_death)
<</dd_do>
Another method to calcuate the time between two timepoints:
<<dd_do>
gen age_years_at_death_v4=datediff(date_of_birth, date_of_death, "year")
<</dd_do>
Now we want to calculate the time between the release of the White Album and the start of the rooftoop concert. We will run into a problem, because the two dates are not in the same format, the release of the white album is a date (i.e. no time included), and the start of the rooftop concert contains time.
<<dd_do>
gen date_release_white_album_date = date(date_release_white_album, "DMY")
format date_release_white_album_date %td
gen start_rooftop_concert_dt = clock(start_rooftop_concert, "DMYhm")
format start_rooftop_concert_dt %tc
<</dd_do>
Therefore, we need to convert the data, see here: https://www.stata.com/manuals13/ddatetime.pdf#ddatetimeSyntaxSIF-to-SIFconversion We decide to convert the date variable date_release_white_album to a clock variable (date and time included), so this will ad a time to this variable.
<<dd_do>
gen date_release_white_album_dt = cofd(date_release_white_album_date)
format date_release_white_album_dt %tc
list date_release_white_album_dt start_rooftop_concert_dt
<</dd_do>
Now, after we have converted the variable, we can calculate the duration in days between the two variables with clockdiff. If we would have converted the variable to date instead of clock, we would have to use datediff instead of clockdiff.
<<dd_do>
gen days_between_white_album_concert=clockdiff(date_release_white_album_dt, start_rooftop_concert_dt, "day")
list days_between_white_album_concert date_release_white_album_dt start_rooftop_concert_dt
<</dd_do>
Just to give some context to this exercise:A song from the white album and here an extract of the rooftop concert.
Until now we worked in the command editor. Normally, we would like to make every step reproducible, therefore, we want to work in so called do-files. Stata has its own do-editor, but you can also work in any other editor you prefer. Please check the new features in the do editor for Stata 17.
Please check out this video on new functions in the do-editor and see here a video on the new bookmark function in the Stata do-editor.
Falls Sie lieber ein Video in Deutsch möchten, finden Sie hier eines.
In smaller projects, it makes sense to put all code within one do-file. In larger projects, I create one project folder that contains several subfolders, amongst them a folder for the scripts. In this script folder, I have several do-files. One large do-file has the advantage that you exactly know where a specific task was done - namely in this very do-file. You can structure your do-file with sections and you can search it easily. However, it still can be overwhelming to have all the code in one and I preferer to split the do-files up by tasks. 
The master do-file is the one do-file you let run in the end. The master do-file will invoke - with the command include all the other do-files . 