Not directly, but you can export your data as a csv file and then import the delimited text file.

Ekeoma, if you interact two dummy variables you produce four outcome groups. You can use Stata's factor notation to do this in different types of models. See help fvvarlist for detail. Here is a quick (and analytically meaningless) example that creates a dummy variable and then uses it in a regression model interacted with another dummy variable. I have included a covariate as well. The plotted results show the relationship between displacement and fuel efficiency across the four groups defined by the interaction term. Hopefully, this gives you some ideas on how to set up your model. sysuse auto, clear * Create dummy variable for this example sum weight, meanonly gen weightdum=weight>=r(mean) * This is the interaction to be estimated tab weightdum foreign, nolabel * Use regress with factor notation (see help fvvarlist) regress mpg displacement i.weightdum##i.foreign margins weightdum#foreign, at(displacement=(75(25)200)) marginsplot, noci scheme(s2color)

Hi Beatrice, I am uncertain exactly what you want to accomplish but I have suggestions for two situations that may help you. The first situation has a typical dataset of recorded values for each observation for several variables. In this situation, you might want to graph the results of some simple descriptive statistics (e.g. means, medians, standard deviations, etc.) In the second situation, you could have a dataset of descriptive statistics (e.g. means, etc.) for variables calculated across some other (string) variable. People are sometimes confused by "graph bar" in Stata expecting something like a histogram. But the help file clearly states, "In a vertical bar chart, the y axis is numerical, and the x axis is categorical." My example uses the auto.dta dataset to demonstrate graphing the means and medians of three variables calculated for two groups. Here is example 1: sysuse auto, clear summarize mpg trunk turn graph bar (mean) mpg turn trunk /// (median) mpg turn trunk, /// over(foreign) blabel(bar, format(%4.0f)) /// name(bar1, replace) I believe this accomplishes graphing means and medians for variables across or over a categorical variable. But, if your dataset contains summary statistics, you can create the same graph shown in example 2: clear input str10 foreign byte (mpgmean mpgmed turnmean turnmed trunkmean trunkmed) "Domestic" 20 19 41 42 15 16 "Foreign" 25 25 35 36 11 11 end graph bar mpgmean turnmean trunkmean mpgmed turnmed trunkmed, /// over(foreign) blabel(bar) name(bar2, replace) Note here that I have simply created a new dataset of summary statistics, an unnecessary step if I have access to the original raw data (e.g. example 1). Hopefully, this helps you a bit. If I have misunderstood your question, feel free to post again. It would be helpful if you could use a built-in s=dataset like the auto dataset to show me an analogous problem that we can work on together.

Thank you!

Hi Koya. Unfortunately, importing data to any statistical applicaiton can be complicated due to the numerous formats that data are stored in. I suggest that you type "help import" in the command window and read the contents to see if there is an east way t import your data. If you know that you need to use infix, you can type "help infix" for detailed information.

Hi Bugz, Stata has a command that allows you to rename a group of variables. I would rename variables using this command and not use any looping constructs. Here is an example: sysuse auto, clear rename (make price mpg rep78 headroom trunk /// weight length turn displacement gear_ratio foreign) /// (v1 v2 v3 v4 v5 v6 v7 v8 v9 v10 v11 v12) You can read about the command by entering help rename group in the command window. You don't say much about how you want to construct your dummy variables but I can give you two examples. The first uses numeric variables as inputs and divides them at their means, assigning 0 if a value is less than or equal to the mean and a 1 if it is greater than the mean and not missing. I put the variable generation in a loop. Here is the example: sysuse auto, clear foreach var of varlist price mpg weight { summarize `var', meanonly generate `var'dum = `var' >= r(mean) & `var' <. } list p* mp* w* But, if I had a discrete variable like rep78 I would use the tabulate command. Something like this: sysuse auto, clear tab rep78, generate(rep78dum_) tab foreign, generate(foreigndum_)

🤩thank you so much for the very clear and extensive answer. I have several categories in the dummy. So as your suggestion it is the tabulate command I used. gen v114=waters_dummyd tabulate waters_dummyd, gen(water_srca) gen imp_water_dd=1 if water_srca1==1 replace imp_water_dd=1 if water_srca2==0 For this, if there are categories from 1 - 20 I need to retype it 20 times and change the 1 or 0. I was wondering if I should do this way manually or if is there a macro to make this in a few lines.

​@@bugzdesilva8216 I am confused by your code. I belive that your first two lines of code can be replaced by a single command like this: tabulate v114, gen(water_srca) You do not have to create a new variable to tabulate. Here is how this works with the auto.dta dataset: sysuse auto, clear tab rep78, generate(rep78dum_) Because the variable rep78 has 5 categories, a total of 5 dummy variables are created. The variables form a linear combination so something like a regression model you would need to decide on a category to exclude. Nota bene: if using indicator variables in some model, you really should use Stata's factor notation (see help fvvarlist). Your next two commands are confusing to me because I do not know how the original variable, v114, was coded. It looks like you are trying to take two conditions and code them equal to 1 and all other equal to zero. Here is an example, again, using rep78: sort rep78 generate rep78dum = cond(1, rep78==1 | rep78==5, 0) if rep78 <. list rep78 rep78dum Here, the new variable called rep78dum is assigned the value of 1 when rep78 is equal to either 1 or 5. All other values are set to 0 except for the cases where rep78 has missing values. There are other ways of doing this, but this is one way to accomplish this with one line of code. It would be easier for me to help you if you created examples using the auto.dta dataset so I can replicate them. It is difficult to answer questions when I do not know how the data are represented.

@@smilex3 Thank you so much !!! The last example you provided was really helpful to me. It solved the problem I faced. I was generating a binary variable assigning 1 or 0 per category to the main variable (ie: rep78). Your command made my dofile clear and smooth. Glad I'm following your channel !!

@smilex3 And one more clarification please. if we want to generate a categorical variable by using the rep78 variable, giving 1 if rep78==1 and 2 if rep78==2 and 3 if rep78==3 and 0 for both rep78==4 and rep78==5 how could we generate the above categories using the same command. Is it possible.. *sort rep78 generate rep78dum = cond(1, rep78==1 | rep78==5, 0) if rep78 <.

I thnk that I said this correctly. Unfortiunately, itis easy to get lost ina lot of similar terms. For example, in the context of ANOVA, the coefficient of determination is equal to the between group sum of squares divided by the total sum of squares. Inthe context of regression, the between group sum of squares is called the regression or model sum of squares. So, the ratio of the between group, model, or regression sums of squares to the total sum of squares is the coefficient of determination. In ANOVA you might also see this called eta-squared where it is used to decompose the explained variance due to different independent variables and it is used as an effect size measure.

There are two options that can help control the look of the bars. Which one you use depends on what you want your graph to look like and how your variables were created. The options are: ascategory which treat yvars as first over() group and asyvars which treat the first over() group as yvars In the example for the last graph in the video, I have one variable to be graphed over two categories. The default option is ascategory. Copy and run this code snippet: sysuse auto, clear set scheme s2color graph bar mpg if rep78>=3, over(foreign) over(rep78) ascategory name(bar1, replace) graph bar mpg if rep78>=3, over(foreign) over(rep78) asyvars name(bar2, replace) But, what if I had a different set of variables for mpg? One of the variables contains the mpg values for domestic cars and is missing for the foreign ones. The other variable contains mpg values for the foreign cars and is missing for the domestic ones. We can create that variable and graph them like this: separate mpg, by(foreign) generate(mpgfor) graph bar mpgfor0 mpgfor1 if rep78>=3, over(rep78) name(bar3, replace) See how the graphs bar2 and bar3 are identical. But, using the asyvars option I did not need to create and manage any new variables. Regarding your second question, changing the color of a single bar is easy if you only have one over grouping variable. The following code shows the easy solution in the first block of code. The second block shows how there might be an issue with more than one over grouping variable: graph bar mpg if rep78>=3, asyvars bargap(80) over(rep78) /// bar(1, bcolor(navy)) /// bar(2, bcolor(navy)) /// bar(3, bcolor(lime)) /// name(bar4, replace) graph bar mpg if rep78>=3, asyvars bargap(80) over(rep78) over(foreign) /// bar(1, bcolor(navy)) /// bar(2, bcolor(navy)) /// bar(3, bcolor(lime)) /// name(bar5, replace) Here is one more possible solution if you have more than one over group. It involves taking the two over variables and combining them into a single over variable, using the asyvars option, and controlling the colors of the bars individually. Here is an example: egen forrep78=group(foreign rep78), label graph bar mpg if rep78>=3, asyvar over(forrep78) bargap(80) /// bar(1, bcolor(navy)) /// bar(2, bcolor(navy)) /// bar(3, bcolor(lime)) /// bar(4, bcolor(navy)) /// bar(5, bcolor(navy)) /// bar(6, bcolor(navy)) /// name(bar6, replace) If none of the above gets you to where you want to be, you can move to graph twoway bar where you can layer one graph over another. So, youcould create a base graph with the default colors, then add a second graph with just a single bar of a different color.

Yes, that is what is being tested. I don't find this particularly interesting since 0 is often outside the range of my data or is a theoretically uninteresting number. But, if you divide each margin by its standard error you get the z-statistic. Here is some Stata code to play with: sysuse auto, clear regress mpg weight margins, at(weight=(1760(20)4840)) marginsplot, recast(line) noci name(linear1, replace) margins, at(weight=(1760(20)4840)) dydx(weight) marginsplot, recast(line) noci name(linear2, replace) regress mpg c.weight##c.weight margins, at(weight=(1760(20)4840)) marginsplot, recast(line) noci name(inter1, replace) margins, at(weight=(1760(20)4840)) dydx(weight) marginsplot, recast(line) noci name(inter2, replace) The first block of code shows a linear relationship and produces the adjusted predicted values as well as the slope derivatives. Notice that in the case of the derivative, both z and p are the same as in the regression model. This does not hold for interaction terms. The second block of code shows quadratic models for both the adjusted predicted values as well as the slope derivatives. Here you can see that the effect size becomes smaller as weight incrases.

This is one of those it depends situations. If you type help infile in the Stata command window and then click on infile2, you will find all of the details. But here is the short version. There are two ways of using infile, one for free-format files and one for fixed--format files. When you use use fixed-format data files you need a data dictionary that maps the location of variables and observations for Stata. This could mean that you have three files. File #1 are the data. File #2 is the data dictionary. File #3 is optional because you could run the infile command from the command window, ut would typically be a Stata do-file with the infile command. In this scenario you wold run or execute the do-file, The infile command would point Stata to your data dictionary, and the data dictionary would either contain the data or would point Stata to the file with the fixed-format data.

Hi Beakal, the predict produces results base don the model you estimated . So, if you have included your time variable, e.g. year, the values of this variable will influence the predicted values.

I don't think that this can be done without writing your own command to create a contingency table and the associate statistics that you are interested in displaying (e.g. chi2, p-value, etc.). There may be a way to do this using the new table command in Stata V17, but I don't see it after a quick look. While not what you are looking for, the following example use the (new) table command with the tabulate command to produce a table that is maybe a little better looking than the default one in tabulate. Not what you wanted, but maybe you will find it interesting: sysuse nlsw88.dta, clear quietly { table (collgrad) (race) () if race<3, /// nototals /// statistic(percent race, across(collgrad)) /// nformat(%3.1f) collect style header race, title(hide) collect style header collgrad, title(hide) noisily collect preview tab collgrad race if race<3, chi2 local df=(r(r)-1) * (r(c)-1) local chi2=strofreal(r(chi2), "%6.1fc") local pr=strofreal(r(p), "%6.4f") noisily di "Pearson chi2(`df') = `chi2' Pr = `pr' " }

@@smilex3 thank you for your prompt response, I have stata 14.2 . I tried to run this code Its saying option nototals not allowed, Then I tried without nototals. Now it is saying option statistic() is not allowed. I want to learn writing own command. If you can suggest me any link or book, that will be helpful.

Stan Rogers - White Collar Holler (kzbin.info/www/bejne/qKSnnKCMpJeZnZY)

👍

Hi Faisal, I am not certain I understand exactly what you want to do, but I show you some Stata cod below that may point you in a good direction. One way of reading your question is that you want to produce two kernel density plote to compare distributions across time. I show two solutions below. The first uses the addplot() notion and the second uses graph twoway density with the by() option. But, if you really want to plot the density data from one year against the other, I show a third solution using kdensity to generate new variables with the plotted data values that you can then use with any Stata graphing command. Here, I use graph twoway scatter. Hopefully something below helps you: use www.stata-press.com/data/r17/nlswork.dta kdensity hours if year==87, addplot(kdensity hours if year==68) scheme(s2mono) kdensity hours if year==87, addplot(kdensity hours if year==68) scheme(s2mono) two kdensity hours if year==68 | year==87, by(year) kdensity hours if year==68, generate(x68 d68) kdensity hours if year==87, generate(x87 d87) graph twoway scatter d68 d87

​@@smilex3 Thank you very much for your reply. Unfortunately, I could not make you understand what I wanted to do. My problem is; assume, I have income data of 80 individuals in year 1996 and 2002 and I want to analyze the evolution of distribution between 1996 to 2002. I want to do this to see whether their earning differential has reduced or increased or stayed the same in this period. I wrongly told you earlier that I want to do bivariate kernel density analysis. I have done some readings later and I came to know that i have to perform stochastic Kernal analysis instead of bivariate kernel density estimation. Stochastic Kernal estimation will give me a three dimensional graph with year 1996 on x-axis and year 2002 on y-axis and density in third dimension. Normally, this is accompanied with the contour plot as well. Please guide me how this can be implemented in Stata. I will be grateful to you.

Hmmm...I never noticed that before. I looked through the help file but a deeper dive in the pdf manuals might explain the difference. Also, a call to Stata support would probably get you an answer to your interesting question. At first I thought that the default settings might be different across the two plots, but that does not appear to be the case. I created two kernel density plots with the same settings and while quite similar, there are some differences. Here is the Stata code I used to test this idea: sysuse auto, clear kdensity mpg, bwidth(1.9746) kernel(epanechnikov) n(300) /// title("Default kdensity plot") /// scheme(s2mono) /// name(kdens, replace) /// nodraw tw kdensity mpg, bwidth(1.9746) kernel(epanechnikov) n(300) /// title("Default twoway kdensity plot") /// xtitle("Mileage (mpg)") /// note("kernel = epanechnikov, bandwidth = 1.9746") /// scheme(s2mono) /// name(kdenstw, replace) /// nodraw graph combine kdens kdenstw, nocopies ycommon xcommon scheme(s2mono) Looking at the pdf help files I found this "graph twoway kdensity varname uses the kdensity command to obtain an estimate of the density of varname and uses graph twoway line to plot the result." So, it is possible that how the line is produced is different and leads to this issue. I thought this could be tracked down by generating the density data points but this is not an option in twoway kdensity. So, it might just be about how the plot is produced using tw line versus om other method.

@@smilex3 Thank you so much for an answer lots of care and attempt! I just like to spot some details that way, which slows me down on STATA sometimes ^^! But I guess "twoway kdensity" is still the best way to have the 1st look at a variable.

In a situation where you have may independent control variables but your main interest is on the varuabke degree, you may want to do an F-test comparing the reduced model, the model without degree but containing the other variables, to the full model with all the variables. If the variable degree is statistically significant than there should be a detectable reduction in the error sum of squares (and an equal increase in the regression sum of squares). The commands in lines 8 & 9 provide that test. To be clear, this command/test is technically not required for this model as the F-test located in the ANOVA table provides the identical test. But, this is only because there are no other independent variables.

I am glad that this helped you!

Hi Roseani, It is difficult to know exactly what comparisons are important for you to highlight. One method would be to use the "by()" option to graph box. Maybe something like this: sysuse auto, clear graph box length displacement, by(rep78) In this example, length and displacement are two (of your five variables) and rep78 stands in for a variable representing your panels. This produces one graph for each value of rep78 (our for a panel in your case). You can reorganize the graph to put all the graphs in a column or row which might help. This might get you a little further along in solving your problem.