(Stata13): Tips to Building Panel Data in Stata

[@CrunchEconometrix]

KZbin recently changed the way my content will be monetised. My channel now needs 1,000 subscribers. So it would be amazing if you show your support by both watching my videos and subscribing to my channel if you haven’t done so already. Monetising my videos allows me to invest back into the channel with some new equipment so this small gesture from you will be extremely huge for me. Many thanks for your support….CrunchEconometrix loves to teach, help me stay online.

[@sebastianosalarin8925]

Although I am 3 years late, I've just watched some of your videos and I can tell you that you earned it! Just proudly subscribed to your channel to show you my support! Your contents are really well done and clear! Keep it up mate!

[@CrunchEconometrix]

Thanks, Sebastiano for the encouraging feedback. Deeply appreciated!!! 💃 🙏❤️

[@BrianJasonPonce]

This is by far the most helpful video on how to format panel data. Thank you!

[@CrunchEconometrix]

Thanks, Brian for the commendation. Deeply appreciated!!!

[@BrianJasonPonce]

@@CrunchEconometrix I hope you continue to make highly informative videos like this one. Your uploads have made a great contribution to the pursuit of my Master's studies.

[@CrunchEconometrix]

A million likes for your comment, Brian!

[@mohamedeltayeb9269]

Thank you so much. I am doing research on linguistics and this was very helpful. As a Sudanese, I am also very glad to see an intelligent African woman teaching the world about advanced econometric techniques.

[@CrunchEconometrix]

Thanks, Mohamed for the kind words. Deeply appreciated!

[@horveysenyosylvester4285]

Hello, From South Africa, I just finished watching all the nine videos. It is the best teaching I have received on GMM. Cheers!

[@CrunchEconometrix]

Thanks for the encouragement, Horvey. Deeply appreciated!

[@mohamedlamroony2601]

Thanks alot dr , absolultly amazing explaining

[@CrunchEconometrix]

My pleasure, Mohd!

[@vasuagarwal9683]

Thank you for this amazing video! 🥰

[@CrunchEconometrix]

You are so welcome, Vasu!

[@monzamana3127]

# respect from Tunisia Great job, your videos are very helpful and your explanations are understandable. Thanks

[@CrunchEconometrix]

Thanks for the encouraging feedback, Monza. Deeply appreciated! Much love from Nigeria :)

[@nowrinjahan546]

Thank you soo much... Your videos are very very helpful.

[@CrunchEconometrix]

U're welcome, Nowrin... thanks! 🥰🙏

[@lazharlabed1211]

good morning Thanks for the great video, I benefited a lot from it I built a model and got these results Arellano-Bond test for AR(1) in first differences: z = -4.69 Pr > z = 0.000 Arellano-Bond test for AR(2) in first differences: z = -0.54 Pr > z = 0.589 ------------------------------------------------------------------------------ Sargan test of overid. restrictions: chi2(61) = 100.02 Prob > chi2 = 0.001 (Not robust, but not weakened by many instruments.) Hansen test of overid. restrictions: chi2(61) = 64.61 Prob > chi2 = 0.352 (Robust, but weakened by many instruments.) Difference-in-Hansen tests of exogeneity of instrument subsets: GMM instruments for levels Hansen test excluding group: chi2(51) = 53.11 Prob > chi2 = 0.393 Difference (null H = exogenous): chi2(10) = 11.50 Prob > chi2 = 0.320 iv(x4) Hansen test excluding group: chi2(60) = 64.57 Prob > chi2 = 0.320 Difference (null H = exogenous): chi2(1) = 0.04 Prob > chi2 = 0.842 Can the form be accepted or rejected? Thank you

[@CrunchEconometrix]

Hi Lazhar, thanks for the positive feedback and kind words about my video. Deeply appreciated! But you posted on the wrong thread. I covered most of the the basics in the 9 GMM series. Kindly use the information I have provided to evaluate your results. Thanks. Please may I know from where (location) you are reaching me?

[@lazharlabed1211]

@@CrunchEconometrix thanks, I'am from algeria i hanve a nother question: The robust option , should it be used or dispensable, as its addition in my model affected the sig of the model parameters My greetings

[@wanjadouglas3058]

Always amazing Doc. Learnt alot

[@CrunchEconometrix]

Thanks as always, Doc!

[@wanjadouglas3058]

You are just amazing and saved the world

[@CrunchEconometrix]

Thanks for the encouraging words, Sir...appreciated!

[@wanjadouglas3058]

@@CrunchEconometrix always. I may come to Nigeria to give a lecture one day

[@esayasmequanint9539]

Thank You Dr.

[@CrunchEconometrix]

You are welcome, Sir!🥰🙏

[@ibaldaus]

Very clear explanation.

[@CrunchEconometrix]

Thanks for the positive feedback, Iqbal!

[@99evan76]

Tips to Building Panel Data in Stata 1. reshape wide to long stata 2. group classification (industry) 3. group classification 4. 4:30 categorise outcome variable a. Get variable in place then use the xtset 5. Create dummy variables

[@CrunchEconometrix]

Thanks.

[@joshmoodley9466]

Thank You!!! SO helpful

[@CrunchEconometrix]

You're welcome, Josh!

[@asgharkamal4633]

its was very fruitful video for my research study coz i have 64 obs, then create a dummy thank madam without your support its impossible for us to complete our PhD.

[@prof.abdinasiraliabdi2794]

that very generous

[@CrunchEconometrix]

Thanks for your encouraging feedback, Abdinasir! 🥰🙏

[@christopherschroder4686]

This is so so helpful thank you very much! I would recommend maybe to do a video how to merge different data sets, especially panel data sets (with some data sets having more countries and years than others, which need to be dropped). This is what I have to do right now and I couldn't find any video for this. All the best

[@CrunchEconometrix]

Thanks for the encouraging feedback, Christopher. Deeply appreciated! I have also noted the suggestion. I will work on it.

[@atahiroumahamane7587]

Very helpful...Thanks

[@CrunchEconometrix]

Thanks Atahirou, for the encouraging words and feedback. Deeply appreciated! Please may I know from where (location) you are reaching me?

[@atahiroumahamane7587]

@@CrunchEconometrix thanks too Pr. Dr. . I am from Niger Republic but studying in Turkey now.

[@KashifIslamOnline]

aoa.. I have been your viewer and completed my phd by learning your ARDL in eviews and stata...... kindly guide me how i can use your data of [[(Stata13): Tips to Building Panel Data in Stata]]]

[@CrunchEconometrix]

Hi Kashif, thanks for your encouraging feedback. Deeply appreciated. Please know that due to abuse and unethical conduct, all Stata dofiles and some datasets used in my videos are no longer free but available on my website upon payment. Here's the link cruncheconometrix.com/view/datashop.php

[@wanjadouglas3058]

Very helpful. Is it possible to get this raw data and the dofile for practice?

[@CrunchEconometrix]

Thanks for the encouraging feedback. Deeply appreciated. Datasets used for all my videos are available on my website upon payment (though some are free). But all Stata dofiles are available on payment. Here's the link👇 cruncheconometrix.com/view/datashop.php

[@rafaelleal9011]

Perfect

[@CrunchEconometrix]

Thanks Rafael!

[@rebeccagramiscelli8149]

Thank you for the help, from Brasil! =)

[@CrunchEconometrix]

You are welcome, Rebecca! 💞 😍

[@vivien7293]

You've really helped me with understanding STATA in total! Thank you so much! I just wanted to ask, is there a video on how to classify data in Excel that I can use to help me?

[@CrunchEconometrix]

Hi Vivien, thanks for the encouraging feedback. Deeply appreciated. Unfortunately, I don't have such video.

[@enongenebetrand1119]

Amazing video Dr. save me alot of time. deeply appreciate. But i have an issue on how to classify the data in excel interms of regions and income level. Any hint on how to classify panel data in excel?

[@CrunchEconometrix]

Hi Mr. Betrand, I answered your query in this video. Kindly watch it again in addition to my other panel data videos. Thanks

[@kimlais2180]

Thank you very much for your highly informative video! I have a question regarding the categorization for inequality: Could you provide a source where I could find these categorizations? I've been using yours but couldn't find the exact page or source for it. Kind Regards

[@CrunchEconometrix]

Hi Kim, I use the World Bank categorization for countries. Do a Google search. It's out there.

[@rgportocarero7889]

Dear Madam, Again, a very helpful video, thank you. My dataset looks at mean scores of companies from 2009 until 2020, not unlike your data (except companies instead of countries). Company Year Score X 2009 70 ... X 2020 80 Y 2009 60 ... X 2020 65 I am looking at the statistical significance in score increases over the years of the aggregate sample (e.g mean of 2009 scores to 2010 over all companies). In your case: the question would be to calculate the Levene's test on equality of mean between yearly score increases/decreases of trade. My question: I need to perform a Levene's test of equal means (sdt test x == sdtest y) over the covered years to test for equal variance difference. Therefore, I will probably need to perform a test of each year against one another. When I declare data to be panel data by xtset scores years, it gives the error of repeating time values in the sample. I was wondering how to tackle this problem so that when I perform the test is still takes the mean of each year. Much much appreciated!!

[@CrunchEconometrix]

Hi Rg, I'm unable to relate your query with this video. Levene test?

[@rgportocarero7889]

@@CrunchEconometrix Hi, thank you for your response. I am testing for equality of variances in the dataset (=Levene's test). My dataset resembles yours. My question therefore, how I would need to code or test the equality of variances from (take your dataset): Trade from one year (e.g 2009) to another (e.g2010, 2011). First, you would need to calculate the aggregate value per year (2004,...2012) over all countries. Then you would need to compare the variance of these average values per year against one another. (2004 vs 2005, 2005 vs 2010) Thank you.

[@frmabhijit]

In cases where the cross-sectional units are companies, does the process remain the same?

[@CrunchEconometrix]

Yes.

[@yusufovvv]

Hello, I need your help. I was doing research. On the impact of tax reform on business investment in Australia. I have to dummy the years in this one. I did not understand that. Do I have to get the Y for Dummy lash or do I get each of my X's as well? I have one Y 5 X variables help me I just need to submit tomorrow as soon as possible or I will fail my project I hope you will read soon.

[@CrunchEconometrix]

Hi Yusuf, I have so many good videos on DUMMY VARIABLES. Kindly watch them to get ideas on what to do.

[@favkop_bencarson1695]

Hello Doctor, please thank you for your insightful tutorials, i have a question, please since income classifications of countries changes over time, would it be advisable to code for instance Albania as UMinc throughout the whole years under observation?

[@CrunchEconometrix]

Hi Bencarson, use the most recent World Bank classification.

[@basanijamelahoveni6852]

Very useful video, what does it mean you have to reset to historic background in order to create Dummy variables?

[@CrunchEconometrix]

Hi Basani, "reset historic background"? I don't know what you mean.

[@kelennaolemgbe770]

Thank you, your videos are enlightening. Please, I need a clarification on group classification and dummy variables creation. My data does not have any category, how can I then create the dummy variables? Thanks

[@CrunchEconometrix]

Hi Kelenna, categorize your data and proceed to create the DVs.

[@kelennaolemgbe770]

@@CrunchEconometrix Thank you

[@anuradhasaikia9305]

How to analyze unbalanced panel in STATA ma'am?

[@CrunchEconometrix]

Hi Anuradha, it is the same approach with balanced panel.

[@farwakhalid4041]

Hi, thanks for this video. It was helpful to understand how to build Panel Data in Stata. I need some help, I want to add Control Variables (region and industry of companies), but I don't to know how to differentiate it in the data that they are control variables? Is there any specific way to create a difference between independent, dependent and control variables? Or should I upload entire data and Stata will determine itself. Please guide, your help means a lot here. Thanks in advance.

[@CrunchEconometrix]

Farwa, just specify all in the model. Your interpretation tells the readership the nature/type of the variables.

[@farwakhalid4041]

@@CrunchEconometrix Thanks a lot for the response, I have added all the variables in the model. :)

[@debosmitachatterjee9994]

Hi! Many thanks for the video. It was immensely informative. Please may I ask you a question? I have an unbalanced panel dataset. Do i need to convert it to a balanced dataset before running panel regressions? Thank you so much and any advice you offer is highly appreciated.

[@CrunchEconometrix]

Hi Debosmita, you can run regressions with unbalanced panel data.

[@debosmitachatterjee9994]

@@CrunchEconometrix Thank you so much. for the reply.

[@mohamedouedraogo7597]

Hello, Prof. Adeleye, Thank you for this interesting lecture. Greenhouse gas emission (GHG) is my dependent variable. I would categorize it into five parts like (Very Low Emission), (LE), (ME), (HE) and (VHE). How to recode instead in what you did? I use Stata 16. Thank you in advance.

[@CrunchEconometrix]

Hi Mohd, forgive my very late response. Fiirst, you have to obtain the industry/scholarly categorisation for GHG emissions from credible sources, then you can split the sample as shown in the video. Thanks.

[@mohamedouedraogo7597]

@@CrunchEconometrix Hey Prof. Adeleye, Never mind. Okay, I got it and many thanks.

[@nadinekukamujinga5351]

This video is really helpful and easy to understand, is it possible to have some tips for diagnostic check (validity and consistency) of panel data and like in my case i am using GDP growth rate as dependent variable which may have endogeinity....which tests and tips to overcome it?

[@CrunchEconometrix]

Hi Nadine, what analytical software are you using?

[@briankaiche2405]

A very helpful video, thank you so much. But madam, i have a question and it has to do with the data set that you have used in the presentation, i guess its a Swiid data set. I'm also doing a research paper where the gini coefficient is my dependent variable and I've opted to use the Swiid dataset due to its efficacy on coverage of the data points. So basically in the stata.dta of the swiid there are many variables of the gini coefficient that are prefixed starting from _01 to _100 and I'm so much conflicted on which variables among those gini coefficient on desposable income am i supposed to get and represent it as my dependent variable. So, is there a way to compute all those gini coefficients to come up with a single variable that should consist of 1 gin variable that can be used as the dependent variable. Or it is just ok to use the gini coefficients that are provided in the excel format. Please help me, thank you.

[@CrunchEconometrix]

Hi Brian, the SWIID dataset comes with NOTES on each variable. I suggest you use the most applicable inequality index to your research. I use only one Gini indicator.

[@diamonddogsgaming7059]

Hi I have a question regarding the dummy variables. Why do we need a year and regional dummy variable? For example I am doing an analysis on homicide rates in the developing countries and I have created year, regional and religion dummy variables. I am using a fixed vs random effect model and when I run the hausman test without including the year dummy variables, it states fixed effect is appropriate and rejects the null hypothesis. But if I include the year dummies and run the hausman test, it now says random effects model is more appropriate. My question is why do we create year dummy and how do we interpret it in the regression analysis?

[@CrunchEconometrix]

Hi there, watch these videos for answers to your questions: Error Component Models (5 videos), Stepwise Regression (2 videos), and Panel Sub-Sample Analysis.

[@KiranSingh-vf8nc]

Is there anywhere that I could reach you, I have a few questions to ask regarding data cleaning and a model I am trying to estimate for my thesis. Thanks in advance

[@CrunchEconometrix]

How can I help you, Kiran? Kindly post your queries (keep them very brief). Thanks.

[@sebaskalkman3832]

what do you mean with firm fixed effects? what type of effects would that be? With industry fixed effects you test in which industry which companies are operating, but with firm fixed effects i'm not sure... Does that mean if I have e.g. the STOXX600, which included 600 companies, the fixed effect as measured in firm class is just every 600 firms individually?

[@CrunchEconometrix]

Hi Sebas, firm fixed effects capture the inherent differences across the firms in the sample. Please may I know from where (location) you are reaching me?

[@meerrasiq1169]

Respected mam, I just need a little clarification Mam as far as my little knowledge goes , the inequality data (gini) you are using in this video is not available in a balanced form. There are a lot of missing values ( data) in between the years for the majority of the Countries whether you refer to World Bank or Solt. But your data shows no missing values and for all the years data is available. How mam?? When gini data is not available for all 141 countries without missing values.

[@CrunchEconometrix]

There are missing values in the data.

[@ag959]

Hello! the link for the data base is not working... Did you have other data bases to do an interesting panel data analysis? thank you !

[@CrunchEconometrix]

I deactivated the link due to ethical breach. Some datasets are available on my website upon payment of a token fee. Here's the link cruncheconometrix.com.ng/shop

[@Osegbue]

How can someone capture the year dummies on running regression? Are they going to be capture as independent variables? I have a work I have created industry and year dummies? Can I capture them as independent variables on running the regressions? How can I interpret them assuming they are not in Logs?

[@CrunchEconometrix]

Watch my videos on: Error Component Models, Panel Sub-Sample Analysis and GMM video on year dummies. You can adapt my interpretations.

[@frmabhijit]

If we create 7 dummy variables for 7 regions, will it create a dummy trap?

[@CrunchEconometrix]

Hi Abhijit, please watch the following videos. Well explained to guide you. Listen to the explanations and FOLLOW the guides: 1) Building a panel data 2) Error Component Models 3) Sub-samples Analysis Also, watch the videos on DUMMY VARIABLES published on my Teachable paid platform. All these videos are explicitly explained. Take down notes while watching them. Thanks!

[@MuhammadBilal-nv4dz]

why dont we just use i.year insted of creating year dummies?

[@CrunchEconometrix]

Sure you can!

[@muhammadfarooq8967]

l wand to download gini data from all-over the world.. 2000 to 2020 I want help

[@CrunchEconometrix]

Check GCIP.org for global data on income inequality.

[@diamonddogsgaming7059]

Hi I have a question. I have multiple countries in my panel dataset, but if I want to have summary statistics of 2 or 3 different countries of my own choice together, is there a command to do that? I know how to do it for a single country, but if I wish to compare the descriptive statistics of one country with another is there a command to do that?

[@CrunchEconometrix]

Watch my video on "panel data descriptive statistics".

[@Sal9862]

I have my data sorted like you habe with years in collumns ranging from 2000 to 2017 for each individual firm. In your video of shaping wide to long data you have the years horizontally though. Please how can i do this?

[@CrunchEconometrix]

Hi Saloua, I reshaped the data fro wide (World Bank format) to long (Stata format). Watch that video to see what I did. May I know from where (location) you are reaching me?

[@lucomannaro1]

Hi, thanks for your video which was very helpful. I wanted to ask you a question: I have a database with on rows the different countries, and on columns the variables that I need for my anaysis organized by year (e.g. Column A: Var1 2000; Column B: Var1 2001, ...., Column M: Var2 2000, Column N: Var2 2001. etc). I was wondering if it was possible to perform a panel data analysis with such an arrangement of the database, since it's so different from what you're showing us in this video. Thanks in advance for the answer!

[@CrunchEconometrix]

Hi Luka, thanks for the encouraging feedback. Deeply appreciated! Stata will only analyze data in long formats. Watch my video on "Reshape wide to longitudinal data''. Please may I know from where (location) you are reaching me?

[@lucomannaro1]

@@CrunchEconometrix alright I'll watch that. I'm writing you from Italy, and I wrote you to ask for help for my master's degree thesis ;)

[@diamonddogsgaming7059]

Hello, I have a question on how can I export my tables to word from Stata? I am using Stata 14 so putdocx command does not work in this version. For example I have a similar table of the gini category and i wish to move that to word? is there a specific command I can use to move it?

[@CrunchEconometrix]

You can watch my videos on exporting Stata output to Word and Excel. You may find them helpful.

[@omnamovenkatesayachannel8739]

Ma'am, i have 'distance' as one of my independent variables...Now distance does not change over time..so for a particular country the distance value is same for all the years under consideration...But under Fixed Effects this distance variable is getting omitted and at the same time Hausman is suggesting to chosse Fixed Effects...So how to deal with this distance variable?...And distance is an important variable in my analysis which i cannot drop

[@CrunchEconometrix]

Hi Om, if I have this challenge, I will estimate both FE and RE models after which I make the necessary notes in my work. Settles the problem.

[@omnamovenkatesayachannel8739]

@@CrunchEconometrix Thanks for your reply.

[@onlineacademy4079]

@@CrunchEconometrix can you help me I have same problem

[@ksanjana3083]

I want to know how can I construct pseudo panel data using two cross-sectional data with different individuals and time. it's very urgent if u can plz help me.

[@CrunchEconometrix]

No idea.

[@aligohar3436]

How to create ID automatically b coz manually in Excel it is very difficult

[@CrunchEconometrix]

Hi Ali, I've always used Excel to automatically generate IDs. Watch my video on "Reshape wide to longitudinal data" and follow the steps.

[@aligohar3436]

@@CrunchEconometrix thank u so much madam... Problem solved

[@aligohar3436]

@@CrunchEconometrix i have approximately 2500 municipality data for 7 years... I got alot of help from videos... Thank u once again

[@nadahusseinshokry5886]

Is it possible to do steps 1 to 3 on STATA? as my data is already imported into STATA and when I try to import it again in Excel it says too many variables.

[@nadahusseinshokry5886]

I mean too many observations

[@CrunchEconometrix]

Hi Nada, no need to import again since data is already in Stata.