Cross Section Analysis

[Gunnarson Mike]

Hello everyone

Apologies for the beginner's question, but I am struggling with simple cross-sectional analysis.

Could anybody help about how to conduct a Cross Section Analysis on Stata

I am trying to estimate the determinants of firms' productivity by applying cross-section analysis. I have already gathered the required data for the time period in question and determinants as well.

But I was wondering what commands to run on stata to perform a simple cross-sectional analysis and obtain meaningful results.

Any kind of help would be highly appreciated....thanks in advance !

[George Ford]

regress y x1 x2 x3

[Ken Chui]

But I was wondering what commands to run on stata to perform a simple cross-sectional analysis and obtain meaningful results.

Welcome to Statalist!

Simply put, analysis does not operate that way. There aren't "commands" that can create meaningful results, just like there aren't cooking utensils that create a meal. Same command, when applied to a different cause, can create meaningful results and well as completely ridiculous results.

The core lies in what your research questions are (just like what the meal plans are), without knowing that, it's hard to confidently suggest anything. I'd recommend first by checking out some literature on how to plan an analysis first. Here are two that may be helpful to get you started:

https://www.cdc.gov/globalhealth/hea...l_09242013.pdf
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4552232/

[Gunnarson Mike]

Welcome to Statalist!

Simply put, analysis does not operate that way. There aren't "commands" that can create meaningful results, just like there aren't cooking utensils that create a meal. Same command, when applied to a different cause, can create meaningful results and well as completely ridiculous results.

The core lies in what your research questions are (just like what the meal plans are), without knowing that, it's hard to confidently suggest anything. I'd recommend first by checking out some literature on how to plan an analysis first. Here are two that may be helpful to get you started:

https://www.cdc.gov/globalhealth/hea...l_09242013.pdf
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4552232/

Dear Ken Chui,

Thank you for your response..your reply was well noted I checked upon the articles too.

I am currently trying to estimate the determinants of firms' productivity by applying cross-section analysis

Any idea about how to proceed with it ?

I have collected the data and tabulated it as below :

COUNTRY	YEAR	LAB_PRODUCTIVTY	OBS_WRKFORCE	BRIBERY_REQ	PAY_SEC	FIN_CONS	FEMALE_PAR	ELECTRI_CONS	DEAL_GOVTREG	FORMAL_TRAINING
Azerbaijan	2019		12.1	10.1	69	6	28.5	4.9	3.4	37.1
Bosnia and Herzegovina	2019	0.7	33	18.2	48.6	15.6	27.7	24.4	6.5	40.9
Bulgaria	2019	0.3	21.5	15.2	70.6	13.7	42.6	21.3	10.3	24.5
Czech Republic	2019	1.3	46.1	0.4	67	13.5	33.4	57	12.4	38.8

I was wondering how to use this data on stata and obtain some useful results

Thanks for your help

[Gunnarson Mike]

regress y x1 x2 x3

Dear George,

Thanks for your response.

But in the event I have multiple countries as in the above table, how can I proceed with cross section analysis

[Carlo Lazzaro]

Mike:
take a look at -regress- entry in Stata .pdf manual anre related reference (any decent textbook on linear regression or econometrics can do the trick).
That said, from your post #3 I'm not clear with what your query is about.
You are probably aiming at a linear regression with standard errors clustered at country level (due to the lack of further details I assume that your regressand is -LAB_PRODUCTIVTY-):

regress LAB_PRODUCTIVTY OBS_WRKFORCE BRIBERY_REQ PAY_SEC FIN_CONS FEMALE_PAR ELECTRI_CONS DEAL_GOVTREG FORMAL_TRAINING, vce(cluster COUNTRY)

[George Ford]

If you have one observation per country, then proceed with regress. Not much more you can do. Maybe use natural log transformation on variables in levels since there may be very large differences in scale across countries. If you have multiple observations per country, then you are getting into panel data estimation for which there are many ways to proceed. xtreg or reghdfe are suited for that.

[Haile Girma]

How can I formulate a spreadsheet from cross-sectional data that is compatible for analysis by STATA?

[Carlo Lazzaro]

Haile:
see -help import-.