Στατιστικές αναλύσεις με STATA

STATA is owned by StataCorp LLC and its main advantage is the freedom it provides to the user through continuous growing and evolving libraries provided by the company and users through various sites or blogs such as. The Stata blog. Although the main advantage of the program is its code, some users may find it difficult to write the code and manipulate the results.

 

With a simple example we will try to show some of the capabilities of the program through time series analysis. We used the GDP index dataset for the years 1995 to 2017 from ELSTAT .

 

Import dataset command

 

import excel «C:\…\stata_example.xlsx», sheet(«Φύλλο1») firstrow

 

year pop GDP
1995 10562164 8811.0354
1996 10608821 9712.3557
1997 10661312 10759.669
1998 10720566 11684.323
1999 10761705 12431.927
2000 10805796 13071.436
2001 10862146 14011.397
2002 10902005 14993.642
2003 10928091 16371.103
2004 10955163 17682.605
2005 10987352 18133.788
2006 11020393 19768.947
2007 11048499 21061.195
2008 11077863 21844.501
2009 11107024 21385.943
2010 11121383 20324.041
2011 11104995 18642.861
2012 11045040 17311.292
2013 10965241 16475.176
2014 10892369 16401.986
2015 10820964 16381.013
2016 10775989 16377.889
2017 10768193 16736.104

 

Descriptives

label variable year «Year»

label variable pop «Population»

label variable GDP «GDP»

 

Για την καλύτερη γραφική απεικόνιση των δεδομένων κατασκευάσαμε μια νέα μεταβλητή, την pop2

 

generate pop2=pop/100

label pop2 «Πληθυσμός (Εκατοντάδες)»

 

The population change time diagram is generated by the tsline pop2 command and produces the following graph

Similarly, command tsline GDP produces the below graph

 

We observe a similar change of the two sizes characterized by the steady decline in their prices after 2010.

Finally, the simultaneous representation of the two variables with the help of the tsline GDP pop2 command can not render in detail the time changes.

 

One solution is to combine these two graphs with the help of commands

 

line pop2 year, saving(g1)

line GDP year, saving(g2)

gr combine g1.gph g2.gph

 

The first test was performed with the help of a linear regression model with the help of the command

regress GDP year

and the results showed that the model is statistically significant (F (1.21) = 16.45, p <0.001) and of moderate interpretability (R2=43.93)

The comparison of real versus estimated values with the help of commands

predict fitted_values

 

line fitted_values GDP year

shows that linear regression can not capture changes in GDP

 

The second estimate was made using the ARIMA method (p, d, q) and led to the ARMA model (2.0) as it presented the lowest BIC and AIC coefficient. The new assessment was made with the help of the commands

arima GDP, arima(2,0,0)

predict f1

estout, stat(aic bic)

line f1 GDP year

was clearly improved as shown in the graph below

This was a simple example of using STATA for analysis. For any question or if you need help with your analysis you can contact us.

 

Contact us

For any question or for information please fill in the following contact form or send an email to info@basispap.com .

Adress:

P.C. 65201

Snt Barbara

Kavala

Mobile:

6931 258 164

Working hours

Mon-Fri 9am-9pm
Saturday 9am-2pm





Call Now Button