Yes, you can. However, you need to take into account that fact when interpreting the results of your empirical work.

Regards.

I agree with Dr. Angel Martin Roman's advice

Yes, you can perform a linear regression. Just try to control for other covariates so that you would not be subject to an omitted variable problem, as government expenditures alone may not completely explain economic, and try to be careful with the interpretation of your coefficients.

You can, if you have to.

No problem with using percentages or decimal or variables in mixed format for regression. The coefficients adjust themselves to suit the units of measurement. While reporting results clearly mention what units which variable is measured in (decimal, percentage, ppm, etc.) The R-square value does not depend on the unit of measurement either.

Hi Pitrika,

it is possible for you to do the linear regression analysis.

All the best

Yes, you can perform linear regression analysis with percentage data

I think Dr. Angel Martin Roman has said all that needs to be said. Pritika also know that she is going to run the data and everyone knows OLS is what is used. What Dr. Roman noted that others are ignoring is interpretation.

Also, Pritika knows that certain assumptions have to be addresses before Dr. Roman's point comes. She is a smart lady!

yes, but if the total amount of gov expenditure (including budget equality) is variable, you may need to enter another predictor and check the pair correlation and intragroup dispersion share

Yes, as has been argued by my colleagues, just take note of the interpretation. Here the interpretation will be one-on-one? Also, try controlling for other covariates that may correlates with government expenditures and may also affect growth.

Yes, better with control variable.

Nice contribution. special events such as Asia financial crisis and 2008 global crises may treated with dummy variable

Yes Pritika, you can perform linear regression analysis with percentage data for both dependent and independent variables.Philips curve is one good example where both dependent and and independent variables are measured in percentage.

No. Economic growth is measured by GDP and government spending is part of GDP, so there is a problem in accounting there. You can measure in a regression, for example, economic growth based on the share of government spending in GDP.

I would make things very simple: Any change in Government Spending (IV) would influence GDP (DV). You can express this in value terms or in percentages.

As long as the two series are multiplied by 100 this may lead Autocorrelation problem. So I recommend doing a regression without multiply by 100 .I think the results will be better