R help - May 2017 - Error: invalid type (closure) for the variable 'time' - object specific trend

Hey,

I just have trouble adding a object specific time trend with the 
plm-package. I recieve the following error:
*
**"Error in model.frame.default(terms(formula, lhs = lhs, rhs = rhs, 
data = data,  : invalid type for the variable 'time' "*

I used the formula:

/ FE_trend<- plm(log(revenue) ~ log(supply)+ city*time, 
data=R_Test_log_Neu, model="within", effect = "twoways") //
/
Do you have an idea how to fix it?

Toby

	[[alternative HTML version deleted]]

On 11/05/2017 9:37 AM, Tobias Christoph wrote:
> Hey,
>
> I just have trouble adding a object specific time trend with the
> plm-package. I recieve the following error:
> *
> **"Error in model.frame.default(terms(formula, lhs = lhs, rhs = rhs,
> data = data,  : invalid type for the variable 'time' "*
>
> I used the formula:
>
> / FE_trend<- plm(log(revenue) ~ log(supply)+ city*time,
> data=R_Test_log_Neu, model="within", effect =
"twoways") //
> /
> Do you have an idea how to fix it?
I don't know that package and you didn't supply a reproducible example, 
so there are at least these two possibilities:

1.  Your dataframe R_Test_log_Neu contains no column named "time". 
(Remember that "Time" and "time" are different.)  Fix:  use
the correct
column name.

2.  The plm package isn't doing lookup of symbols from the formula first 
in your dataframe.  This would need to be fixed in the package.  A 
workaround might be to name the columns with unique names that won't be 
found anywhere else, e.g. name your column "NeuTime" instead of
"time".

Duncan Murdoch

Hey Duncan,

thank you very much for your quick reply.

_My data used:_

1st column(town):1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,...........,11

2nd column(year):1,2,3,4,5,6,7,8,9,10,11,12,1,2,3.......,12

3rd column (revenue):

4th colum (supply):

I have now renamed my colums and did the regression again. Now there is 
a problem with R-squared, as it is the sum of 1 now with no given std. 
error and t-value. This is probably due to the fact, that I try to 
estimate more parameters than data.

To add a linear trend I found the following formular:*lm(y ~ x1 + 
factor(ccode)*time, data=df)*

I try to I decode it for and use it for my regression: *plm(log(revenue) 
~ log(supply) + factor(town)*time, data=R_Test_log_Neu)*

When I do this regression I will get the original error: "invalid type 
(closure) for the variable 'time' - object specific trend"

With the notation"time" not my colum is meant, but probably the
command
"time" in R.

Can you follow my thoughts?

Tobi






Am 11.05.2017 um 17:23 schrieb Duncan Murdoch:
> Duncan Murdoch