Dummy Variable Help

[Becca88222]

I'm using Eviews6 student version

This is my dilemma:

I Have data from 1990-2008 for Salaries in Japan.
I have 4 education levels: college, advanced (masters, etc), high school and Junior high...
and I have the salary for each of those 4 education levels for male and female
I have all of that on excel...so I have 8 columns of data

This is my issue: what I want to do is generate an equation where I can compute the salary if i put in 0 for male, 1 for female and 1,2,3,4 for the corresponding education levels

So pretty much what I want to do is to estimate a regression inputting:

ls c salary education gender time


And i have NO CLUE how to create those dummy variables for time and education....plus I have no clue how to create a variable for salary !!

PLEASE HELP!!!!!!!!!!!

[EViews Gareth]

Your question isn't particularly clear. Without knowing what form your data is currently in, it is hard to say how to transform it.

I have no idea, for example, what you mean by "how to create a variable for salary", given you already said that you have salary information.

[Becca88222]

Sorry If i wasn't clear

What I mean is that I don't have one separate variable for salary, what I have is this table in excel

Year College Degree Advanced Degree High School Junior High College Degree Advanced Degree High School Junior High
1990 169.9 145.4 133.0 117.0 162.9 138.1 126.0 107.1
1991 179.4 155.1 140.8 123.5 172.3 146.5 133.2 114.8
1992 186.9 160.9 146.6 128.0 180.1 152.4 139.5 117.2
1993 190.3 165.1 150.6 131.5 181.9 155.6 142.4 120.1
1994 192.4 166.6 153.8 134.9 184.5 157.7 145.5 122.8
1995 194.2 165.1 154.0 142.0 184.0 158.7 144.7 125.5
1996 193.2 166.8 154.5 146.6 183.6 158.7 146.1 130.8
1997 193.9 168.9 156.0 141.6 186.2 161.0 147.3 131.8
1998 201.5 174.4 156.5 143.4 186.3 161.8 147.9 129.0
1999 196.6 170.3 157.6 144.1 188.7 162.2 148.3 139.5
2000 196.9 171.6 157.1 140.6 187.4 163.6 147.6 137.5
2001 198.3 170.3 158.1 145.1 188.6 163.8 148.7 134.0
2002 198.5 169.5 157.5 141.9 188.8 164.3 148.8 125.7
2003 201.3 169.8 157.5 151.2 192.5 163.5 147.0 133.5
2004 198.3 170.7 156.1 145.4 189.5 164.2 147.2 137.6
2005 221.0 196.7 170.3 155.7 216.6 189.3 164.2 148.0
2006 224.6 199.8 171.2 157.6 226.0 190.8 166.8 149.4
2007 224.6 198.8 171.2 158.8 226.7 191.4 166.9 150.8
2008 226.2 201.3 171.6 160.0 223.6 194.6 168.6 154.3



Columns 2,3,4,5 are the salaries for Males in Japan according to year and education level
columns 6,7,8,9 are the salaries for Females in Japan according to year and education level...

This is what I want to be able to do:

I want to set up a regression where I can calculate the salary given education sex and year

So I guess what I am doing here is setting up two dummy variables: sex (where 1= female and 0=male) and education: (where you can input 1 for college, 2 for advanced, 3 for high school and 4 for junior high)

Thankssss

[EViews Gareth]

This is one of those cases where it is probably easier for an EViews novice to do it in Excel rather then EViews.

If you want to do it in EViews, post the Excel file.


Edit: Just to be clear - it is probably easier for you to create the dummy variables in Excel, and then bring it into EViews to do the estimation. Especially since you're using the student version.

[Becca88222]

Here is the excel file

Thanks for your help!

[Becca88222]

How would I create the dummy variables in excel?

[startz]

Copy and paste so you have one long series for salary for both men and women. Then make a new column for the dummy and enter 1 for the male entries and 0 for the female entries.

[Becca88222]

and how would i differentiate between education levels in the list?
and how would i create the dummy for the education level?

[startz]

Make a column that's 1 for college salaries and 0 otherwise. Same for other levels.

[Becca88222]

I think that I'm not being clear on what I want.......

What i want to do is generate a regression with the variables salary, education and sex so that I can do a hypothesis that proves that women, for every education level, are paid less than men in Japan....

Maybe I'm doing this the wrong way, I don't know

[startz]

So you need one column that lists all salaries. Another column that is 1 or zero according to gender. Another column that is 1 or 0 according to whether the salary was for college level. Etc.

[Becca88222]

Okaayyyyyyyyy

And how do i account for the year?
The info is from 1990-2008

[startz]

Include a column with the year number!

The trick in all this is that you need one long colunm for each variable you want to use in your regression.

[Becca88222]

Thank you!!

So for the column for the year number its okay to repeat years for example

1990
1991
1992
1993
2008

and then directly underneath:

1990
1991

So forth?

[startz]

You got it!