OLS Regression

[Guest]

Hey folks
i'm writing a term paper about the influence of parenthood on life satisfaction. Since it's my first term paper with STATA ever i'm pretty confused with the OLS-regression and the variables (yes i'm still a poor little bachelors student).
I'm using a german dataset called ALLBUS with 3490 observations. Since i'm only interested in parents, I dropped all non-parents from my sample by creating a variable parent no/yes and then "drop if parent==0" which leaves 2230 obs. Then i created variables "general life satisfaction"=dependent (0-10; only dropped the missings), "number of children" = independent (1-9; including biological and adopted/step-children), and two seperate dummy-variables "child_biological" = independent (1=at least one biological child) and "child_adoptedstep" = independent (at least one stepchild or adopted child). I'm controlling for gender (dummy), age (centred), relationship yes/no (dummy), employment status (categorial), household income (centred and log) and status of health (1-5).
When i regress e.g. general life satisfaction and number of children, adjusted r-squared is negative and really really small (<-0.00), the same thing happens with the dummy variables for biological and adopted children so for all my independent variables but not for the variables which i control for.

What did I do wrong? Is it okay to just drop all the non-parents from the sample? Might there be a problem with the adopted/step-children since theres only 60 people who have at least one? Do i have to recode the variable for number of children into a categorial variable since there's very few people who have more than 4 children (less than 50)?

It might be pretty basic and pretty confusing at the same time but I'm happy for all the help that i get!

Greetings, Guest
(ps. sorry for my english I'm from Germany)

[Carlo Lazzaro]

Guest:
welcome to this forum.
For statistical assignment (like term paper), see thttps://www.statalist.org/forums/help#adviceextras.
That said, even replying on marginal respects of your question is difficult, as you describe what you're experiencing (and it take too long to read what you wrote) but you do no provide (as per FAQ again):
-what you typed and what Stata gave you back (via CODE delimiters, please);
- an example/excerpt of your data (via -dataex-, please).

[Guest]

Hi Carlo,
i read the extra advice, does the point "homework" mean that i shouldn't ask a question concerning my term paper since it'll be graded (although the grade won't be relevant for my final bachelors degree grade)?
And if I post an example of my data i guess i'd have to translate the variables and lables into english before posting it?

Kind regards,
Guest

[Carlo Lazzaro]

Guest:
1) yes and no (in my opinion, at least. Other contributors may legally have different takes on this point). Yes covers your queries that follow your attemtps to provide/sketch Stata code(s) and output. No covers all the rest.
2) yes, I think it would be easier for those who are not German mother-tongue (or fluent) to understand what's the matter with your data.

[Clyde Schechter]

Let me weigh in on the term paper issue here. The reason for the norm that we do not help with homework problems here is that homework assignments normally as given with the understanding that they will be completed by individuals without any outside assistance. By contrast, we do give advice to people working on master's theses or doctoral dissertations--because it is generally understood that some degree of involvement with outsiders is necessary and expected. Undergraduate senior theses are often, though not always, similarly associated with permission to involve others.

A term paper falls into a gray area. If the paper was assigned with the expectation that it would be completed independently, then it would be inappropriate to provide assistance here. But if the paper was assigned with the understanding that the student is permitted, perhaps even encouraged, to seek help from others, then it would be appropriate for Forum members to ask and answer related questions.

I suggest that Guest consult her school's policies that apply to this term paper to determine which category it falls into. If there are no directly applicable school policies, then she should ask the teacher who assigned it what the expectation is.

While I agree with Carlo that providing example data with the variable names translated to English would be helpful to most members of the Forum, the most important thing is to show the data. Variable names are often meaningless alphanumeric strings (a practice I dislike and discourage, but it happens frequently, especially with large-scale surveys). And even when they are in English, they sometimes are jargon words that, to non-specialists in the topic of the research, might as well be in a foreign language.

[Guest]

Unfortunatly, I think asking the school or the teacher would take too long since it's pretty urgent.
But thanks anyways for the help!
I'll find a way to solve the problem myself hopefully.

Greetings!