# Questions tagged [two-step-estimation]

Models in which a complicated function of data is estimated in the first step, and plugged again into another estimation model of primary interest in the second step

44 questions
Filter by
Sorted by
Tagged with
32 views

### Neutral number in XGBoost algorithm prediction

Does a concept of "neutral number" in machine learning algorithms exist? To make it clearer: suppose we have a logistic regression with only one feature, the "neutral number" is ...
1 vote
27 views

### Excluding second-stage covariates from first stage

I have a question about the inclusion/exclusion of certain covariates in a 2SLS model. My concern is that my instrumental variable, Z, influences Y through two pathways: X1 and X2. If I am ...
• 11
27 views

### Are these two estimated regression coefficient asymptotically equivalent? If not, which one is more efficient?

Suppose I have $Y=\beta_1X_1+\beta_2X_1X_2+g(X_2)+u$, where $E(u|X_1,X_2)=0$ and $S=g(X_2)+e$ with $E(e|X_2)=0$. I have a random sample $\{Y_i,X_{1i},X_{2i},S_i\}_{i=1}^n$. Suppose I first use a ...
• 2,966
35 views

### Can I use the same regressors for Instrumental variables and the outcome variable?

I'm reading a research paper that uses an IV, they regress A, and B, to estimate C, then they use estimated C and actual B to regress on D an outcome variable. This seems to violate the only through ...
67 views

### Two step maximum likelihood

Suppose I have a likelihood maximisation problem $$\hat{\theta} = \max L_n(\theta;y)$$ where $\theta = [\theta_1, \theta_2, ...., \theta_k]^T$. What if I would estimate instead estimate the ...
• 363
455 views

### Control function bootstrapped standard errors

Using a control function approach allows to incorporate non-linear first-stage specifications. A problem when doing this approach is that the naive standard errors are too small since they do not ...
• 363
87 views

### Using a regression coefficient as independent variable

Two continuous variables, $Z$ and $X$, are measured at 60 time points during a week for a given person. At the end of the week, the persons value of a variable $Y$ is measured. This is repeated during ...
• 1,838
1 vote
57 views