Questions tagged [misspecification]
Problems with model specification, such as missing variables/predictors, wrong functional form, wrong variance or covariance structure, etc.
121
questions
5
votes
1
answer
170
views
No more than $n$ moose, but how many?
Introduction
I am thinking about how to estimate the number of individual moose from wildlife camera photos. I have the latitude and longitude position of each observation, along with a datetime of ...
0
votes
0
answers
17
views
How should I use a larger dataset with fewer variables to help create a good model of a smaller dataset with more variables?
I have two data sets, DF1 and DF2.
DF1 has millions of observations, and 20 variables.
DF2 has ~100,000 observations, and 40 variables. The DF1 variables are a subset of the DF2 variables.
I want to ...
0
votes
0
answers
24
views
Link test for pooled logistic regression
After logistic regression of the cross-sectional data sets, link test _hatsq shows insignificant. However, when I pool the same two data sets, the link test using the same set of variables regression ...
0
votes
0
answers
20
views
misspecified models coverage and efficiency
Suppose I have model $M$ generating data $Y=\beta_0+\beta_1X+\beta_2Z+\beta_3W$ with all $\beta$'s known. Instead of using model $M$, I used misspecified models $M':Y=\beta'_0+\beta'_1X+\beta'_2Z$, $M'...
0
votes
0
answers
19
views
Oaxaca-Blinder in Macroeconomics?
I'm currently trying to have a look at the effect of renewable energy consumption and innovation on economic growth. One technique I was looking at for showing this was running an Oaxaca-Blinder ...
3
votes
1
answer
101
views
Ramsey RESET Test
The Ramsey RESET test uses the fitted value of y to test nonlinearity, for example:
$$
y_i=x_i\beta+\epsilon
$$
$$
\hat{y_i}=x_ib
$$
$$
y_i=x_i\beta+\gamma\hat{y}^2_i+u_i
$$
Test if $\gamma=0$
Why do ...
16
votes
4
answers
1k
views
Priors that do not become irrelevant with large sample sizes
This may be a weird question. My colleagues and I are working on a medical estimation problem, where relevant prior knowledge regarding plausible values of some physiological parameters exists. In ...
1
vote
1
answer
29
views
The concept of a specification test
There is a confusing facet of the concept of a specification test.
According to many textbooks, a specification test is roughly a statistical test investigating whether assumptions in a statistical ...
2
votes
0
answers
49
views
Equation for a DiD/Event Study Combining Several Simple DiD
I am trying to do a Difference in Differences/Dynamic Event Study but my data doens't easily lend itself to a classic staggered treatment DiD nor a simple 2X2 DiD. My dataset has 1-min readings from ...
2
votes
1
answer
52
views
Is this a valid model specification?
I am working with weekly, highly seasonal data (period=52) and fitting the following regression using lm:
$$
y_t=\rho_1y_{t-1}+\rho_2y_{t-52}+\beta x_t+\...
0
votes
1
answer
2k
views
Linktest results interpretation
linktest's rule of thumb is that _hat should be statistically significant while _hatsq ...
3
votes
0
answers
78
views
A misspecification error with linear models that can complete reverse the direction of an effect, has this been described, has this a name?
Linear models are ubiquitous in economic, social, health and nutritional sciences and the starting point for much research and many articles.
However, there is a problem with linear models. When the ...
2
votes
0
answers
49
views
Why are parametric models usually robust to mild misspecification?
Statisticians often use parametric families of models, for example normal distributions with unknown mean and variance. However, nothing in real life is perfectly normally distributed (or distributed ...
8
votes
4
answers
1k
views
What is the benefit of regression with student-t residuals over OLS regression? [duplicate]
Sometimes I see advice to fit regressions with student-t residuals rather than using OLS (which is equivalent to assuming normally distributed residuals) if the distribution of the residuals is heavy-...
1
vote
0
answers
34
views
Not sure of model choice - significant p-values but very low R2 adj
I am writing a thesis in which I'm trying to see what the relationship is between how innovative a patent is (as determined by a series of measures based on NLP that) and the financial value of that ...
0
votes
0
answers
13
views
Changing relation between independent and dependent variable over time
I am currently trying to run a linear regression model where the effect of one of my independent variables (Xit) on my dependent variable (Yit) changes over time. Specifically, I suspect the ...
1
vote
0
answers
24
views
Do insignificant variables result in a specification error?
I am trying to understand omitted variable bais better. I know that it detects irrelevant variables, but are irrelevant variables and insignificant variables synonymous here? If I have a regression ...
0
votes
1
answer
274
views
The exact meaning of "correctly specified"
I will cite a paper I am reading: (Bugni, Federico A., Ivan A. Canay, and Xiaoxia Shi. "Specification tests for partially identified models defined by moment inequalities." Journal of ...
2
votes
1
answer
78
views
Concerns about pre-trend stability testing
In this discussion, @Thomas Bilach well explains the equation to test pre-trend stability.
$$
y_{kt} = \alpha_k + \lambda_t + \delta_{-2} d_{k,t-2} + \delta_{-1} d_{k,t-1} + \delta d_{kt} + \delta_{+1}...
3
votes
1
answer
763
views
What is the consequence of misspecification in logistic regression
In linear regression, the Ramsey RESET test can be used to test if the model is misspecified. The Gauss-Markov theorem allows us to understand the consequences of misspecification in linear regression....
1
vote
0
answers
97
views
Is traditional negative binomial regression robust to model misspecification or not?
By "traditional" NBR I mean NB2, i.e. the one modeling variance as a quadratic function of the mean, with the formula: $Var(Y)=E[Y]*(1+\alpha*E[Y])$. I have found contrasting statements in ...
2
votes
1
answer
125
views
Does the problem of multiple testing also apply to the testing of assumptions?
When dealing with statistical tests, sometimes we can run into cases where many assumptions would apply and consequently would need to be tested. In complex models, testing many assumptions at 5% may ...
0
votes
1
answer
459
views
Simple explanation of Takeuchi’s information criterion?
Takeuchi’s information criterion is said to be the generalization of AIC to misspecified models. That publication presents DEGREES OF FREEDOM FOR NONLINEAR LEAST SQUARES ESTIMATION. From that source:
...
3
votes
1
answer
391
views
ARMA/GARCH statistical significance of estimated parameters
My question is general and is concerned with ARMA-GARCH modeling.
When performing the joint estimation of the ARMA and GARCH parts, some works tend to not be concerned with the statistical ...
1
vote
1
answer
188
views
Wald Statistic Simplification - 2 Restrictions?
(Specify the wald statistic as: W=$(R\hat{\theta}-r)'[R\hat{V}(\hat{\theta})R']^{-1}(R\hat{\theta}-r)$
Where r is a vector of restrictions, R imposes the restrictions on $\hat{\beta}$, $\hat{\theta}$ ...
0
votes
1
answer
28
views
Issues in testing a simple linear relationship. Collinearity? Misspecification? Any other insight?
I have a theoretical model saying that Y should be equal to:
Y = X + c * (W - X) + (Z1 - Z2),
where c is a given constant.
Here, it may be important to say that X is measured with error.
Someone ...
1
vote
0
answers
104
views
Misspecification in the outcome model with regression adjustment for propensity score?
The propensity score is a popular tool used to control for confounding by covariates $C$ on the effect of an exposure $A$ on an outcome $Y$. There are several ways to incorporate the propensity score ...
2
votes
1
answer
815
views
Question about regression using a rate vs a count as the dependent variable
Let's say I wanted to run a regression of mortality on an x variable, so I could see 3 different possibilities I can do
I define a rate: $y = \left(\frac{\#deaths}{population}\right)$
and regress
$$...
0
votes
0
answers
102
views
Bayes estimates and model misspecification
Consider a misspecified model:
$$P \sim \text{SegmentedUniform}(0,1) \\ Y \mid P \sim \text{Binomial}(N,P).$$
Where SegmentedUniform has uniform density on intervals $$(0.1, 0.2), (0.3, 0.4), (0.5, 0....
1
vote
1
answer
1k
views
Event Study with two treatments
Lets say I wanted to include two different treatments in a diff in diff at the same time, so I could have:
$y_{i,t} = \lambda_i +\tau_t + Treat1_i*post_t+Treat2_i*post_t + \eta_{i,t}$,
where $\...
1
vote
1
answer
98
views
Role of misspecification by biased data in the generalization error
I am confused with the role that model misspecification plays in the generalization error, in particular when the misspecification is due to a biased (non representative) training dataset. To clarify ...
4
votes
1
answer
357
views
Consequences of choosing the wrong GLM response distribution
In Chapter 10 of McElreath's Statistical Rethinking (2nd edition), he argues that the response distribution for a GLM should be chosen to maximize entropy given a set of constraints on the response ...
0
votes
1
answer
81
views
Practical numerical example illustrating how bayesian p-values should be used for model checking?
Can you give a practical numerical example illustrating how bayesian p-values should be used for model checking?
3
votes
3
answers
470
views
Bayesian inference with false models: to what does it converge?
This is the second follow up question from these two previous questions:
Bayesian inference and testable implications
How do I perform an actual "posterior predictive check" in this model?
...
3
votes
0
answers
1k
views
How do I perform an actual "posterior predictive check"?
This question is the follow-up of this previous question: Bayesian inference and testable implications.
For concreteness, consider the following bayesian model. This model is not to be taken ...
1
vote
0
answers
140
views
How bad are short-term forecasts of ARIMA models of ARFIMA processes?
Suppose I have a set of economic time series that appear to be a unit root process. I difference them, and fit an ARMA model to the differenced series. Suppose however that the true data generating ...
0
votes
0
answers
48
views
Bayesian model averaging when none of the models is well specified
Usually, from what I've read of Bayesian Model Averaging (BMA), a typical assumption is that one of the models is well specified...
However, what will happen when all models are misspecified? What ...
2
votes
0
answers
37
views
SEM: Why is there no dependable connection between the size of SEM residuals and the type or degree of model misspecification?
Kline (2016) p. 278 writes
[Y]ou should know that there is actually no dependable or trustworthy
connection between the size of the residuals and the type or degree of
model misspecification. For ...
1
vote
0
answers
60
views
Are there any practical differences in fitting a random slope vs. fitting an interaction term in the intercept in lmer
I am trying to fit a few models with the form
...
1
vote
0
answers
150
views
Buld prediction intervals for glmmTMB
I'm using glmmTMB to build a mixed effect logistic model from which I want to draw predictions.
The predict() function applied to a glmmTMB model allows extracting the ML prediction and the relative ...
0
votes
1
answer
1k
views
Path analysis model fit statistics. RMSEA = 0.106
Hi folks,
My path analysis is using primary data that I've collected for a uni project. After removing the very non-signif variables, I have a model (which I hypothesised) that seems to work pretty ...
0
votes
0
answers
577
views
how to interpret the smoothed effect of time in GAMM models
I have recently taken a plunge into the world of Generalized Additive Mixed Models (GAMMs) after learning that including the quadratic and cubic fixed and random effects of time to understand students'...
3
votes
0
answers
296
views
Over "specification" of a statistics model?
For a simple example, I am fitting data with a likelihood function generated by a normal distribution. The first model is the normal distribution with two-parameters. The second competing model is the ...
4
votes
1
answer
223
views
Non-Linear regression and variance misspecification
Given a non-linear regression model for cross-section data
$$y_i = f(x_i,\theta_0) + \epsilon_i,$$
where it is assumed that $\mathbb E[y_i\lvert x_i] = f(x_i,\theta_0)$, I understand that it is a ...
2
votes
2
answers
2k
views
$R^2$ is too high- reasons? [closed]
What are the reasons of too high values of $R^2$? I only know that its value increases when the number of independent variables increases. What are the other factors?
1
vote
0
answers
14
views
Question on econometric specifications transformation [closed]
Two groups of forecasters are trying to forecast outcome variable y by issuing forecast X. Denote the two groups of forecasters with Z={1,2}. We want to know whether group 1 is more accurate in ...
2
votes
1
answer
164
views
Bayesian inference with simple models and many data points: impact of priors and the number of data points
The setting. Let us assume I would like to perform Bayesian inference for a low-dimensional model on a large dataset, i.e., there are many more data points than there are parameters to identify. Let ...
1
vote
0
answers
84
views
When regularizing based on an informative prior, how to give model a little more freedom to partially reject regulariziation
I am new here I hope this question is appropriate.
I am modelling a spatial domain, whereby I have repeated measures at n locations. I make a bayesian linear model at each n locations based on about ...
2
votes
0
answers
41
views
Random effects for mutually inclusive grouping factors
I am trying to fit a model on a set of data (e.g. 10,000 observations with 20 explanatory variables). The observations belong to 30 groups, G1, G2, G3, ... G30, so I need to account for the grouping ...
3
votes
0
answers
289
views
Relative importance different model diagnostics in LMM/GLMM
I may be over thinking things but I have been stuck on this issue for the last few days. I am currently modelling data with a large number of clusters but cluster size is limited to 2, with a skewed ...