Short Question of Statistics: Part-II

Basic Statistics 

Teacher Education: A comparative Perspective (8626): Download

Question: What is the difference between the Residual and Error? 

Answer: R^2 is the coefficient of the determination; which describe how much percent source of variation in the dependent variable is describe by the linear relationship with the independent variable. Large the value of R^2 is better. 

R-Adjusted^2 describe how much percent of the variation in the dependent variable is describe by the other factors (error term) or, how much percent the model will be able to accept the independent variable. Small the value of R-Adjusted^2 is better of the model.

Statistics: Short Question Part-I: Click here

Question: Describe in brief generalized variance functions?

Answer: Generalized variance functions (GVFs) ar provided in a very range of surveys to

calculate commonplace errors. Definition: a formula for variance estimation created supported a regression model for the variance.

Advantages: The GVF is also used once insufficient info is provided within the public-use knowledge files to permit direct calculation of standard errors. the information collector will calculate the GVF, and sometimes has a lot of information for estimating variances than is released to the general public. A GVF saves an excellent deal of your time and speeds production of annual reports. it's conjointly helpful for coming up with similar surveys within the future.

Disadvantages: The model relating vi to ^ti may not be applicable for the number you're

interested in, leading to associate unreliable estimate of the variance. you need to use caution concerning using GVFs for estimates not enclosed once calculating the regression parameters. If a subpopulation has a curiously high degree of agglomeration (and thence a high deff), the GVF estimate of the variance is also a lot of too tiny.

Question: When Two-phase sampling is useful?

Answer: this is often helpful once the variable of interest y is comparatively costly to live, but a related to variable (x) may be measured fairly easily and accustomed improve the exactness of the estimator of ty. it's going to even be accustomed regulate for non-response, to sample rare populations, or to improve the sampling frame.