## 23 Questions

What is a common problem in statistical analyses involving real data sets?

Missing data

In which type of studies is missing data a common problem?

Non controlled, observational studies

What statistical technique is used to estimate parameters in a population different from that in which the data was collected?

Inverse probability weighting

What is the main limitation of regression imputation method?

Imputed values within each variable are not equal due to the values of the remaining variables.

According to the text, what does IPW do when applied to deal with missing data?

It weights each observation by the inverse of the probability of being sampled.

What assumption do MI techniques make about missing data?

Missing data can be replaced by predictions derived from observed data.

What is the MICE algorithm used for in multiple imputation?

Imputing realistic values to the missing data and propagating the uncertainty.

How are the missing values for each variable handled in the MICE algorithm?

By using predictions from fitted regression models after performing a regression model analysis.

Which statistical technique uses inflation of weight for under-represented subjects due to a large degree of missing data?

Inverse probability weighting

In the context of linear regression analysis, what is performed for each imputed dataset in multiple imputation?

A typical regression analysis with Yi as the dependent variable and all other variables as independent predictors.

What can be performed using the mice package in R according to the text?

Multiple imputation analysis.

What is Var(✓) MI used for in multiple imputation?

Captures the uncertainty of the imputations and inflates the error of the estimate accordingly.

What is the consequence of missing data being MNAR?

Sample size reduction and biased parameter estimates.

What method replaces each missing datum with the sample mean, median, or mode of the variable computed using available data?

Mean/median/mode substitution

What is the primary assumption for estimates to be unbiased under complete cases analysis?

Data is missing completely at random (MCAR).

Which type of missing data pattern implies that the probability of an observation being missing does not depend on the value of the observation or any other variables in the dataset?

MCAR (Missing completely at random)

What is the consequence of using mean/median/mode substitution for missing data?

Unrealistic imputed values and underestimation of variance.

What is the primary characteristic of MNAR (Missing not at random) data?

The probability of missing data depends only on the value of the observation itself.

What are the consequences when using complete cases analysis?

Sample size reduction and biased parameter estimates.

What characteristic differentiates MAR from MCAR?

The probability of missing Xi depends on observed data, not missing data.

What is one of the potential consequences when data are MCAR?

Sample size reduction and statistical power reduction.

What is a primary characteristic that defines MAR (Missing at random) data?

The probability of missing Xi depends on observed data, not missing data.

What is one of the potential consequences when data are MNAR?

Sample size reduction and biased parameter estimates.

Learn about dealing with missing data in health statistics in this quiz based on the B.Sc. Degree in Applied Statistics. Explore the different types of missing data and methods to handle them, presented by Jose Barrera from ISGlobal Barcelona Institute for Global Health.

## Make Your Own Quizzes and Flashcards

Convert your notes into interactive study material.

Get started for free