Categorical Data Analysis Quiz Answers 2024

What kind of variable is "blood type"? - answers-Nominal What kind of variable is "gender"? - answers-Nominal What kind of variable is "opinion"? - answers-Ordinal What kind of variable is "weight"? - answers-Ratio What kind of variable is temperature? - answers-Interval For ordinal variables, order matters and difference between levels is important to the researcher? t/f - answers-False If n is large and pi is small, a Poisson distribution with mu=npi provides an approximation to the binomial distribution? t/f - answers-True Which of the following is not an example of a Poisson Distribution? - answers-None of the above Specificity is the probability has a positive test result given that he/she has the condition? t/f - answers-False 1. The random variable Y is distributed as binomial. Given that n=20 and pi=0.5, find the mean and . of Y respectively. - answers-20*0.5=10=mean variance =20*0.5(1-0.5)=5 sqrt of 5= 2.24= ANS: 10 and 2.24 What is the Z value and P value respectively? - answers--3.16 and 0.0016 What is your conclusion from question 11? - answers-Reject null (Ho) Compute the marginal probability of lung cancer. - answers-50/140 35.71% Compute the marginal probability of non-smoking. - answers-78/140 55.71% Compute the conditional probability of lung cancer among smokers. - answers-32/62 51.61% Compute the odds ratio of having lung cancer for smoking compared to non-smoking. - answers-32*60/18*30=3.56 Compute the relative risk of having lung cancer among smoking subjects compared to non-smoking. - answers-32/62= .52 18/78=.23 .52/.23=2.24 The relative risk is more meaningful then the difference in proportions when the proportions are far from zero.t/f - answers-False Which of the following is not used to compare proportions? - answers-Odds Ratio The Relative Risk is approximately the same as the Odds Ratio for events with low prevalence? T/F - answers-True The specificity is the true positive rate while the sensitivity is the true negative rate? T/F - answers-False Which of the following represents the power of a study? - answers-1-False Negative Rate (Type 2) Which of the following is more powerful to test for independence of ordinal variables? - answers-CMH Linear Trend Test Which of the following is False about the odds ratio? - answers-The asymptotic SE of odds ratio can be expressed as (1/a+1/b+1/c+1/d)1/2 The confidence interval of RR tends to be conservative while the CI of difference is proportion tend to be liberal. T/F - answers-True Which of the following is true about the Pearson Chi-Squared test for a contingency table? - answers-All of the Above CMH-1 - answers-CMH Linear Trend Test RelRisk - answers-Relative Risk Fisher - answers-Fisher exact Test CellCHI2 - answers-Each Cell's Contribution to the Pearson Measures - answers-Odds Ratio CHISQ - answers-Pearson Chi Square The sample Pearson correlation between maternal alcohol consumption and congenital misinformation r=0.03(n=7,367). Compute the linear trend statistic and p-value. - answers-M2=6.63 p=0.01 Which of the following distributions belongs to the natural exponential family? - answers-All of the Above The offset variable cannot be a class variable. T/F - answers-True Suppose the dependent variable Y is on the vertical axis and the independent variable x is on the horizontal axis. - answers-The scatter plot will generate four points if both Y and X are dichotomous and coded as 1 and 0 ANOVA Model and Logestic Regression Model do not belong to the family of generalized linear models.T/F - answers-False The link function g(mui)=mui is called the identity link and the link function g(pii)=log(pi/1-pii)) is called the canonical link. T/F - answers-True The identity link function is used if we build a logistic regression model using proc genmod? T/F - answers-False Which of the following is false about the logistic regression model? - answers-None of the Above In SAS which of the following procedures should be used for GLM for count data? - answers-Proc Genmod Which of the following is true about the scale option in proc genmod - answers-All of the Above Linear probability regression model has a larger range for the linear combo of Xmus. Therefore , it is better than probit regression model.T/F - answers-False Which of the following models corresponds to a Poisson Loglinear model: - answers-log u=α + βx The HL test for goodness of fit is used for grouped data and ungrouped data as well:T/F - answers-False What tests are adequate for goodness of fit for grouped data?: - answers-Both A and C Which of the following is true about the Hosmer-Lemeshow test - answers-All of the Above In PROC LOGISTIC in SAS, the LACKFIT option in the MODEL statement procedures - answers-The Hosmer-Lemeshow goodness of fit test The "scale=None" equation proc logistic procedures - answers-The deviance and Pearson Chi Square If the estimated slope Beta-0.1 and SE(Beta)=0.95 for a simple logistic regression model, then the 95% confidence interval of the slope Beta equals - answers-(-0.09,0.29) Logistic regression requires all variables used in data analysis to be categorical, while linear regression requires all variables used in data analysis to be continuous. t/f - answers-False In logistic regression, both the Wald Test and the Likelihood ratio test can be used to test for a single regression coefficient. T/F - answers-True For Logistic regression model, the dependent variable should be: - answers-Nominal or Ordinal For akaike information criterion, the larger the value, the worse the fitted model? T/F - answers-True For the backward selection, the model starts with all of the predictors of interests? t/f - answers-True An ROC curve is a plot of sensitivity against false positive rate? t/f - answers-True Which of the following is false about AIC - answers-Cannot be used for models that have likelihoods of Poisson distributions. Which of the following is true about the multicollinearity in logistic regression? - answers-All of the Above Multicollinearity can be detected by examining - answers-A and C Which of the following indicates multicollinearity? - answers-correlation coefficient=-0.71 Which of the following indicates an influence - answers-All of the Above Which of the following is not a selection algorithm in regression - answers-none of the above The Concordance Index c, also known as the c statistic, represents the AUC, the area under the ROC curve. The closer is to 0, the better the predictive capabilities of the fitted model? T/F - answers-False The difference of AIC between the full model and the reduced model is 0.2413 - answers-The reduced model is preferred AIC or BIC can be used to compare underlying distributions assumptions? T/F - answers-True Multicollinearity can be a problem of all regression modeling - answers-incorrect conclusions about relationships between independent and dependent variable may be drawn due to multicollinearity In statistics, multicolinearity is a phenomenon in which two or more predictor variables in a multiple regression model are highly correlated t/f - answers-False VIF TOL - answers-True