Study description: In a study of violent victimization of women and men, Porcere
ID: 3254105 • Letter: S
Question
Study description: In a study of violent victimization of women and men, Porcerelli et al. (A-11) collected information from 679 women and 345 men ages 18 to 64 years at several family-practice centers in the metropolitan Detroit area. Patients filled out a health history questionnaire that included a question about victimization. The following table shows the sample subjects cross-classified by gender and whether the subject self-identified as being “hit, kicked, punched, or otherwise hurt by someone within the past year.” Subjects answering yes to that question are classified “violently victimized.”
What are the study questions?
What is the outcome of interest (dependent variable)?
What are the predictors of outcome (independent variables)?
What statistical model was used?
What statistical hypotheses were tested to address the study questions?
Were the null hypotheses rejected at the level 0.05?
What conclusions should be made based on the statistical analysis results?
In addition to summary questions 1-7 answer the following questions:
What is the odds ratio (with corresponding confidence interval) associated with the gender difference for the rates of victimization?
What is the interpretation of this odds ratio?
The LOGISTIC The LOGISTIC Procedure Model Information WORK.VICTIMIZED Data Set Response Variable (Events victimized Response Variable ITrials) total binary logit Model optimization Technique Fisher's scoring Number of observations Read 2 Number of Observations Used Sum of Frequencies Read 1024 Sum of Frequencies Used 1024 Response Profile Total ordered Binary Value outcome Frequency 1 Event 119 2 None vent 905 Class Level Information Design Class Value Variables Sex F 1 0 Frequency Distribution of CLASS Variables outcome Class Value Event Nonevent Total sex F 68 611 679 51 294 345 Model Convergence Status Convergence criterion (GCONV-1E-8) satisfied.Explanation / Answer
The study question is to predict the probability of a person being 'Violently Victimized'.
The outcome of the interest is probability of a person will be classified as Violently Victimized'.
The predictors of the outcome is basically two variable based the Category of Sex. 1. Binary if the sex is F. 2. Binary if sex is M.
As the outcome variable is binary, i.e. 1 or 0 outcome type, so a logistic model has been used.
The statistical hypothesis addressed to study the questions is: Whether the independent variable Sex is important to predict that a person will be violently victimized. That is the Null hypothesis : coefficient of Sex F is equal to zero vs Alternative Hypothesis : coefficient of Sex F is not equal to zero
From the table of coefficient of Sex F we see that, Pr > ChiSq = 0.0253 < 0.05, which means the null hypothesis is rejected.
From, the coefficient of Sex F, we clearly conclude that, the variable is effective in predicting the outcome, but from the Somer's D perspective, it is very less, so the overall model performance needs to be improved.