Median Procedures Research Articles

Gender-role classification and demographic relationships: A function of type of scoring procedures

This study investigated the effect of different median split (combined-sex, single-sex, weighted, and unweighted) procedures on Bem Sex Role Inventory gender-role classifications and the effect of these procedures on demographic relationships. Results revealed significantly different gender-role classifications for all procedures except single-sex weighted and unweighted. No significant demographic relationships were found for men as a function of type of scoring procedures, but significant relationships for SES and marital status were found for women when single-sex weighted and unweighted median splits were used. Discussion focused on recommendations for further research using various median split procedures.

Mediastinitis after cardiovascular surgery.

Postoperative mediastinitis complicated 21 (3.4%) of 616 median sternotomy procedures at Beth Israel Hospital (Boston, Mass.) between 1975 and 1979. These cases were analyzed by means of a case control study to identify host and operative risk factors and to characterize the clinical features of mediastinitis. Eighteen patients with mediastinitis (86%) had serious underlying noncardiac diseases, as compared with 14 (33%) of 42 noninfected controls (P = .001). Reoperation was positively associated with infection (P = .03). All patients had abnormal sternal wounds and fever; sternal instability and mediastinal widening were unusual. Twelve patients (57%) were bacteremic. Twenty-four organisms were recovered from the 21 patients with mediastinitis; 13 of the isolates were gram-positive, and 11 were gram-negative. Infections due to gram-negative bacteria appeared earlier and were more likely to be bacteremic (70%). All gram-negative isolates and five of six isolates of Staphylococcus epidermidis were resistant to the antimicrobial agent used perioperatively. Patients were treated with extensive debridement and appropriate antibiotics. The mortality rate was 24% (five of 21). Long-term complications in survivors were not seen.

Robust regression using repeated medians

Abstract : The repeated median algorithm is a robustified U-statistic in which nested medians replace the single mean. Unlike many generalizations of the univariate median, repeated median estimates maintain the high 50% breakdown value and can resist the effects of outliers even when they comprise nearly half of the data. Because they are calculated directly, not iteratively, repeated median procedures can be used as starting values for iterative robust estimation methods. For bivariate linear regression with symmetric errors, repeated median estimates are unbiased and Fisher consistent, and their efficiency under Gaussian sampling can be comparable to the efficiency of the univariate median. (Author)

The median procedure in cluster analysis and social choice theory

Classical approachs for fitting and aggregation problems, specially in cluster analysis, social choice theory and paired comparisons methods, consist in the minimization of a remoteness function between relational data and a relational model. The notion of median, with its algebraic, metric, geometrical and statistical aspects, allow a unified treatment of many of base problems. Properties of median procedures are organized according to four directions: stabilities and axiomatic characterizations; Arrow-like properties; combinatorial properties; effective computational possibilities. Finally, interesting mathematical problems, related to the notion of median are surveyed.

Comments on median split procedures for scoring androgyny measures

With Bem's shift from advocacy of t-difference to median split procedures for scoring the Bem Sex Role Inventory (BSRI), all the major androgyny measures are now scored according to median splits. These comments are directed to ward differences in interpretation of the scale as a result of varying scoring methods, and the psychometric problems introduced by median split procedures.

Hyperactive and normal children and the error, latency, and double median split scoring procedures of the matching familiar figures test

To further evaluate the construct validity of Kagan's Matching Familiar Figures Test (MFF), three different scoring methods—the MFF error measure, the MFF latency measure, and Kagan's double median split—were used to compare learning disabled and behavior disordered boys at two age levels with their normal peers on the frequency with which they were designated as impulsive. In the comparison of 7-year-old and 12-year-old learning disabled boys with the normal children, the MFF error measure identified a greater frequency of the learning disabled boys as impulsive at both age levels. No differences occurred between the learning disabled and the normal boys at either age level on the latency measure. Kagan's double median split procedure differentiated only between the older learning disabled and the normal boys. Thus, combining the latency measure with the error measure detracted from the effectiveness of the error measure. None of the MFF scoring procedures differentiated between the behavior disordered and the normal boys at either age level. Although the findings tend to support the results of studies with normal children in indicating that the error measure is more sensitive than the latency measure, the failure of the error measure to designate a greater frequency of the severely behavior disordered boys as being impulsive than are their normal peers raises further questions about the construct validity of the MFF, regardless of the scoring procedure used.