Bronwen Perley-Robertson, Kelly M Babchishin, L Maaike Helmus
{"title":"The Effect of Missing Item Data on the Relative Predictive Accuracy of Correctional Risk Assessment Tools.","authors":"Bronwen Perley-Robertson, Kelly M Babchishin, L Maaike Helmus","doi":"10.1177/10731911231225191","DOIUrl":null,"url":null,"abstract":"<p><p>Missing data are pervasive in risk assessment but their impact on predictive accuracy has largely been unexplored. Common techniques for handling missing risk data include summing available items or proration; however, multiple imputation is a more defensible approach that has not been methodically tested against these simpler techniques. We compared the validity of these three missing data techniques across six conditions using STABLE-2007 (<i>N</i> = 4,286) and SARA-V2 (<i>N</i> = 455) assessments from men on community supervision in Canada. Condition 1 was the observed data (low missingness), and Conditions 2 to 6 were generated missing data conditions, whereby 1% to 50% of items per case were randomly deleted in 10% increments. Relative predictive accuracy was unaffected by missing data, and simpler techniques performed just as well as multiple imputation, but summed totals underestimated absolute risk. The current study therefore provides empirical justification for using proration when data are missing within a sample.</p>","PeriodicalId":8577,"journal":{"name":"Assessment","volume":null,"pages":null},"PeriodicalIF":3.5000,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11490059/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Assessment","FirstCategoryId":"102","ListUrlMain":"https://doi.org/10.1177/10731911231225191","RegionNum":2,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/2/7 0:00:00","PubModel":"Epub","JCR":"Q1","JCRName":"PSYCHOLOGY, CLINICAL","Score":null,"Total":0}
引用次数: 0
Abstract
Missing data are pervasive in risk assessment but their impact on predictive accuracy has largely been unexplored. Common techniques for handling missing risk data include summing available items or proration; however, multiple imputation is a more defensible approach that has not been methodically tested against these simpler techniques. We compared the validity of these three missing data techniques across six conditions using STABLE-2007 (N = 4,286) and SARA-V2 (N = 455) assessments from men on community supervision in Canada. Condition 1 was the observed data (low missingness), and Conditions 2 to 6 were generated missing data conditions, whereby 1% to 50% of items per case were randomly deleted in 10% increments. Relative predictive accuracy was unaffected by missing data, and simpler techniques performed just as well as multiple imputation, but summed totals underestimated absolute risk. The current study therefore provides empirical justification for using proration when data are missing within a sample.
期刊介绍:
Assessment publishes articles in the domain of applied clinical assessment. The emphasis of this journal is on publication of information of relevance to the use of assessment measures, including test development, validation, and interpretation practices. The scope of the journal includes research that can inform assessment practices in mental health, forensic, medical, and other applied settings. Papers that focus on the assessment of cognitive and neuropsychological functioning, personality, and psychopathology are invited. Most papers published in Assessment report the results of original empirical research, however integrative review articles and scholarly case studies will also be considered.