COMPOSITE SCORES INDEXES SCALES TYPOLOGIES WHY USE COMPOSITE

  • Slides: 10
Download presentation
COMPOSITE SCORES INDEXES, SCALES, & TYPOLOGIES

COMPOSITE SCORES INDEXES, SCALES, & TYPOLOGIES

WHY USE COMPOSITE SCORES ¡ May not be a “clear” single indicator ¡ Offers

WHY USE COMPOSITE SCORES ¡ May not be a “clear” single indicator ¡ Offers an increase in variability ¡ Makes the analysis of data more “efficient”

Index or Scale ¡ Index l l ¡ Scale l ¡ Most “common” Cumulative

Index or Scale ¡ Index l l ¡ Scale l ¡ Most “common” Cumulative score (perform mathematical operation Assign values to a “pattern” of responses (note: not all items are “equal” Note: both indexes and scales are rank-order composite measures

Index Construction ¡ Item Selection l l ¡ Scoring l l ¡ ¡ Variability

Index Construction ¡ Item Selection l l ¡ Scoring l l ¡ ¡ Variability vs “Adequate number of cases in categories Item weighting (equal/unequal) in the index Empirical Examination l ¡ Face validity Unidimensionality General vs Specific Variability Bivariate vs Multivariate Missing Data Validation

Scales and Scaling ¡ Empirically based l ¡ Responses are “scaled (not the survey

Scales and Scaling ¡ Empirically based l ¡ Responses are “scaled (not the survey questions/items Intensity of particular items l Is there a pattern in how the items tap a concept?

Scale Construction ¡ Not all items are “equal” l l ¡ “intensity structures” among

Scale Construction ¡ Not all items are “equal” l l ¡ “intensity structures” among items assurance of “ordinality” Examples: l l l Thurstone Scale Bogardus Social Distance Scale Likert Scale Semantic Differential Guttman Scale

Typologies ¡ Researcher can summarize the “overlap/intersection of two or more variables l ¡

Typologies ¡ Researcher can summarize the “overlap/intersection of two or more variables l ¡ Typically driven by some combination of “theory” and “data” ¡ Life-course criminality ¡ Probation example Difficult to use a typology as the dependent variable l Question: why do particular cases “fit” in a particular typology (predicting typologies)

Using SPSS to Create an Index Locate the data file ¡ Identify the variables

Using SPSS to Create an Index Locate the data file ¡ Identify the variables ¡ Define the “New Variable” ¡ Define the “mathematical operation” ¡ Select “Transform” ¡ Select “Compute” ¡ Enter commands ¡ Save file ¡

Characteristics of Items ¡ Recoding of some items may be necessary ¡ Not all

Characteristics of Items ¡ Recoding of some items may be necessary ¡ Not all items (statements/questions) are in the same “direction” ¡ Managing missing data ¡ “Don’t Know” is problematic

Verification of Composite Scores ¡ Factor Analysis ¡ Reliability Coefficient l l Item-Item Correlation

Verification of Composite Scores ¡ Factor Analysis ¡ Reliability Coefficient l l Item-Item Correlation Item-Total Correlation