Indexes and Scales Social Research Methods 2117 6501

Indexes and Scales (指標及量表) Social Research Methods 2117 & 6501 Fall, 2006 November 20, 2006 1

Recall: many social concepts are complex and multi-dimensional • • • So we often use multiple indicators to measure a concept. Composite measures (複合式的測量 ) of variables: indexes and scales Data-reduction devices (資料化約的 具） 2

Also remember some important issues about measurement: • Attributes and variables: mutually exclusive and exhaustive • Unidimensionality (單一向度): all the items in a scale or index fit together (i. e. , they measure the same thing. ) 3

Indexes vs. Scales 相同之處: 不相同之處: • Ordinal measures of • an index: a variables (次序的測量) combination of items into a single score (個 • Composite measures 人屬性積分的累計) of variables (having • a scale: assigning more than one data scores to patterns of item) responses (各問項間 有相對的強弱程度之 分，所以依程度給予 問項不同的配分比重) 5

Scales are generally superior to indexes. 1. Indexes are used more often in social research. 2. More discussions are on scales. 6

Index Construction Four main steps: • Item Selection (題目的選定): each item has face validity • Examination of Empirical Relationships (實證關係 的檢驗) • Index Scoring (將數個題目給分以建立指標) – Handling Missing Data (遺漏值的處理) – Weighting (加權): assign each item different weights • Index Validation (有效化、檢查指標的效度) 7

Index Construction • Index Scoring – Decide the desirable range – Assign scores—items treated as equal (建議：相同比 重的配分） – May need rates and standardization (標準化) • Handling Missing Data – – – If few cases, exclude them Have reasons to replace missing data: educated guess May find an interpretation Assign the middle value or assign values at random A somewhat difficult choice: bias vs. influencing the nature of the findings (處理遺漏值有時很困難，必須在 不造成更多偏差及能有效分析資料兩者之間決定) 8

Example of index 9

Index Validation (註: 我們會在討論量 化資料分析時再詳細討論這個議題) • Item Analysis（題目分析） – How the index is related to the individual item • External Validation（外在的效度） – Examine the relationship of an index to other indicators of the same variable • Bad index vs. Bad Validators – Check the relationship between the validating items and the individual items in an index 10

Scale Construction (量表的建立) A scale: consider different intensities of items in a composite measure (考慮指標題目不同的強度) 11

Examples of Scales • Likert Scale (summated-rating or additive scales)(總和評分或加法量表) – Questionnaire item containing response categories such as “strongly agree, ” “agree, ” disagree, ” and “strongly disagree” is often called as a Likert scale. – One of the most commonly used formats in questionnaire design – The value of a Likert scale: unambiguous ordinality of response categories (答案選項的清楚順序) – Be aware of the problem set (回答傾 向)(or called response bias) (回答偏頗) 12

Examples of using the Likert Scale to create indexes 13

Examples of using the Likert Scale to create indexes 14

Examples of Scales • Bogardus Social Distance Scale: to determine the willingness of people to participate in social relations with other kinds of people (測量人們與他人建立各種 社會距離的意願) • Thurstone Scale: generating groups of indicators of a variable that have at least an empirical structure among them (變項 的指標題目間至少有一實證結構) 15

Examples of Scales • Semantic Differential (語意差異): ask respondents to choose between two opposite positions • Guttman Scale: some items may prove to be more extreme indicators of the variable than other – Coefficient of reproducibility (複製係數) – Based on the structure observed among the actual data 16