Soc 5 - Haveman - Week 5 - Spring 2013

Front 1

Two ways of measuring prejudice?

Back 1

1. The Godfrey-Richman ISM scale (GRISM) - pen and paper measure
2. Implicit Association Test (IAT) - behavioral measure

Front 2

Levels of measurement?

Back 2

1.Nominal - Differences between categories
2. Ordinal - Categories can be ordered.
3. Interval - Equal distance between categories is expressable in standardized units.
4. Ratio - a true zero point exists

Front 3

Validity

Back 3

The strength of conclusions we can draw from our research results.

Front 4

Construct Validity

Back 4

1. Do measures capture your theoretical concepts?
2. Two components: divergent validity and convergent validity.

Front 5

(Statistical) Conclusion Validity

Back 5

Is there a relationship between the IV and DV?

Front 6

Reliability

Back 6

Consistency of measurement or the degree to which your measurement instrument works the same each time it is used under the same conditions with the same unit of analysis.

Repeatability of measurement over:
1. Time
2. Across research subjects (your unit of analysis)
3. Across measurement instruments

Front 7

3 Types of Reliability

Back 7

1. Inter-Rater - Two observers will record the same score
2. Internal Consistency - within test people respond consistently.
3. Test-Retest - People should score about the same each time.

Front 8

4 Ways to check reliability?

Back 8

1. Test-Retest
2. Split Half
3. Use Established Measures
4. Check reliability of Research Workers

Front 9

Index

Back 9

Indicators combined in an index are given equal weights.

Subjects are scored on the basis of their total number of positive responses.

Front 10

Scale

Back 10

Indicators are combined to use any intensity structure that may exist among the indicators being combined.

Scored on pattern of responses.

Front 11

5 Types of Scales?

Back 11

1. Bogardus social-distance scale
2. Likert scales - rates strength of agreement
3. Semantic Differential Scales - rate something in terms of two opposite adjectives
4. Guttman Scales - a positive response on the scale likely means a person gave a positive response to all preceding responses.
5. Thurstone Scales - Judges score indicators. Throw it disagreeable ones. Average out agreeable scores for weight.

Front 12

Conceptualization

Back 12

Mental process where fuzzy and imprecise notions are made more specific and precise.

Creation of empirical statements from your theoretical statements.

Front 13

Indicator

Back 13

Observation we choose to consider a reflection of a variable we wish to study.

Front 14

Concept

Back 14

constructs derived by mutual agreement from mental images

Front 15

Dimension

Back 15

specifiable aspect of a concept

Front 16

3 Different definitions of a concept

Back 16

1. Real Definition - Does not exist
2. Nominal Definition - definition give without any claim that it represents a "real" entity
3. Operational Definition - specifies precisely how a concept will be measured.

Front 17

Progression of Measurement

Back 17

1. Conceptualization
2. Nominal Definition
3. Operational Definition
4. Measurements in the real world

Front 18

Operationalization Choices?

Back 18

1. Range of variation
2. Variation between extremes

Front 19

3 Common pitfalls of indexes and scales

Back 19

1. Referring to indexes as scales
2. Data items that form a scale with one set of observations may not form a scale with another set of observations
3. Use of scaling techniques does not ensure creation of a scale

Front 20

4 Main steps of Index Construction

Back 20

1. Selecting Items
2. Examining empirical relationships
3. Scoring the index
4. Validating the index

Front 21

How do we select items for a scale?

Back 21

1. Face validity
2. Unidimensionality
3. General or Specific
4. Variance - try to select items that divide people equally in terms of the variabe OR select items with different variation.

Front 22

How do we examine empirical relationships of items on a scale?

Back 22

1. Bivariate relationship: If an item doesn't relate to several other measures then drop it. If an item item two items strongly relate then only keep one
2.Multivariate relationship among items: cross tabulations done by computers.

Front 23

How do we score and index?

Back 23

Should we weight things equally or unequally? Only unequally for compelling reasons.

Front 24

How do we handle missing data?

Back 24

1. If only a few cases, then exclude them.
2. May have grounds for treating missing data as one of the responses
3. Careful analysis of missing data may yield meaning
4. Can assign the middle value or mean of range. Conservative because it reduces the likelihood it will relate to other variables in ways you have hypothesized.
5. Assign a proportional value (If test taker answers 2 out of 4 questions positively then give a score of 2/4 for the missing indicator)

Front 25

2 ways of validating an index?

Back 25

1. Item analysis - assessment of each measure to see if it makes an independent contribution rather than duplicating another item.
2. External validation - process of testing the validity of a measure by examining its a relationship to other measures.