• 1. Asking questions
• 2. Defining Concepts
• 3. Operationalizing
• 4. Measurement

Abstract ideas that represent or symbolize a quality of the political or social world.

Concepts can be relatively abstract or relatively concrete. The more abstract the more difficult it is to come up with agreement on a definition.

• 1. Social class
• 2. Political participation
• 3. Liberty

The process of coming to an agreement about what concept mean and represent with regards to social and physical reality

Or the mental process whereby fuzzy and imprecise notions are made more specific and precise

The variable empirically captures the variation within the concept

• Concept: education
• Variable: level of education

1. The attributes / values should be exhaustive, covering the whole range of a variable.

Ex. Religious affiliation: the various affiliations should be included. Ex 1. Catholics 2. Protestants 3. Muslims etc.

2. Attributes / values should be mutually exclusive.

Ex. From above, a person is either a Catholic or a Protestant, but not both.

The means by which we assign each individual case to the different values of the variable

• Example:
• Concept: Political participation
• Variable: Level of political participation
• Values: High or low levels of political participation
• Indicator: Attending a political rally

The process of moving from a conceptual definition to a measure (or set of measures) that allow the researcher to empirically observe the construct

1. Conceptualization - we have an agreement about the definition of the concept

2. Designate variables of the concept, and the values on those variables

3. Look for indicators of those variables

A definition that provides a way to measure an abstract concept for all cases in the study by specifying how the concept is to be measured.

• Example:
• For the purpose of the study, trust will be measured based on whether an individual would leave her wallet unattended.

Conceptual definition: acting with the intent of directly influencing the political system

Variable: levels of political participation (high/low)

Indicator: being a member or not being a member of a political party

Measure: ask participants if they are members. Their answer is the indicator of the variable.

1. Whereas a single indicator may misclassify a person, multiple indicators help to offset this problem

2. Multiple indicators can capture finer distinctions of the concept present in the subject

NOIR

Nominal: variable with just discrete categories that cannot be ranked. Difference in kind, not degree.

Ex.religion- Protestant, Sikh, Muslim etc.

Ordinal: discrete categories that can be ranked, but with no fixed distance between the attributes of the variable

Ex. Low, medium and high (social class, conservatism, prejudice satisfaction).

Interval: can be logically ranked, and in which the distance separating the attributes has a meaning and is consistent. No "real" zero value

Ex. Celcius scale, IQ score

Ratio: has all the characteristics of the interval measure and a true zero point. Can draw mathematical inferences between attributes of the variable

Ex. Number of TV's, Kelvin scale, income

1. Precision: How fine are the distinctions between the attributes composing a variable

2. Accuracy: how much the measure reflects the actual attribute of a variable.

3. Reliability: whether a particular measurement yields the same results when applied repeatedly to the same object or person

4. Validity: the extent to which an empirical measure adequately reflects the agreed-upon meaning of the concept.

Are you measuring what you think you are measuring.

High validity: measuring economic wealth based on income and assets

Low validity: measuring love for your parents as time you spent on the phone with them

Every time you step on the scale it weighs 50 lb. But you know you weigh around 170 pounds. The scale may be reliable, but it is not valid

1. Inductive reasoning: using empirical evidence to help form the definition of a concept

2. Extrapolation: using deductive reasoning by identifying one or more postulates, then deriving expectations and conclusions based on the claims.

• 3. Intuition: thinking about a concept in a way no one else has previously done.
• Be careful to check the literature for examples of other research that used similar conceptual definitions.
• Explain why yours is superior.

Abstract concept: liberty

Concrete concept: region

Some degree of meaning or understanding becomes lost because we are limited by our variables to those aspects of the concept that can be measured empirically.

• Ex. The concept of benefits of university education.
• You might be a more thoughtful person after completing your education, but this is a variable that cannot be easily measured.

Use multiple variables to cover more of the concept

Ex. The main variable in political participation is voting, but other variables such as lobbying, party membership, protest activity and Community participation can also help the fill in the gaps

1. How have other researchers measured this variable?

2. Are your indicators consistent with the literature in the area?

1. Face validity: on the face of it, does the measure logically measure the concept and can you justify your selection of the measure?

2. Convergent validity: When comparing indicators designed to measure the same variable, the two indicators should yield similar results for most cases.

3. Discriminant (divergent) validity: if two indicators predict opposing views, they should yield different results for most cases.

4. Predictive validity: does the use of the measure help us to predict outcomes? To test predictive ability, we need to Pilot test the measure in an appropriate population.

To test the reliability and validity of the measures

the survey questions themselves.

• Close questions
• 1. responders can answer quickly
• 2. easy to compare the responses of different individuals
• 3. data entry is less complex

Disadvantage: these questions often asked the respondent to give a simple response to an extremely complex issue

1. Use neutral language: (do not indicate bias)

2. Be clear:

3. Avoid response sets: (questions that force respondant to move back and forth between agreement and disagreement).

4. Keep response categories mutually exclusive and exhaustive: ensure that all relevant responses are included including 'no opinion'.

Also make sure there is no overlap between categories.

5. Select the highest reliable level of measurement: when possible, we should use interval level measures above ordinal level measures

6. Pay close attention to question order: preceding questions which provide context may contaminate what comes after

7. Minimize defensive reactions: questions about specific income could make respondents defensive, and more likely to lie.

A complex measure that combines responses from more than one question in the creation of a new measure.

• Ex. Orientation to feminism:
• is an index based on several questions such as " it is a woman's right to choose an abortion" and " lesbians and homosexuals should not be allowed to teach in school" etc.

See page 113 for example.