Surveys...Scale development--- DeVellis,1991

Scale use

Consider the latent variable

Consider the manifest variable

The more the better

Reading difficulty

Writing the scales:

Avoid double barreled items.

"I support civil rights because discrimination is a crime against God."

Ambiguous pronoun references

" Murderers and rapists should not seek pardons from politicians because they are the scum of the earth."

Misplaced modifiers

"Our representatives should work diligently to legalize prostitution in the House of Representatives."

Positively and negatively worded items.

You can use both to keep the respondent conscious of what they are reading. But, you may also create confusion for the respondent.

Focus is on a single affirmative response

Scale formats.

Thurstone:

You can have items that correspond to different intensities of an attribute, spaced to represent equal intervals and could be formatted with agree -disagree responses.

Calibration of answers is possible

Finding items that rise to specific levels of a phenomena is difficult.

Guttman scale:

The focus is on a transition from affirmative to negative responses .

Here when you respond positively to one level of a hierarchy, the implication exists that you satisfy the criteria of all lower levels of that hierarchy.

Ie.. Do you smoke? Do you smoke more than 10 cigarette? Do you smoke more than a pack a day?

Cause and effect can not be assessed in these two cases.

Likert statements: Often used to measure opinions, beliefs and attitudes.

You must state the opinion, attitude, belief, or other construct under study in clear terms.

Declarative sentences---followed by response options that indicate varying degrees of agreement with or endorsement of the statement.---responses should have roughly equal intervals with respect to agreement.

Six responses is not uncommon.

Strongly disagree, moderately disagree, mildly disagree, mildly agree, moderately agree, and strongly agree.

Semantic differential: Used with one or more stimuli.

Here we identify the target stimulus and follow it with a list of adjective pairs.

Each pair represents opposite ends of a continuum.

Stimuli:

Reporters

Honest ___ ___ ___ ___ ____ ___ ___ Dishonest (Unipolar set)

Quiet ___ ___ ___ ___ ____ ___ ___ Noisy (Bipolar)

You may use bipolar adjectives that express opposite attributes, such as friendly and hostile

Unipolar adjective look at a single attribute.

Visual Analog

No Pain at All ___________________The worst pain I ever experienced.

Single item....Hard to test reliability

Binary options...limited variability ...Two answer... Variability limited

Response options:

How many is enough?

Open ended

Closed ended

If a scale fails to discriminate differences in the underlying attribute, its correlations with other measures will be restricted and its utility will be limited.

Variability can be increased with increasing scale items.

Variability can be increased with numerous response items.

Respondents may not be able to understand distinctions between responses.

Often to fix this you may present the responses in a continuum.

Odd and even number of responses.

Lay out of questionnaire is important and can help the ambiguity in responses.

Get an expert to check the scale.

Mild statements are not very useful.

Survey questions

Use simple language..

Long vs short questions.... Clarity is the key

Double barreled questions

Double negatives.. Avoid

Length of lists....

Recalling things from the past.....cuing

Hypothetical questions

Specificity of questions.

Open vs closed ended.

Filtering questions..

Question ordering

Multiple questions for a topic

Page layout..white space

Pretesting

Checking for variation...meaning..task difficulty..respondents interest and attention....

Flow and naturalness of the sections

Order of questions

Skip patterns..

Timing

Respondents interest and attention

Respondents well being.