measurement, scales and attitudes. nominal ordinal?
TRANSCRIPT
Measurement, Scales and Attitudes
Nominal
Ordinal?
Interval
Ratio
Types of Scales
• Nominal - Identification only– Ex (players numbers, male=1 female=0)
• Ordinal -Ranking– Ex (grades?) SEI score
• Interval –Ranks and distinguishes intervals– Ex (temperature)
• Ratio – absolute quantities– Ex (weight, degrees Kelvin)
You Try
• Which college is your major housed in?1. CLS 2. CBA 3. SAH
• How much money did you spend on alcohol at the bars in the last 7 days?_____Dollars
– Or answer categories1. $0-10 2. $11-20 3. $21 +
Practice Problems
Appropriate Statistical Analysis
• Scale matters.
– Usually
Index Measures
• Conglomerates of questions
• Mapping multiple responses to a single metric
• Consumer Sentiment Index
Good Measurement
• Reliable
• Valid
• Sensitive
Wrong
Validity-reliability bulls eye (Babbie, 1998)
Bothvalid & reliable Reliable,
but invalid
Reliability
• Degree to which measures are free from error
Reliability
• Repeatability– Test-Retest method- High correlation suggests
stability/reliability
Reliability
• Internal Consistency – Split-half method- take results form even
questions and compare to odd number questions– Equivalent form- asking different but equivalent
questions of a group, and comparing their answers on the separate questions
Validity
• the ability to measure that which you intend to measure
• Reliability is a necessary condition for validity, not sufficient– Example - if the scale always reads 5 pounds too
much. It is reliable, but not a valid measure of weight.
Types of Validity
• Content (Face) validity- agreement that a scale accurately measures that which it is intended to
• Criterion validity- the ability of a measure to correlate highly with another measure of the same construct– Concurrent validity- measures made concurrently – Predictive validity- correlates with future measure
• Construct validity- The ability of a measure to confirm a network of related hypothesis
Types of Validity
• Content (Face) validity- agreement that a scale accurately measures that which it is intended to
Types of Validity
• Criterion validity- the ability of a measure to correlate highly with another measure of the same construct– Concurrent validity- measures made concurrently – Predictive validity- correlates with future measure
Types of Validity
• Construct validity- The ability of a measure to confirm a network of related hypothesis
Sensitivity
• a measurements ability to measure variability in stimuli
• Without variability in response we have nothing of interest.– We can classify but not understand or explain
Attitude Measurement
• Attitude – an enduring disposition which contains these components:– Affective (emotional)– Cognitive (reason)– Behavioral (action)
• Example (sexual identity)
• Hypothetical Construct
Male Sex Behavior and Identification
Total 33 10 1,213 1,256 normal/straight 1 0 64 65 something else 1 2 4 7 bisexual 4 3 1 8 homosexual 22 0 2 24 heterosexual 5 5 1,142 1,152 identification exclusive both male exclusive Total rs sexual year gender of sex ps in the last
Col1 exclusively maleCol 2 bothCol 3 exclusively female
Female Sex Behavior and Identification
Col1 exclusively maleCol 2 bothCol 3 exclusively female
Total 1,531 7 19 1,557 normal/straight 87 1 0 88 something else 7 0 5 12 bisexual 4 2 2 8 homosexual 1 0 8 9 heterosexual 1,432 4 4 1,440 identification exclusive both male exclusive Total rs sexual year gender of sex ps in the last
. tab sxident sexsex if gender==0
Male Sexual Attraction and Identification
Total 1,396 43 9 11 37 1,496 normal/straight 74 2 0 0 3 79 something else 5 1 2 2 1 11 bisexual 2 3 5 1 0 11 homosexual 1 1 0 8 17 27 heterosexual 1,314 36 2 0 16 1,368 identification only wome mostly wo both wome mostly me only men Total rs sexual is r attracted to males or females
Female Sexual Attraction and Identification
Total 7 11 15 50 1,803 1,886 normal/straight 1 0 0 0 116 117 something else 3 0 1 1 7 12 bisexual 0 2 5 3 0 10 homosexual 2 9 1 0 0 12 heterosexual 1 0 8 46 1,680 1,735 identification only wome mostly wo both wome mostly me only men Total rs sexual is r attracted to males or females
. tab sxident attract if gender==0
Practice Problems
• Drinking Alcohol
• Religion
• Music
Concept Measurement• Awareness: measure of knowledge; understanding;
familiarity• Behavior: measure of actions/choices that took place• Motivation: measure of why people behave as they do• Opinion: measure of belief or attitude• Preference: measure of likes/dislikes• Desire: measure of wants• Interest: measure of concerns/curiosities• Intention: measure of anticipated behavior• Demographic: measure of respondent’s characteristics• Perceptions of above?
What is Most Appropriate?
Methods of Measuring Attitudes
• Rating– Likert Scale – carefully constructed attitudinal
measure which asks people for their agreement with a statement
– Example: Please rate each of the following, on a scale from 1 to 10, with 1 being least important and 10 being most important…
Methods of Measuring Attitudes
• Ranking– Rank choices from most important to least
important
• Example: Of the following 10 items please rank them in order of importance, with 1 being least important and 10 being most important.
Source: http://blog.vovici.com/blog/bid/18228/Ranking-Questions-vs-Rating-Questions
• Sorting– Asks respondents to sort items/names into groups
• Multiple Choice– Provide alternative responses that R might
choose.
Monadic
Q. How satisfied are you with your current job?
1. Very Satisfied
2. Somewhat Satisfied
3. Not Very Satisfied
Comparative
Q. Compared to your current job how much responsibility did you have at your last job?
1. More
2. About the Same
3. Less
How many categories are needed?
Balanced
Q. How satisfied are you with your current cable service?
1. Very Satisfied 2. Somewhat Satisfied 3. Somewhat unsatisfied 4. Very unsatisfied
Unbalanced
Q. How satisfied are you with your current cable service?
1. Very Satisfied 2. Somewhat Satisfied 3. Somewhat unsatisfied
Forced Choice
• Does not allow the respondent to offer no opinion, which is different than a neutral opinion
Unforced Choice
• Allows respondents to opt out of providing an opinion.
JOHN ALLEN PAULOS
Unless we know how things are counted, we don’t know if it’s wise to count on the numbers.
Albert Einstein
Not everything that can be counted counts, and not everything that counts can be counted.