Measurements and Scale
Radio Data example
-What is your annual household income? $____________
Coefficient alpha (most common test of reliability)
average of all possible split-half coefficients resulting from different splittings of the scale items. -"Nunnally (1978) Fish > .7 reliable
Number of scale positions (or points)
-Most common numbers are 5 & 7. -Some scales go as low as 3, and others as high as 21. -It depends upon: -Level of specificity required to test the hypothesis, and - level of specificity that the target respondents can reasonably provide.
reliability
extent to which a scale produces consistent results.
Test-retest Reliability
Respondents are administered scales at 2 different times under nearly equivalent conditions. "Do it twice at different times" "Double the work of respondent"
Measurement
Rules for assigning numbers to objects to represent quantities of attrubutes.
Construct Validity
addresses the question of what construct or characteristic the scale is actually measuring.
The relationship between Reliability and Validity is?
a scale can be reliable (1st), but NOT valid In order for a scale to valid, it must also be reliable. in other words: -RELIABILITY is a necessary but insufficient condition for VALIDITY.
Reliability vs. Validity
- old rifle: Neither reliable nor valid (inconsistent) -New Rifle that is sighted poorly: Reliable but not valid (consistant) -New Rifle that is sighted accurately: Both reliable and Valid
Semantic Differentail Scales
-A series of numbered (usually 7 point) bipolar rating scales. "Answer on each line" -Takes up lots of room, have to answer on each line. -Biopolar adjectives (Good, and Bad), anchor both ends (or poles) of the scale. -A weight is assigned to each position on the rating scale. -Scores are: 7, 6, 5, 4, 3, 2, 1 or +3, +2, +1, 0, -1, -2, -3.
other considerations
-Consider including a "Don't Know" or "Not applicable" Response Category -Doing so reduces the degree of item Non-Response Bias, and -provides a better understanding or respondents' answers. -odd or even number of categories? (Common in odd) -Verbal Decription-how in-depth? (tell people how to use the scale) m -Physical From to use? (left or right, below lots of space)
The Likert Scale
-Extremely popular means for measuring attitudes. "Abstract Constructs" -Respondents indicate their own attitudes by checking how strongly they agree/disagree with statements. -Response alternatives: -strongly agree, agree, uncertain, disagree, strongly disagree. -Gneraly use either a 5- or 7-point scale -Psychographic, brand based on attitude, brand loyalty.
Continuous Rating Scale Example
-Left of scale negative, Right of scale positive -Cant measure, used in political polls -Not used often, very quickly, lack of perspective general idea. Ex: I believe the lvl of services offered by Sirius XM Radio is___________?
Paired comparison Scaling
-Respondent is presented with 2 objects at a time -Then asked to select 1 object in the pair according to some criteria. -Data obtained are ORDINAL in nature -Arranged or ranked in order of magnitude -Easy to do if only a few items are compared. -If numbers of comparisons may become fatigued and no longer carefully discriminate among them. -Dont know the degree of that preferences
Constant Sum Scaling
-Respondents are asked to allocate a constant sum of units among a set of stimulus objects with respect to some criterion. -Units allocated represent the importance attached to the objects. -Data obtained are interval in nature -Allows for fine discrimination among alternatives. -Hard to do, assign points based on criteria (100)
Rank order scaling
-Respondents are presented with several objects simultaneously -Then asked to order or rank them according to some criterion. -Data obtained are ordinal in nature --Arranged or ranked in order of magnitude -Commonly used to measure preferences among brands and brand attributes Ex: rank football teams in order 1,2,3.
Number of items in an Itemized rating scale
-The number of items (questions) in many itemized rating scale is 3 to 5, though many have up to 20. -At least 3, common 5 or 6 -It depends on the construct being measured.
Ratio Scale Example
-Zero has an absolute meaning -Zero = Absence of the property being measured -What is your credit card balance? $______________
Interval Scale Example
-Zero is just another scale position -Zero = scale point between -1 & 1. -What is your attitude toward credit cards? -3,-2,-1,0,+1,+2,+3 OR 0 1 2 3 4 5 6 7
Stapel Scales
-they place a single adjective as a substitute for the semantic differential when it is difficult to create pairs of bipolar adjectives. -Advanatge and disadvantage of a Stapel Scale, as well as results, are very similar to those for a semantic differential. -Tends to be easier to conduct and admin. -Fast Telephone research, takes space up. Ex: Department Store Name: +3,+2,+1, or -1,-2,-3
Alternative-form Reliability
2 equivalent forms of scale are constructed, then tested with the same respondents at 2 different times. -2 versions of scale, 2 different times
Concept (or Construct)
A generalized idea about a class of objects, attributes, occurrences, or processes.
Scale
A quantifying measure-a combo of items that is progressively arranged according to value or magnitude. Purpose of quantitatively represent an item's, person's, or event's place in the scaling continuum.
Relatively CONCRETE constructs
Age, gender, number of children, education, income. "Easy to measure, single measure & yr they were born in"
Content Validity
Aka: Face validity, "Valid on its face" Subjective, but systematic evolution of the representativeness of the content of a scale for the measuring task at hand (Satisfaction)
Relatively ABSTRACT constructs
Brand loyalty, personality, channel power, satisfaction. "Multiple questions, hard to measure"
The scales of measurement
Data collected at higher levels can be represented at lower levels; however, data collected at lower levels CANNOT be represented at higher levels. -Need: Hypothesis & level of Data Nominal, Ordinal, Interval, & Ratio
Unbalanced Scorecard
Ex: Surfing the Internet is: ____Extremely Good ____Very Good ____Good ____Somewhat Good ____Bad _____Very Bad -Expect answers to be on 1 side than the other. More positive than negative Ex: more negatives for cell phones for seniors, more positives for cell phones for college students.
Balanced Scorecard
Ex: surfing the internet is: _____Extremely Good _____Very Good _____Good _____Bad _____Very Bad _____Extremely Bad
Criterion Validity
Examines whether measurement scale performs as expected in relation to other variables selected as meaningful criteria I.E., predicted and actual behavior should be similar Results from scale, actually measuring behavior? "What people are actually going to do"
Discriminant Validity
Extent to which a measure DOES NOT correlate with other constructs from which it is supposed to differ. Ex: no correlation like hair and lawn mower.
Nomological Validity
Extent to which scale correlates in theoretically predicted ways of measures of different but related constructs. Ex: trust & sastifaction (related but close)
Convergent Validity
Extent to which scale correlates positively with other measures of the same construct.
Validity
Extent to which true differences among the objects are reflected on the characteristic being measured
Reverse Scoring
In a multi-item scale, positively wording some of the questions, and negatively wording others.
Graphic Rating Scale
It presents respondents with a graphic continuum. For children, special needs or poor eye sight people. (Takes up space)
Ordinal Scale (Order)
Measurement in which numbers are assigned to data on the basis of some order (ex: more than, greater than) of objects. -also take the properties of nominal scale (ID) -Brand preference: Rank the following mascara brands. (1=Most preferred, 4=Least preferred ) -Order of Magnitude: but not the degree of magnitude -Cannot determine the exact amount (to the dollar, don't know degree of magnitude) Average: Median (midpoint), Mode
Nominal Scale (identification)
Measurement in which numbers are assigned to objects or classes of objects solely for the purpose of identification. -Gender: 1=female, 2=males -Brand Purchase:Did you purchase Cover Girl mascara? 1=yes, 2=No -Average: Can compute Mode (value occurring most frequently)
Interval Scale (Rating)
Measurement in which the assigned numbers legitimately allow the comparison of the size of the differences among and between numbers. -also take on the properties of nominal (ID) and ordinal (order) scales -Average: Mean (ARITHMETIC AVERAGE), median, mode
Ratio Scale
Measurement that has a natural, or absolute zero and therefore allows the comparison of absolute magnitudes of the numbers. -Can be put into categories,( what is your annual income? $_______________________ -Also take the properties of nominal (ID), ordinal (rank order), and interval (ratings) scales. -Units sold: How many cases of each of the following mascaras were sold last month? -Avg: Geometric & Harmonic Means, Mean, Median, Mode
internal consistency reliability
consistency with which each item represents the construct of interest used to assess the reliability of a summated scale
Comparative Scales
involve the direct compassion of 2 or more objects against 1 another.
Split-half reliability
items constituting the scale divided into 2 halves, and resulting half scores are correlated
Non comparative scales
objects or stimuli are scaled independently of each other.