1. Measurement is the assignment of numbers or other symbols to characteristics of objects according to certain specified rules. Measurement can be scalable, specific and can also include a description, an order, the distance or origin. Measurement can be a unique identifier, a preference of something over another or even a generalization or precise starting point. 2. The basic characteristic of a scale are description, order, distance, and origin, together they define the level of measurement of a scale. The description is the unique identifier that is used for each of the values being measured on the scale. Order can be the size or preference of what is preferred over another. Distance is the difference of what makes what is currently known about what is being scaled to what is being measured. Origin is having a set or known fix point to start what is being measured. 3. The primary scales of measurement are nominal, ordinal, interval, and ratio. Nominal scale is what is labeled to identify or have a description of what is being measured. Ordinal scale is ranking or the positioning of the unit being measured. Interval scale is the measuring of the equal values assigned to the units which allows for comparison between equal units. Ratio is the combination of nominal, ordinal and interval scales, with a starting point of zero that creates a statistical measurement of ranking, identifying and ordering. 4. The differences between a nominal scale and an ordinal scale are as follows. A nominal scale labels or has a description which identified what is being measured while the ordinal scale is the actual ranking or the positioning of what is being measured. 5. The implications of having an arbitrary zero point in an interval scale would impact the ability to measure equal values or differences of what is being measured. This is to remain consistent of each item being measured against each unit, if there was a fixed arbitrary zero point of being measured, no measurements could be completed between the different units being measured. 6. The advantages of a ratio scale over an interval scale is a ration scale has all the needed primary scales of measurements which are nominal, ordinal and interval scales with a fixed zero point. This ratio scale has the advantage that all statistic techniques can be applied with a positive constant and not an arbitrary number which would not allow a ration to be determined. Interval allows for a comparison which is more in general than the hard data presented by a ratio. In an interval the zero point is not fixed, as it is in a ration, which allows for more arbitrary data and scale values so there can be a range of differences. 7. The comparative rating scale is a mainly used in marketing research and involves the direct comparison of stimulus objects or the preference/judgement of what is liked/not liked. 9. What are the advantages and disadvantages of paired comparison scaling? The disadvantage of the comparative rating scale is there can be no generalization of data if comparing object A to object B. The disadvantage continues is the inability to generalize the data further than what is being compared. To compare object C there has to be another comparative rating scale completed. The advantages are the scales are easy to understand and have few assumptions, very basic.
Chapter 14 page 284 1 – 7, 9 and 12 – 19 Exercises

1. What is the semantic differential scale? This is a seven point rating scale with defined end point that are associated with bipolar labels such as hot/cold, young/old, colorless/colorful. For what purposes is this scale used? This is a rating scale used to measure the person respond on people or product concepts or what they mean to the person. 2. The Likert scale is a widely used rating scale that has the responders indicate their degree of agreement or disagreement to a series of statements about a stimulus object. There usually is five response categories from Strongly Agree to Strongly Disagree. The responder will have a more favorable attitude to that which has the higher score of agreement. 3. What are the differences between the Stapel scale and the semantic differential? The differences between the two are the Stapel scale has ten categories numbering from -5 to +5 without the zero (0) in the middle. This scale rates how accurate or inaccurate each unit is rated and does not require a pretest of wording to understand the test and can be done over the phone without visual contact. The semantic differential is a seven point rating scale with endpoint of hot/cold, formal/informal, and colorless/colorful. This scale measures concepts of people and products which respondents can understand. Which scale is more popular? The Semantic differential is to be more popular since many believe the Stapel is confusing visually and difficult to apply to context, it is not used very much. The Semantic differential has wording that is understandable and respondents understand concepts better than ranking numbers when completing surveys. 4. What are the major decisions involved in constructing an itemized rating scale? When constructing an itemized rating scale there are six major decision points that must be discussed. The number of scale categories to use, balanced versus unbalanced scale, odd or even number of categories, forced versus nonforced choices, the nature and degree of the verbal description and the physical form of the scale. 5. How many scale categories should be used in an itemized rating scale? In an itemized rating scale, the suggestion of five to nine should be used. Why? Most people cannot handle more than a few categories due to attention span or time. Depending on what the survey is on and the questions will help ensure the best possible survey can be used. Having an inaccurate portrayal of what is being surveyed will not result in information that is useable. 6. What is the difference between forced and nonforced scales? Forced scales is when people have to make a choice because there is no “not applicable” or “no opinion” or neutral choice to make. Nonforced scales give the option of “not applicable” or “no opinion” or neutral choice to make so many “no opinions” will be generated. Data accuracy is not very good. 7. Should an odd or even number of categories be used in an itemized rating scale? With the odd number of categories used, usually the middle choice will be the no opinion either way. To force the issue better would be to use the even category when one would be force to choose and not have just a middle answer.
9. How do the nature and degree of verbal description affect the response to itemized rating scales? Verbal description has the scale categories of verbal, numerical and even picture descriptions and how to label each scale is a decision process. Being just verbal in the category may not be the best way to proceed but a balance is needed to correctly assign values for the scale to receive the best response.
12. What is reliability? Reliability is when a scale continually produces consistent results, from repeated events, to be reliable.
13. What are the differences between test-retest and alternative forms reliability? Test-retest is when respondents are given the same identical scale at different times under the same conditions usually two to four weeks apart. In the alternative-forms reliability, two equal forms of the scale are used, not identical, and using a different scale.
14. Describe the notion of internal consistency reliability. Internal consistency reliability is used to assess the reliability of a scale where several of the items on the scale are totaled to form a total score. This is used to measure different types of a multidimensional like an image.
15. What is validity? Validity is the differences in an observed scale reflect the true differences of what is being measured rather than random errors being measured.
16. What is criterion validity? Criterion validity reflects if a scale will perform as expected during a time period and can include demographics and psychographic characteristics. How is it assessed? Criterion validity is assessed when data on the sale and the variables from the scale are collected at the same time for example the predicted and actual purchase are compared to predict the validity.
17. How would you assess the construct validity of a multi-item scale? To assess the construct validity of a multi-item scale one is measuring and is attempting to answer theoretical questions of why something works using theory. This scale uses convergent, discriminant and nomological validity for multi-items.
18. What is the relationship between reliability and validity? The relationship is what a true score model is. If the measure is perfectly valid then the results are perfectly reliable and the opposite also holds true for this relationship.
19. How would you select a particular scaling technique? In choosing a particular technique research is the key of what you are trying to measure and the desired outcome. The scale that would result in the highest success rate would be the one I would choose. A single item is more manageable if I am only looking for one aspect to scale.