MKT ch 6 scales of measurement

Interval

Metric. DIFFERENCES (INTERVALS) BETWEEN SCALE VALUES HAVE MEANING AND THEY ARE EQUAL. This is the primary difference between an ______ scale and an ordinal scale. A PARTICULAR SCALE MARK HAS NO ABSOLUTE MEANING Although an _______ scale reveals the differences between the scale marks, the scale values or marks have no absolute meaning. (winner has ten more beauty traits than first runner up, who has ten more than second runner up; but (0) and (1) and (2) have no absolute meaning numerically. NO TRUE ZERO Ex: the Fahrenheit temperature thermometer

Semantic Differential

Respondents are asked to check which blank or cell between a set of bipolar adjectives or phrases best describes their feelings.

Percentage

Ss are asked to indicate the percentage (from 0 to 100) that something exists. DOES have a true zero, as well as having known, equal distances between the scale marks.

Staple

The _____ scale differs from the semantic differential in that: 1.) adjectives or descriptive phrases are tested separately instead of simultaneously as bipolar pairs 2.) points on the scale are identified by number 3.) there are 10 scale positions

Nominal

The ______ scale (also called a Categorical scale) arbitrarily/randomly assigns a number to each response. Scale number has NO meaning. NO obvious relation between the quantity of the construct being measured and the numerical value assigned to it. Ex. Your marital status? (construct is marital status) (1) Single __ (2) Married ___ ----or---- (1) Married ___ (2) Single ___ Ex. "Check the category that best describes your occupation" (1) Unemployed__ (2) Part-time __ (3) Unskilled blue collar ___ (4) Skilled blue collar ___ (5) Office/clerical/sales ___ (6) Manager ___ (7) Professional ___ *****If the construct were social class instead, it would be "operationalized," and occupation would be a key indicator of social class. The scale would then be ordinal.****************8

Paired Comparison

The advantage of the _____ _______ is the ease of the respondent's task in evaluating only two objects at a time. However, the serious drawback of this technique is even when the number of objects is modest (e.g. comparing 15 auto models), a disproportionately large number of pairs have to be compared (If there are 15 objects, a factorial equation yields 105 pairs!!)

Metric

When we know the exact difference between the scale marks, as with an interval scale, this is the key characteristic of a ______ scale.

Likert Summated-Ratings

With this scale, researchers write a number of statements that relate to the issue or object in question. Respondents are asked to indicate their degree of agreement or disagreement with each statement in the series. The response categories represent various degrees of agreement and are assigned scale values. A total score can then be calculated for each respondent by adding/summation or averaging the scores across iteams.

Ratio

_____ is the highest order scale, and only has TWO characteristics. 1. DIFFERENCES BETWEEN SCALE VALUES ARE KNOWN AND EQUAL, which is also true of the interval scale. 2. TRUE ZERO. Differentiates a ____ from an interval scale. ***In other words, a ____ scale is an interval scale with a natural origin.**** Ex. "How many TV's in your residence? _____ " respondent can put "0" and it be a true zero, a total absence of the construct.

Order

_____ refers to the level of sophistication of the data resulting from a scale. Thus, the higher the ____ of the scale type, the higher level or more sophisticated the statistical analysis one will be able to perform on the resulting data. Likewise, a lower ____ scale will generate data with which a researcher CANNOT do very much statistically or analytically.

Bipolar

_____ scaling technique is anchored by words or indicators at the far ends. Having an odd-point scale allows for neutral responses (use even-point if you want to force respondents to lean more or less towards a particular way)

Ordinal

______ scaling techniques. 1.) RANKING 2.) FORCED RANKING (alternatives are given, asked to rank according to construct i.e. "preference" of soda). 3.) PAIRED COMPARISON 4.) SEMANTIC SCALE 5.) PICTURE SCALE

Nominal

_______ scaling techniques: 1.) Multiple Choice 2.) Binary (always "yes" or "no")

Interval Scaling Techniques

1. Bipolar 2. Semantic Differential 3. Likert Summated-Ratings 4. Staple 5. Graphic Ratings

Ratio Scaling Techniques

1.) Percentage 2.) Direct Quantification 3.) Constant Sum 4.) Reference Alternative

Semantic

A ______ scale has only words (no numbers) for the response categories. The scale is ______ because respondents are choosing their responses one the basis of words. Because we never know the difference between words, the _____ scale is ordinal, rather than interval.

Ordinal

All rankings are ______, but not all ______ scales are rankings.

Picture Scale

An alternative to a semantic scale is a graphical scale. Such scales are particularly useful for children and those populations where literacy is low. Ex. smiley scale

Paired Comparison

In a _____ _________, the respondent will be asked to compare only TWO objects at a time. For each pair, the researcher will count the number of times each brand was selected, and then arrange the brands in rank order of preference.

Order of the Measurement Scales

LOWEST ORDER (NONMETRIC--- the differences between the scale marks are unknown and/or unequal) a. Nominal (lowest order scale) b. Ordinal (second lowest) HIGHEST ORDER (METRIC ---- which means known, equal distances or intervals between the scale marks) a. interval b. ratio

Quantitative

Measurement and scaling, by definition, connotes ________ measurement (numbers). When assigning numbers to such a scale, it is a good idea to assign the highest number to the most positive category.

Ordinal

Nonmetric. With a _______ scale, THE HIGHER THE NUMBER, THE MORE (OR LESS) THE CONSTRUCT EXISTS. This is the main difference between ______ scales and nominal scales. ABSOLUTE SIZE OF THE NUMBER HAS NO MEANING. The winner of the beauty contest won, but her value (here it is 0), does not say she's "one" whole value more beautiful than the first runner up (1). However, THE NUMBERS HAVE MEANING IN THAT AS THE SCALE NUMBERS BECOME GREATER, THE CONSTRUCT OF BEAUTY DECREASES. But we don't know to what extent. DIFFERENCES BETWEEN SCALE VALUES HAVE NO MEANING. It is impossible to determine how much more or less one scale value is relative to another. Ex: classic beauty contest

Graphic Ratings

On these scales, Ss indicate their rating by placing a check at the appropriate point on a line that runs from one extreme of the attribute to the other. The actual scale value would then be inferred by measuring the length of the line from the left origin to the marked position.

Interval

The following are examples of ________ scales: 1. how many tv's are in your residence? (1) 0-1____ (2) 2-3____ (3) 4-5____ (A person could not have a tv and mark (1) but we still wouldn't know if they had no tv or just one. plus, the scale marks have equal, known distances) 2. What is your overall attitude toward Ruby Tuesday? Very Favorable___:___:___:___:___:Very unfavorable (When assigning numbers to such a scale, it is a good idea to assign the highest number to the most positive category.) ****Even if "very unfavorable were scaled with a "0," that is not a true zero. There is not such a thing as a zero in the construct of attitude.

Ordinal

The following scale is ___________. "Your age?" (1) less than 35 ___ (2) 35-55 ___ (3) over 55 ____ ***Scale characteristics: the scale is not in equal intervals, therefore, it cannot be interval. The scale numbers have meaning in that as you go from (1) to (3); the construct of age is increasing, Therefore, it cannot be a nominal scale.

four

There are _____ measurement scales/ ways to scale the responses to a fixed question. Each type yields its own particular type of data. So there are ___ types of measurement scales are ____ types of data.

Constant Sum

This scale has a natural starting point (zero) and asks Ss to allocate a given set of points among several attitude objects. (Ex. allocate 100 points among the following, like Dr. Shanahan did with our project)

Reference Alternative

This scale is very similar to the Constant Sum scale in that it requests Ss to allocate points among objects based on how they feel toward them. It is different from the constant sum in that it: 1.) it gives a reference point (such as 100) 2.) the sum does NOT have to total a certain amount Ex. "Given that Harvey's = 100 pts, assign points to the other restaurants listed"