Tải bản đầy đủ
13…Types of Attitude ScalesTypes of attitude scales

13…Types of Attitude ScalesTypes of attitude scales

Tải bản đầy đủ

4.13

Types of Attitude Scales

129

contains a number or a brief description about a particular category. These categories are ordered in terms of position of the scale and respondents have to select
one category that they feel best describes the object. It is easy to develop itemized
category scales. Exhibit 4.7 gives an itemized category scale where a hotel customer is asked to indicate the level of satisfaction for service provided.
Exhibit 4.7: Itemized Category Scale
Given below is a itemized category scale ranging from highly satisfied to highly unsatisfied.
Please select one of the following options based on your satisfaction levels of the hotel service.

Highly Satisfied

Considerably Satisfied Reasonably Satisfied

Unsatisfied Highly Unsatisfied

Adapted from Tull and Hawkins (1998), p. 380.

4.13.3 Rank: Order Scales
Rank order scales are comparative scales, where the respondent is asked to rate an
item in comparison with another item or a group of items against each other based
on a common criterion. For instance, a respondent may be asked to rank three
motorcycle brands on attributes such as cost, mileage, style, pick-up and so on.
Although it is easy to develop a rank-order scale, it has some disadvantages. It is
very difficult to include every possible brand or attribute on a scale. Therefore, a
respondent may rate a brand as number one, but it might not be his first choice as
the brand he prefers may not have been included in the list at all. Sometimes,
respondents may feel that the attributes used to construct the scale are not relevant
to judging the subject under research. One major shortcoming is that the researcher
will not have any clue about why the respondent has given a particular rating for
items listed on the scale. Exhibit 4.8 shows a rank order scale for ranking different
brands on motorcycle on specified attributes.
Exhibit 4.8: Rank Order Scale used for Analysing Motorcycles
Please rank the following brands of motorcycles with 1 being the brand that best
meets the characteristic being evaluated and 7 being the worst on the characteristic being evaluated. Let us now start rating these brands basing on their
affordability, first. Which brand has the highest affordability? Which is second?
(Record the answers below).

130

Brand of motorcycle
Hero Honda
TVS
Bajaj

4 Scales and Measurement

Affordable cost

High mileage

Stylish

Great pickup

Adapted from Tull and Hawkins (1998), p. 384.

4.13.4 Q-Sort Scales
When the number of objects or characteristics to be rated is very large in number,
it becomes difficult and tedious for respondents to rank order. In such cases, Q-sort
scaling is used. Here, respondents are asked to sort out various characteristics or
objects that are being compared into various groups so that the distribution of the
number of objects or characteristics in each group follows a normal pattern. For
instance, let us consider that the designing team of a toy manufacturing company
has come out with hundreds of new product ideas with slight variations. The
research team’s task is to find out from customers which combination of features is
the best and will generate maximum sales. To accomplish this, Q-sort scaling is
the best method. The procedure followed is:
Respondents are given a set of cards, usually varying from 80 to 120 cards,
containing different categories of items to be selected from. For instance, if
respondents have to rate 100 different products according to their tastes and
preferences, each respondent will be given about 100 cards containing a product
and its features. Respondents are then asked to segregate the cards into 10 stacks
so that the first stack contains a set of cards that are highly preferred by respondents. The 10th stack will contain a set of cards that are least preferred by them.
The individual stacks in between (2nd and 9th) should be prepared by the
respondent in such a way that they range from higher preference to lower preference. Once the stacks are ready, the cards in each stack should be arranged in the
respondents’ order of preference, based on criteria like features of a product,
communication processes and customer service. This gives the best and the worst
product in each stack. The disadvantage of this process is that it asks a lot of time
and effort on the part of respondents.

4.13.5 Comparative Scales
In the itemized category scale, respondents select a category that they feel best
describes a product. The problem here is that respondents may select a category
based on their own perceptions. For instance, respondent A might select a category
based on his or her view of an ideal brand, respondent B may pick a brand based
on knowledge of an existing brand and respondent C might choose based on some

4.13

Types of Attitude Scales

131

other criteria. Ultimately, the selection process lacks uniformity. To overcome this,
comparative scales have been developed, where the researcher provides a point of
comparison for respondents to provide answers. Therefore, all respondents will
have a uniform point of comparison for selecting answers. For instance, rather than
asking a person to evaluate the quality of sweets in one sweet shop in Hyderabad,
the respondents will be asked to evaluate the quality of that sweet shop in comparison with another sweet shop in Hyderabad. Exhibit 4.9 gives a comparative
rating scale.

Exhibit 4.9: Comparative Scales
Given below is the scale ranging from excellent to very poor. If you were asked to
rate the sweet shop ‘X’ in comparison to sweet shop ‘Y’ in Hyderabad. Which one
will you choose. If you choose excellent then select the first option.

Excellent

Very Good

Good

Both are same

Poor

Very poor

4.13.6 Paired Comparison Scales
In paired comparison scales, respondents are asked to select one of two items in a
pair based on preset criteria. As each item is compared with all other items, the
number of times an item is selected from a pair gives its rank. The higher the
number, the better is the rank. In this method, the shortcoming of rank order
scaling is overcome, as it is easy for respondents to select one item from two rather
than ranking a long list of items. Another advantage is that the problem of order
bias is eliminated as no set pattern is followed while providing respondents the
pairs. A typical paired comparison scale for toothpaste is shown in Exhibit 4.10.

Exhibit 4.10: Paired Comparison Scale for a Toothpaste
Please select one item each from the following pairs that is most important to
you for selecting a toothpaste.
a. Fights decay
a. Affordable
a. Longer germ protection

b. Affordable
b. Longer germ protection
b. Fights decay

132

4 Scales and Measurement

4.13.7 Constant Sum Scales
In constant sum scales, respondents are asked to divide a given number of points,
usually 100, among two or more attributes based on the importance they attach to
each attribute. These scales are often used in place of paired comparison scales to
eliminate the long lists in paired comparisons. Here, respondents have to rate an
item in relation with all other items. Ranking for each item is based on the points
assigned by the respondent to the items. The disadvantage of this approach is that
the researcher is limited to giving 10 items for the respondent as a higher number
of items will confuse the respondent. Exhibit 4.11 shows the constant sum scale
where respondents are asked to rate 10 characteristics of a supermarket for a total
sum of 100 points.

Exhibit 4.11: Constant Sum Scale Used for a Supermarket
Given below are the ten characteristics of a supermarket. Please give each
characteristic some point(s) based on your assessment, so that the total points
add up to 100. The higher number of points allocated to a particular characteristic, the higher its importance to you. However, if any particular characteristic is unimportant to you, then you need not assign any points to it.
However, it is essential that all points given add up to 100.
Characteristics of a supermarket
The supermarket is conveniently located
The supermarket has enough range of products
All the items in the store are conveniently located
Sales persons are cooperative
Aisle space is comfortable
Prices are very much affordable
The ambiance in the store is pleasing
Soft music played in the store is entertaining
Billing counters are sufficient
Parking facility is adequate

Number of points
_______
_______
_______
_______
_______
_______
_______
_______
_______
_______
________
100 points

4.13.8 Pictorial Scales
Here, the different types of scales are represented pictorially. The respondents are
asked to rate a concept or statement based on their intensity of agreement or
disagreement, on a pictorial scale. Pictorial scales have to be developed carefully
so that respondents will not have problems selecting appropriate responses. These

4.13

Types of Attitude Scales

133

scales are generally used for respondents who cannot analyse complex scales, such
as young children or illiterates. Typical pictorial scales are a thermometer scale or
a scale depicting a smiling face. Exhibit 4.12 shows the smiling face scale for
measuring the effectiveness of an advertisement campaign for a chocolate.

Exhibit 4.12: Smiling Face Scale

(Verbal Instructions)
Face 1 should be selected, if you did not like the chocolate ad at all and face 5 should
be selected if you liked it very much, now how did you like it?

Adapted from Neelankavil et al. (1985)

4.13.9 Continuous Scales
Continuous scales are those where respondents are asked to rate items being
studied by marking at an appropriate place on a line drawn from one extreme of
the scale to the other. These scales are rarely used in marketing research as they do
not give accurate results and the scoring process is complicated. This scale’s only
advantage is that it is very easy to develop. For instance, if a fast food outlet such
as Pizza Hut wants to find out whether customers are satisfied with its overall
service, then a continuous scale can be developed as shown in Exhibit 4.13

Exhibit 4.13: Continuous Rating Scale
Given below is a continuous scale ranging from 0 to 100 points. You have to indicate a point that
best describes how you rate the overall service. If you rate it the best then it would be 100.
How would you rate the overall service of Pizza hut?
Best -------------------I-------------------------------------------------------------------------------- Worst
100

90

80

70

60

50

40

30

20

0

134

4 Scales and Measurement

4.13.10 Multi-Item Scales
Let us move now to multi-item scales. These scales are used when it is difficult to
measure people’s attitudes based only on one attribute. For instance, to measure
respondents’ attitudes towards the Indian Railways, if you ask them only whether they
are satisfied with Indian Railways or not, it will not suffice. People may say that they are
satisfied on an overall basis, but there might be number of factors that they find
unsatisfactory. Thus, it is impossible to capture the complete picture with one overall
attitude-scale question. To measure individual attributes, a number of scales have been
developed that can measure a respondent’s attitude on several issues on a scale ranging
from most favourable to least favourable. The Semantic, Likert, Thurstone and differential scales are some examples that follow such measurement techniques.
Developing multi-item scale involves certain crucial steps that have been discussed in
Exhibit 4.14.

Exhibit 4.14: Developing a Multi-item Scale

Exhibit 4.14
Initial Development of a theory

Generating a number of Items

Selecting a final set of Items with the Help of
experts

Identifying the appropriate attitudinal scale to
be used

Adding some items for improving the validity
of the scale

Administering the items to an initial set of
respondents

Analyzing the Items

Optimizing the Final Scale Length

4.13

Types of Attitude Scales

135

4.13.11 Semantic Differential Scales
Semantic differential scales are used to describe a set of beliefs that underline a
person’s attitude towards an organization, product or brand. This scale is based on
the principle that individuals think dichotomously or in terms of polar opposites
such as reliable/unreliable, modern/old fashioned, cold/warm.
The respondents are asked to rate an attitude object on a set of itemized, sevenpoint rating scale, bounded by bipolar phrases or adjectives. The initial process of
developing a semantic differential scale starts with determining the object to be
rated. Once this object is determined, respondents are asked to express their
attitudes towards the object, using the dichotomous pair on a scale. Such points are

Exhibit 4.15: A Semantic Differential Scale for Measuring the Attitudes
of Respondents for a Newspaper
Given below is a semantic differential scale, where a respondent has chosen his
options on a 7 point scale based on 9 bi-polar categories.
1 2 3 4 5 6 7
Contemporary

. . . . . . .

Old Fashioned

Comprehensive
News Coverage

. . . . . . .

News coverage is very
limited

. . . . . . .

Uninteresting Supplements

. . . . . . .

Unbalanced coverage of news

Interesting
Supplements
Balanced coverage of
news
Quality of language is
high

. . . . . . .

Quality of language is
poor

Too much
international
coverage

. . . . . . .

Too less international
coverage

Excellent Editorial Articles . . . . . . .

Worst Editorial Articles

Appealing to youth

. . . . . . .

Not at all appealing to
youth

Not inclined towards any
political party

. . . . . . .

Inclined towards a political
party

136

4 Scales and Measurement

then plotted on a graph. This is the most efficient technique for determining the
strengths and shortcomings of a product/service or a company in the market.
While designing the scale, care should be taken that all negative or positive
adjectives or phrases do not appear on one side. This avoids a person from picking
either only positive or negative phrases.
Another problem that should be addressed while developing a seven-point
semantic scale is the response of four. If the respondent selects four for all items,
then it becomes neutral without indicating any specific direction. Exhibit 4.15
represents the semantic differential scale administered for measuring the attitudes
of respondents towards a newspaper. It can also be used for comparing the
products with that of the competition. Consider four brands of cars being rated on
the same scales, as shown in the Exhibit 4.16.

Exhibit 4.16: Semantic Differential Scale for Comparing Four Brands of
Cars
Below given is the semantic differential scale rated by a responding by comparing 4
brands of cars.
Mitsubishi (L)
Skoda
Lancer
Octavia (O)
Hyundai (E)
Elantra
Fast
Large

EL

Honda (C)
City

O
L

C

Slow

E

O

C

Plain

E

L

O

Inexpensive

C

L

E

Small
C
O

Stylish
Expensive

Adapted from Tull and Hawkins (1998), p 390.

4.13.12 Stapel Scales
A Stapel scale is an attitude measure that places a single adjective or an attribute
describing an object in the centre of an even number of numerical values. In
general, staple scales are constructed on a scale of 10 ranging from -5 to +5,
without a neutral point (zero). The respondent is asked to rate attributes on this
scale.

4.13

Types of Attitude Scales

137

Stapel scales are similar to semantic differential scales, but here there is only
one pole (single adjective) rather than bipolar adjectives. This scale is useful for
researchers to understand the positive and negative intensity of attitudes of
respondents. The numeric value assigned to an adjective shows how well it
describes the object. The higher the positive value, the better it describes the
object.
One big disadvantage is that the respondent might select all attributes on a
positive or negative range. A Stapel scale that is designed to measure the attitude
of passengers towards an airline is shown in Exhibit 4.17.
Exhibit 4.17: Stapel Scale for Measuring the Attitudes of Flight Passengers
Below given is a staple scale designed to measure your attitude on three
attributes. Please circle one number from the following three columns that
best describes your attitude towards them.
+5
+4
+3
+2
+1
Friendly cabin crew
-1
-2
-3
-4
-5

+5
+4
+3
+2
+1
Comfortable interiors
-1
-2
-3
-4
-5

+5
+4
+3
+2
+1
Accurate timings
-1
-2
-3
-4
-5

4.13.13 Likert Scales
Likert scales consist of a series of statements where the respondent provides
answers in the form of degree of agreement or disagreement. This expresses
attitude towards the concept under study. The respondent selects a numerical score
for each statement to indicate the degree of agreement or otherwise. Each such
score is finally added up to measure the respondent’s attitude. The various steps
involved in developing a Likert scale are given below.
• Identify the concept that needs to be measured
• Develop a series of statements (say, 100) that articulate respondents’ feelings
towards the concept
• Every test item is categorized by the respondent as generally favourable or
unfavourable based on the attitude that needs to be measured

138

4 Scales and Measurement

• A pre-test is conducted to measure the intensity of the favourable or unfavourable attitude of respondents towards each test item. The scale would have
intensity descriptors like, highly favourable, favourable, neutral, unfavourable
and highly unfavourable. These responses are given a numerical weight.
The total attitude score is represented by the algebraic sum of the weights of the
items. To make the measuring process uniform, the weights are consistently
assigned. For instance, if 5 were assigned to reflect strong agreement with a
favourable situation, then 5 should be assigned to show strong disagreement with
an unfavourable situation too.
After the results have been obtained, the researcher selects items that reveal a
clear discrimination between high and low total scorers by identifying the highest
and lowest quartiles based on total scores. Subsequently, mean differences are
computed for these high and low groups.
Finally, a set of items is chosen that represent the greatest difference between
the highest and the lowest mean values.
Likert scales are very popular among researchers for measuring the attitudes of
people. But, in practical situations, commercial researchers are more concerned
with finding the respondents attitudes towards individual components, rather than
overall positive or negative attitudes of respondents. For instance, the manufacturer of a brand of shoes will be more interested in finding out why people are not
buying the brand rather than respondents’ attitudes towards shoes in general.
A typical Likert scale is discussed in the Exhibit 4.18.

Exhibit 4.18: Likert Scale
A Likert scale for evaluating the attitudes of customers, who have not
used a vacuum cleaner, but are aware of its existence, is given below.
Here are some statements that describe how customers might feel about
vacuum cleaners.
Please indicate your agreement or disagreement. For each statement given
below, please circle the appropriate number to indicate whether you:
1-strongly agree, 2-agree, 3-neutral, 4-disagree and 5-strongly disagree.
Strongly Agree Neutral Disagree Strongly
agree
disagree
The product is costlier
I do not find time to use a vacuum
cleaner
Advertising of the product is not
convincing enough

1
1

2
2

3
3

4
4

5
5

1

2

3

4

5
(continued)

4.13

Types of Attitude Scales

139

(continued)
Strongly Agree Neutral Disagree Strongly
agree
disagree
I have never used a vacuum cleaner
I am satisfied with the way I am cleaning
my house right now
Using a vacuum cleaner is cumbersome
Competitor’s vacuum cleaner has better
features
The initial enthusiasm to use a vacuum
cleaner dies down at a later stage and it is
permanently kept in the store room
The demonstration of the product given
by the salesperson is not effective

1
1

2
2

3
3

4
4

5
5

1
1

2
2

3
3

4
4

5
5

1

2

3

4

5

1

2

3

4

5

4.13.14 Thurstone Scales
In Thurstone scales, researchers select a group of 80 to 100 items indicating the
different degrees of favourable attitude towards a concept under study. Once items
are selected, they are given to a group of judges, who are asked to categorize them
according to how much they favour or disfavour them. The judges are asked to
treat intervals between categories as equal and analyse each item without
expressing their own attitudes. Once the results are obtained, all those items that
have a consensus from the judges are selected and items where there was no
consensus are eliminated. These results are then distributed uniformly on a scale of
favourability. This scale is then administered to a set of respondents for measuring
their attitude towards a particular concept. Although the thurstone method is timeconsuming as it involves a two-stage procedure, it is easy to administer. This
method comes under criticism because the scale values are developed based on the
attitudes of the judges.

4.13.15 Profile Analysis
Profile analysis is a process where two or more objects are rated by respondents on
a scale. Profile analysis can be considered as an application of the semantic
differential scale. Comparing different objects visually, based on different attributes, is possible in this approach. The major disadvantage is that it is very difficult
to interpret the profiles as the number of objects increases. The profile analysis is
used in Exhibit 4.19 to compare the three jeans brands.