This document outlines the steps for constructing a Likert scale, which is commonly used to measure attitudes, beliefs, and perceptions in agricultural extension research. It discusses defining the construct and dimensions, collecting and editing items, deciding on response categories, pilot testing with experts, conducting item analysis including computing t-values and reliability/validity testing, and standardizing the scale. The key steps are: 1) defining the construct and dimensions, 2) designing the scale including collecting/editing items and response categories, 3) pilot testing with experts, 4) administering to a sample and conducting item analysis, and 5) establishing reliability and validity to standardize the scale.
2. The method of Summated Ratings (The Likert Scale)
The most popular scale used in agricultural extension research
It is widely used for measurement of attitudes, beliefs, emotions,
feelings, perceptions and personality.
The Likert method of scale construction also called as the method of
summated ratings because each response to a statement (may be
considered a rating) are summated over all statements (Bird; 1940,
p.159)
Ordinal level of measurement
3. Informal Criteria of good items (Edward, 1957)
Avoid items that refer to the past rather than to the present
Avoid items that are factual
Avoid statements that may be interpreted in more than one way
Avoid statement which are irrelevant to the psychological object
Avoid statements that are likely to be endorsed by almost everyone or
by none
Statement should be short, rarely exceeding 20 words
Contd..
4. Informal Criteria of good items (Edward, 1957)
Statements containing universals i.e. All, always, none, never etc.
often introduce ambiguity and should be avoided
Words such as only, just, merely should be used with care and
moderation
Follow simple form of sentences rather complex or compound
Avoid the use of double negatives and Jargon
Items should be clear, well written, and contain a single idea
Keep the items short and the language simple and straightforward
for wide acceptance.
6. STEPS
1. Define Construct: the construct of interest must be clearly and
precisely defined
e.g. Identification of construct : Scale construction to identify the
training need of dairy processors to enhance their competency for
value chain development
Operationalization of dimensions under the construct: Major
dimensions identified under this construct were
Marketing (product, price, place and promotion),
Technical (input, infrastructure, finance, legal and managerial),
Information and
Social Responsibility.
7. 2. Design Scale: the scale itself is designed which involves deciding on the
exact format of the scale, including selection of response categories and
writing of instructions.
e.g. Collection of items: Items collected based on review of literature,
consultation with the experts from various institutes (say 135 statements
collected consisting 100 positive and 35 negative statements).
Editing of items: The statements collected were cautiously edited by
following the 14 informal criteria suggested by Edwards (1957). Thus, a
total of 110 statements were taken out of 135 statements.
Deciding response categories: Selecting five point continuum viz.,
Strongly agree, Agree, Undecided, disagree and strongly disagree with
the score of 5,4,3, 2 and 1, respectively and reverse for the negative
statements
8. 3. Pilot-test: The edited statements send to the experts who are asked to
critique the scale through Google forms or personally. They should
indicate which items are ambiguous or confusing, and which items
cannot be rated along the dimension chosen.
e.g. Relevancy test of items: Finally the 110 statements on a five point
continuum were sent to 55 Judges. Whereas, only 34 responses were
received.
Relevancy Weightage (RW) and Mean Relevancy Score (MRS) were
calculated
Finally the statements having relevancy weightage of ≥0.82 and mean
relevancy score ≥4.14 were selected for item analysis. Thus, a total of 60
statements selected for the item analysis.
9. 4. Administration and item analysis:
Item analysis is done to find those items that form an internally consistent
scale and to remove those items that do not (Spector, 1992).
e.g. The selected 60 items after expert relevancy test were administered to
a random sample of 40 dairy processors in non-sample area. The items
were rated on five point continuum
Computation of ‘t’ values: Based on total individual score, the judges were
arranged in descending order. Upper and lower 25 % of the judges i.e. 10
respondents (dairy processors) with the highest total score and 10
respondent with the lowest total score were selected.
These two groups used as criterion groups to evaluate the discrete
statements. Calculate t value for each items and select those items having
≥ 1.75 t value
The t value is a measure of the extent to which a given statement
discriminates between the high and low groups of respondents for each
statement
10.
11.
12.
13.
14. Standardisation of the scale
Reliability of scale : Reliability is the ability to give consistent, stable and
accurate measurement score in repetitive testing with same instrument.
The Split half method: The scale is reliable the value is more than .70
The Pearson product moment correlation (r) between odd and even
scores was 0.647
Spearman- Brown (1910) prophecy formula
15. The value of Cronbach’s alpha calculated and found to be 0.855
Cronbach’s alpha is the average of all possible split-half correlations
(Cortina, 1993).
It assumes the average covariance among non-parallel items is equal
to the average covariance among all parallel items
16. Validity of scale : Validity means ability of an instrument to measure what
one intended to measure
Content validity of scale is the representative or sampling adequacy of the
content, the substance, the matter and the topics of a measuring
instrument.
Lynns’ method (1986) used for quantifying the content validity, the 40
selected statements (‘t’ value >1.75) were given to 6 experts (processors).
The 6 experts were selected as with increase in number of experts the
likelihood of achieving total agreement decreases.
The 4-point scale was used to avoid a neutral and ambivalent midpoint as
per Davis (1992)
1= not relevant, 2 = somewhat relevant,
3= quite relevant and 4= highly relevant.
S-CVI/Ave =0.930 (excellent)
17. Advantages of Likert-like scale
It is relatively simple
Less laborious
Convenient
Large number of statements included
Both negative and positive statements minimize the possible
response sets of subject (only favourable or unfavourable)
18. Limitations of Likert-like scale
Uni-dimensional-: it gives only certain amount of choices
Previous questions can influenced the response of any further
questions that will be asked
Depends on participants honesty and knowledge of particular topic
19. • Identifying Construct and its
dimensions
Define Construct
• Collection and editing of items
• Deciding response categories
Design Scale
• Relevancy testing
Pilot Testing
• Administering scale to non-sample
experts
• Computing t- value
Item Analysis
• Reliability
• Validity
Standardization
20. References
Shruti; J.P. Sharma; R.R. Burman; R. Gills and M. Singh 2019. Scale Construction
to Identify Training Needs of Agripreneurs to Enhance their Competency for
Value Chain Development: A Methodological Approach. Journal of Community
Mobilization and Sustainable Development, 14(1):33-40
Edward, A.L. 1957. Techniques of Attitude Scale Construction. Vakils, Feffer and
Simons Inc, New York.
Kerlinger, F.N. 1973. Foundations of Behavioral Research. (2nd ed.) Holt,
Rinehart and Winston.
Likert, R.A. 1932. A technique for the measurement of attitude. Archives of
Psychology, 22(140): 1-55.
Lynn, M.R. 1986. Determination and quantification of content validity. Nursing
Research, 35: 382–385.
Spector, P.E. 1992. Summated Rating Scale Construction: An Introduction Sage
University, Papers Series. Quantitative Applications in the Social Sciences; No. 7-
82.
Singh, M.; B.K. Singh; P. Singh; R.R. Burman and Shruti. 2018. Development of
Scale to Measure Attitude of farmers towards IARI-Voluntary Organizations
Partnership Extension Model: A Methodological Approach. Journal of
Community Mobilization and Sustainable Development, 13(2):221-226.