What is Scale ?
A series of categories or items arranged in a progressive manner (in a continuous spectrum) based on magnitude or value is called 'scale'. Different responses of individuals are quantitatively placed on such scales. Every scale includes a lowest point, a highest point and some intermediate points (representing a particular activity or attitude). If the first point or item on the scale is of highest degree, the second point is higher in nature, than the third point; consequently, the third pint is higher in nature than the fourth point and so on.
The process of developing a continuum for placing the measured objects is called scaling. These measured objects may involve interests, opinions. attitude and other responses of the individuals. Thus, the next operation involved in measurement is the scaling process. In this, different qualitative aspects are associated with quantitative components.
Scale of Measurement in Research Methodology :
Measurement is followed by scaling procedure in a research activity. Researchers always face the problem of how to measure or quantify abstract concepts and how to relate one individual's response to that with another response. Hence, the problem is two-fold. First is, how to quantify a response in absolute terms, and the second is, how to relate it with other responses. This problem is resolved by scaling process, or assigning the numbers or values to responses and positioning them on a common scale. Scaling can then be defined as, "creating a continuum with two extreme limits and few immediate values between them". Hence, scaling is the process or procedure of quantifying and measuring abstract concepts like attitudes, opinions, emotions, etc. Scaling also quantitatively determines the position of an individual in a range of values.
Definition of Scaling
According to Edwards, scaling can be defined as,
"Procedure for the assignment of numbers (or other symbols) to a property of objects in order to impart some of the characteristics of numbers to the properties in question"
In a scale, the values are progressively placed according their value or magnitude in such a manner that moving forward in a scale will depict an item to be always higher than the immediate lower one.
Process of Scaling
Among all the commonly used direct techniques to measure attitudes, like word-association tests, sentence-completion tests, story-telling. performance of 'objective' tasks. physiological reactions, etc., the 'self-report method' is the most popular. In this method, the respondent is asked to give his opinion about a particular object freely. Described below is a step-by-step guide on how to construct a self-report :
1) Gathering Related Variables :
Once the subject of inquiry has been decided, it is time to collect all possible variables and related statements, which are then examined to remove those that are imprecise and ambiguous, so that the questionnaire is easily comprehensible and accurately answerable. Utmost care should be taken to ensure the use of simple, easily understandable language,
2) Shortlisting the Variables :
Once the pool of variables or questions has been finalized, it is time for scrutinizing the variables to ensure that the question or variable can be included in the scale. Some of the criteria for selecting the variables are as follows :
- Relevant with the enquiry,
- Indispensible,
- Easily comprehensible,
- Accurately answerable,
- Covers all the possible dimensions and angles of the subject of enquiry.
Before finalizing, its efficacy is needed to be tested either through a sample tested on a dummy or by assigning the task to a group of experts.
3) Formation of Scale :
Once the variables have been finalized, the scale is then tested for its validity and reliability, before being put to use in a survey.
Scale Construction Techniques
Techniques used in developing set of items or categories in a progressive manner (in a continuous spectrum) so as to measure an individual's attitude towards given event or object, refers to scale construction techniques. Different scale construction techniques in research methodology used in scaling are as follows :
1) Arbitrary Scales :
In this technique, large number of items or categories, which are able to represent the concept under study are collected and (then) measured in terms of same characteristic. Subjective selection of the researcher plays an important role here. Firstly, few related statements or items are selected by the researcher, and then these statements are filtered for being included in the measuring instrument. A list of filtered statements is presented to the respondents and they are required to tick the most suitable one.
2) Cumulative Scales :
A series of statements is used in developing scale for the respondents in cumulative scales. These are also called 'Louis Guttman's scales. These scales require respondents to show their agreement or disagreement over the statements. These scales are called cumulative scales because a cumulative series is formed by different statements here. For example, the confirmation of an extreme position item should also lead to the confirmation of less extreme position items.
3) Consensus Scaling :
This method of scaling was given by L.L. Thurstone. In this, a panel of judges uses different attributes like attitude. relevance and ambiguity factor so as to select items or categories. Generally, differential scales are created by this method, which focus on measuring attitudes of respondents concerning specific issues like religion, war, etc. These are the most suitable scales for recording only a single attitude of the given concept. A high level of cost and effort is needed to create such scales. Judges' own attitudes may affect their functioning of assigning values to different statements.
4) Item Analysis :
In this type of scale construction technique, a group of respondents is given a test created by selecting different individual items. Following steps are involved in this technique :
- Firstly, large number of items expressing their positive or negative aspects is selected.
- Then, a group of respondents assigns scores to them.
- Then for each respondent, total scores for all items are calculated. Then using these scores, they are divided into four constructs. Two middle constructs are then eliminated from the calculation.
- Then average score is calculated separately for each item.
- T-test is then used to compare average of items of different groups to select Items with noteworthy t-values.
In this, different statements expressing the positive or negative attitude towards the selected object are used to create the scale and the respondents are asked to check the most suitable one. Thus, the agreement or disagreement of respondents over different statements is recorded. The most common example of item analysis is Likert scale. In this, several degrees of agreement or disagreement (generally five, but sometimes three or seven) are used by the respondents to respond to 'different statements. This type of scale is very easy to develop.
5) Factor Scales :
Several different techniques are involved in constructing such scales. The purpose of using different scales is to study multi-dimensional items, to find the relations between different dimensions and to identify any other dimensions of the items. In the end, limited set of factors are available each having interrelated dimensions. Scales developed through this technique are as follows :
i) Q-Sort Technique :
In this, the respondent is asked to sort the different given statements into prefixed categories.
ii) Semantic Differential (SD) :
Factor analysis of assumed interval scales results in semantic differential scales.
iii) Stapel Scale :
It is a non-verbal rating scale having even number of scales. Single adjectives are used in items or categories. It is necessary to measure both the dimensions consequently. There is no need of assuming the ratings or the interval equality.
iv) Mull-Dimensional Scaling :
In multi dimensional scaling, a multi-dimensional space is assumed to be present. A set of techniques are developed to deal with such multi-dimensional space.
v) Standardized Instruments :
In this, an available measuring instrument is selected for data collection. No new measuring instruments are developed. The opinion of expert is taken for selecting such standardized instruments.
Criteria for Good Scale
Any scale must satisfy the following criteria to prove its efficacy :
1) Validity :
Validity is the most critical aspect. It establishes or indicates the extent to which the scale does what it is supposed to do. Does it measure what it was intended to measure? In other words, whether or not the test has been useful or worthwhile.
2) Reliability :
This indicates whether or not the findings of a measuring instrument can be relied upon. To know this, the researcher needs to answer some questions, such as, Are the outcomes accurate? Can they be replicated? Are the findings consistent?, etc. While reliability contributes to validity, it is not true the other way round. A reliable instrument may or may not be necessarily valid.
3) Practicality :
Practicality refers to ease and economy, of constructing the measuring instrument as well as administration and interpretation of the test outcome. It measures the achievability, and practicality of an instrument. The benefits or results should justify the costs.
4) Sensitivity :
This refers to how well the test is standardized. It tests how much an instrument is able to measure accurately. For example, a test which requires respondents to merely say 'yes' or 'no', may not be very sensitive. On the other hand, if the respondent is asked to rule on a '5' or "7"-point scale, the scale may be considered to be highly sensitive. It may however be noted that all scales need not be highly sensitive. It largely depends upon the requirement of the test.
5) Generalisability :
This refers to whether or not one can generalize the findings, i.e., whether or not the findings can be applied to both similar and different situations. Whether or not the sample or respondents selected can be said to represent larger population. For example, exit polls can be said to have a high degree of generalisability.
6) Economy :
This factor emphasizes on the aspect that whether the instrument, is economical to be constructed and conducted. Tests are expensive to develop and administer. Therefore, the results or benefits of the tests must justify or warrant the costs.
7) Convenience :
A measuring instrument is convenient, if it is easy to conduct. The instruments that provide guidelines to use are much more convenient than those instruments that do not include this feature. Generally, it is considered that the requirement for convenience increases with the level of complexity of a measuring instrument.
Importance of Scaling
In business or management research, scaling is very crucial for the research process. It helps in measuring. and analyzing attitudes of different individuals. The exact behavior of an individual is reflected by such attitude analysis. Number of attitude measuring scales has been developed by researchers. For example, in order to measure the attitude of an individual about a particular tourist place, product or election candidate, i.e., if he visits, buys or votes, respectively, a suitable scale is developed. The different facts describing the importance of scaling are as follows :
1) Attitude Scoring :
Scaling is particularly used for altitude scoring of an individual. In scaling, with the help of an individual's responses a particular number or point is selected on given scale, which represents the attitude of that individual.
2) Broad Application :
Different management research processes as well as scientific inventions use scaling as their crucial element. Data collection methods like interviews, observations, surveys, etc,, also use scaling for attitude measurement of respondents.
3) Hypothesis Testing :
Scaling is also functional in hypothesis testing. Without effective measure of attitudes of different respondents, it is not easy to test hypothesis about the population.
4) Dimension Checking :
Scaling is useful in determining the dimensional aspects of different quantitative concepts or items. It helps in checking whether a particular item is single dimensional or multi-dimensional.
5) Others :
- It is an essential element of the exploratory research.
- It is used to check whether a set of questions is measuring single aspect or multiple aspects.