 # Measurement Scales Paper

Examples of measurement scales and their use in a questionnaire

In marketing research, the respondent’s answers and comments are needed in a measurable form. To enable this, researchers have developed a scale of measurement to allow for the definition and grouping of these variables. Each of the scales has unique measurement properties that determine its range of use statistically. These significant properties include magnitude, which refers to the capability of determining the difference in quantity between data. The second property is the equal interval which is a trait concerned with the existence of uniform intervals between categories. While the third property, absolute zero, refers to the end of the scale or to the point where zero has been placed (McDaniel & Gates, 1998). The four scales of measurement discussed here are nominal, ordinal, interval and ratio. These four are usually arranged in a hierarchy with nominal being the most primitive and ratio occupying the other end.

Nominal scale as its name suggests, is based on names. In this scale, data such as individuals and products are placed in categories while the researcher pays no consideration to order. It does not posses any of the above named qualities that are of importance to a researcher. The values assigned to data are arbitrary and represent difference in quality but do not express magnitude. In statistics, it is the lowest in the hierarchy of scales. Nominal scale supports only two types of statistics – mode and chi square (Anderson, Sweeny & Williams, 2008). An example of its use in a questionnaire may be in areas where the data involved is classification data or in instances where the response is yes/no. For instance, the data of the gender of respondents may be coded using numerals during data entry. Male may be represented by four while female may be represented by seven. As discussed earlier, order is not of importance, therefore it would make little sense to say female is greater then male.

Ordinal scale occupies the next spot in the hierarchy of measurement scale. In contrast to the nominal scale, the ordinal scale ranks the data based on continuity of data within a given category. For example in a questionnaire, when asked to make a list of preferred movie theatres, one is making an ordinal scale. The Likert scale is a well-known example of ordinal scale in a questionnaire. The range of responses in a Likert scale was originally from a choice of five starting from strongly agrees and ending in strongly disagrees (McDaniel & Gates, 1998). Preference scores are also another example of ordinal scales. The shortcoming in ordinal scales is in its inability to show the interval between ranks. For example, a list of one’s best clothing brands only informs of the order of preference but does not inform of the difference in magnitude between the first and the second placed or the last and the second placed. Therefore, ordinal scale enables one to appreciate the general direction of rank but not the magnitude in difference between ranks (Anderson et al., 2008). In statistical analysis of ordinal scale, two measures are applicable, the measures of mode and median.

The interval scale is the third in the rank of measurement scales. In this scale, the magnitude of data is established and the intervals between ranks are shown. Unlike in the ordinal scale, the interval is measurable since the units in use are equal. It does not posses an absolute zero point. The zero value is established arbitrarily. The lack of zero point makes it difficult to express interval scales in ratios. An example of such as scale is the Fahrenheit temperature scale. The difference between sixty and eighty degrees is equal to the difference between twenty-three and forty-three degrees .In contrast, in the absence of absolute zero, it makes it incorrect to claim that fifty degrees is twice as hot as twenty five degrees. In a questionnaire, the use of a normal survey rating scale is an example. In such a survey, the respondent may be asked to rate their agreement to a particular statement on a scale of five ranging from strongly agree to strongly disagree in a similar way to a Likert scale which in some instances is considered an interval scale. The interval scale provides more opportunities quantitatively. This is because a wider range of statistics can be derived. The mean is the most notable addition .The other statistical measures are standard deviation, variance and analysis of correlation and regression.

The ratio is the highest scale in the measurement hierarchy. It has all measures that are of importance in research including the absolute zero. The presence of absolute zero enables better analysis of data. This is because it enables the researcher to compare the difference in magnitude between ranks and different data. This can be carried out by subtraction, addition and other mathematical calculations. In addition, a ratio scales bears characteristics of all the other scales and can be converted into both ordinal and nominal scales. Examples of data that are expressed in a nominal scale in a questionnaire include age, height and weight. Another common example is in money, it is in order to say that two hundred dollars is twice as much as a hundred dollars while at the same time it is still possible not to have money thus have zero dollars. The statistical information that can be derived from ratio scale is similar to that of the interval study (Anderson et al., 2008).

In conclusion, the choice of any measurement of scale is dependent on sound judgment. Several factors should be put in consideration amongst them include the statistical possibilities of a particular scale, the assumptions of the chosen statistical analysis and the statistical analytic model to be used. As a rule, the model that offers the most information is chosen (Malhotra, 2009). This applies even in application of these models in questionnaires.

Works Cited

Anderson, D. R., Sweeny, D. J., & Williams, T. A. (2008). Statistics for business and economics. Stamford, Connecticut: Cengage Learning

Malhotra, N. K. (2009). Marketing Research an Applied Orientation. London: Pearson

McDaniel, C. D., & Gates, R. H. (1998). Marketing research essentials: Marketing Research Series. New York: Taylor & Francis