Measurement Voss and Parasuraman (2003)

¶ … Measurement Voss and Parasuraman (2003) and Zikmund (2003) define measurement as the process by which a property of a phenomenon of interest of the researcher is described by assigning numbers to it in a valid and reliable manner in order to convey information regarding the property that is being measured. Performance in the research class which the property being measured by the instructor assigning them grades from a to F, numbers say from 1 to 10, percentages from 0 to 100 or two letters such as P. And F. depending on their performance.

A student who is given a grade of a is one who has performed excellently while F. represents students who have failed. In terms of numbers, 10 would represent a student with outstanding performance while a student below 5 has failing performance. For percentages, 100% represents a perfect score and below 59% represents failing performance. In terms of two letter, a P. would mean a student has a passing mark while an F.

represents a student with a failing mark Zikmund, 2003() Difference to concept of the construct Constructs refer to the measurement of concepts using multiple variables. In measurement, only one variable which is the academic performance of the student is used while in a construct, multiple variables are used such as academic performance, extra-curricular performance and class attendance Zikmund, 2003() Levels of scale measurement There are four types or levels of scale measurement which are nominal, interval, ordinal and ratio level scales.

The nominal scale assigns a particular value to an object for the purposes of identification or classification. This scale is qualitative therefore the value of the variable does not have to be a number. They are used where a qualitative scale is the only true measure even though they may be elementary. Any variable can be assigned any label without introducing error to the measurement.

Examples of usage include sports uniform numbers and school bus numbers Zikmund, 2003() The second type of scale is the ordinal scale which allows variables to be organized according to how much of a particular concept they possess. This means that it is a ranking scale. A good example of this scale is ranking of students in school or in financial management where stocks are ranked as 'buy', 'hold' or 'sell'.

Though the person would know how the stocks are ranked, they do not know by how much they outrank each other Zikmund, 2003() The interval scale is the third level of scale and has both ordinal and nominal properties. They also capture differences in quantities of a concept. As in the examples given earlier for the ordinal scale, not only would the person how the stocks are ranked but also by how much they outrank each other.

So the stocks would be presented say on a scale of 1.0 to 10.0 where 10.0 is a sell position, 1.0 is a buy position and 5.0 is a hold position. So a stock with a rank of 10.0 outperforms another that is ranked 8.0 by 2.0 points Zikmund, 2003() The last type of scale is the ratio scale which represents the highest form of measurement since they have both ordinal and nominal values in addition to representing absolute quantities.

The difference between interval and ratio scales is that the former represents relative meaning while the second represents absolute meaning. In the example of stocks presented above, we can say that a stock that has a position of 8.0 is twice as valuable as the one with a position of 4.0 Zikmund, 2003() Grading measuring scales When students are graded using numbers and percentages is an interval scale since it gives the performance of students as a numeric value on a scale from 1 to 20.

This scale provides low chances for error since by interpreting the performance on the scale it is possible to get a statistical understanding of the performance of each student. The same goes for grading using a letter corresponding to their performance from a for excellent performance to F. For failing performance. These methods of grading are also ordinal scales since they represent an ordered arrangement of performance of the students. Grading using two letters corresponding to performance, P for pass and F.

For fail is a nominal scale and it gives a high chance for error since the letters do not have any arithmetic properties and are simply labels for the students. Additionally, it does not say anything about the relationships that exist between the students' performance. This is also an ordinal scale since it represents an ordered arrangement of performance Zikmund, 2003() Question 6 i. The summated scale score for the respondent is 7+5+8+6 which is equals to 26. ii. Based on the average composite scale score, the score is 6.5. iii.

None of these items needs to be reverse-coded since none of the variables state the opposite of the variable intended. All items are on the same scale and none of them is on a negative scale, therefore, they do not need to be reversed. Attitude Attitude is defined as a disposition in the respondent to respond consistently to certain aspects of the world such as people, actions and objects.

There is a general consensus regarding the definition of attitude as the cognitive component that represents the knowledge of an individual regarding a particular attribute and its consequences. Differences between rating and ranking Ranking is defined as the ordering of the variables based on the overall preference or a particular characteristic. Rating, however, is estimation of the magnitude or extent to which some particular characteristics exist. Both ranking and rating have been used in many studies to measure attitude.

It has been found that both reduce language and response bias significantly but ranking is the better solution since it eliminates both extreme and middle response styles and acquiescence and dis-acquiescence completely Harzing et al., 2009. Ranking is also better since it is easily understood by respondents. Reverse cook recording in SAS or SPSS SPSS has a recode function which can be used to perform simple recording by entering the old and new scale values.

Recoding can also be done using a simple mathematical formula as below: Xnew value = (n+1) -- Xold value Therefore, for a scale of 1-5, the applicable formula will be Xnew value = 6 -- Xold value Advantage of numerical scales over semantic differential scales Semantic differential scales is where a series of seven-point rating scales which use bipolar adjectives are used to anchor the beginning and end of each of the attitude scales used.

Numerical scales, on the other hand, simply use numbers instead of verbal descriptions or semantic space in order to identify options or categories of the responses. It also uses bipolar adjectives in the same way that semantic differential scales do. Semantic differential scales are quantitative measures of attitudes unlike numerical scales which are qualitative measures of attitudes. The numerical scale is more advantageous since it is easier for respondents to understand and fill and is as effective as a true semantic differential.

Semantic differential scales are also more difficult to prove their validity as a result of the difficulty in finding semantic opposites Zikmund, 2003() Issues to consider when choosing a measurement scale Choice of the measurement scale to use depends on the definition of the manager's problem, and the linkages that exist between the choices that have already been made. However, there are certain techniques that help the researcher make the decision. The first is whether ranking, sorting, rating or choice is the best technique.

This decision is largely determined by the problem definition and the type of statistical analysis that is desired. The second is whether a monadic or comparative scale should be used. Here, the researcher needs to decide whether to include a standard of comparison in the verbal portion of the scale. Third is the type of category labels for the rating scale. The researcher may choose to use verbal, numerical and unlisted choices labels.

Fourth is the number of scale categories or response positions that are needed in order to accurately measure attitude. Fifth is whether a balanced or unbalanced rating scale should be used. In a balanced scale, there is an equal number of positive and negative response categories therefore there is a neutral point or point of indifference at the center of the scale. In an unbalanced scale, there are more response categories at one end leading to an unequal number of positive and negative categories.

Sixth is whether a scale that forces one to choose from the predetermined options should be used. A forced-choice rating scale requires the respondent to answer through allowing the midpoint of the scale to be used to indicate that the respondent is unaware of the indifference that exists. Lastly is whether a single or index measure should be used. Single measure can be used for simple concepts that do not vary from context to.