Statistics Allowable With Nominal, Ordinal

¶ … statistics allowable with nominal, ordinal and interval scales.

Nominal is a counting operation and its descriptive statistics is "frequency in each category, percentage in each category mode." Ordinal is a rank ordering and its descriptive statistics is "median range, percentile ranking." Interval is an arithmetic operation on intervals between numbers and its descriptive statistics is "mean, standard deviation and variance." Understanding descriptive statistics necessitates specifically looking at the type of data that are being described. The nominal scales only place numeric labels on non-quantitative concepts, for example, dogs have the value of "1" and cats have the value "2." Many categories or groups are actually nominal, such as racial group and gender. In some research, for instance, the study counts the number of individuals who are in a specific category, such as living in a designated city. Ordinal scales are ranked in a way that compares one to another, with a highest and lowest. An example is the tallest and shortest children in the school. It is not possible to perform ordinal data with mathematical computations. Interval scales allocate specific values to something, so that the intervals are equal, such as a six-point attitudinal scale. In this case, mathematical operations can be performed. With ratio scales, there is the interval with a true zero point, such as weight or the number of something in a room. Then ratios can be determined.

Difference between validity and reliability. The purpose of conducting a study is to come up with accurate measurement results. This is why research must be both reliable and valid; the two are interrelated. Reliability is the consistency of the measurements, or how well the study can be repeated. Does the same measurement yield the same results when repeated? Reliability cannot be calculated, only estimated. Validity is whether the test is measuring what it expects to measure. if, for example, the researchers are measuring a table that is six feet wide, they measure the table with a measuring tape and find it is six feet. They measure it again and again and consistently get six feet. The tape measure is yielding reliable results. The tape measure includes inches and feet, so it should also yield valid results. If the researchers measure the table with the "right" tape measure, it should yield a correct measurement of the table's width. In other words, when conducting research, it is necessary to use measurement tools that yield consistent responses when asked time after time and that yield accurate responses from the participants.

Difference between conceptual and operational definition. Conceptual definitions define a concept with the use of other concepts, which makes measuring difficult. An operational definition specifically identifies at least one observable condition or event, so the researcher knows how to measure that condition or event. The operational definition must be reliable and valid. For instance, if a researcher wanted to know about a person's enjoyment for his or her job, the conceptual definition would reflect interest for an enjoyment and satisfaction in his or her job and the ability that he or she has to apply skills. The operational definition would ask this person how true each statement is about his or her job: very true, somewhat true or not very true at all. (e.g. 1) I find my work enjoyable; 2) I can use my own unique skills and abilities; 3) I gain satisfaction knowing that I have the opportunity of demonstrating my best abilities to others.

Why would a researcher want to have more than one question to measure satisfaction with a job? By asking the same question in different ways, it helps the researcher insure that the answers are correctly defined. This makes the response more concrete and increases the ability to be reliable and valid. It must always be remembered that an operational definition is a matter of degree; there is an infinite factors that can be considered. The goal is not to have a completely concrete definition, but rather one that can be replicated by another researcher. The more information that is found the less subjectivity and possible bias. There is more information gained on the reason for enjoyment, for example, as noted above with "I enjoy my work," and "I find satisfaction in my work because I can demonstrate my best abilities to others." a) the researcher's assumption is not valid. The person could be ordering the magazine for a variety of reasons, such as wanting to buy a car or looking for the most cost-efficient green products. The purpose of validity is to know that you are asking the right question.

A b) the assumption may not be reliable, since it may not test the same over and over again. Reliability depends on consistency.

A c) This is reliable and valid since it is consistent and is measuring what is desired to be measured, the frequency of magazine reading.

6) prices on a stock market -- nominal

Marital status, married or never married -- nominal

Whether or not respondent has been unemployed -- this is a yes or no, nominal

Professional rank; assistant professor, associate professor, professor -- ordinal

Chapter 14

In this book, attitudes are defined as enduring dispositions to respond consistently, in a given manner, to various aspects of the world, including persons, events and objects. The three components of attitudes are effective, or emotions/feelings involved, cognitive, or awareness/knowledge, and behavioral or predisposition to action. These attitudes are hypothetical constructs, meaning that they are variables that are not directly observable but must be measured indirectly.

However, there are two difficulties with the word "attitude." First, the word attitude is not definitive, such as a number. It can be and has been defined in different ways. For example,