Quiz 11: Measurement
Measurement is process of assigning numbers to the empirical objects or events based on the set of rules. These objects or events are measured using variables, which are further classified as properties or as objects. Researcher measures the indicants of the properties or objects. For instance: if a person is considered as an object, then properties of that object like height, attitude, age etc can be measured. a. The properties used for the measurement of the object laundry detergent are color, odor, quality of the powder, availability of quantity etc. These properties are classified using the different type of scales. For instance: The color of the powder is given name or label like blue, black etc. Further, respondents can be asked how strongly they like the color of the laundry detergent using 3point -likert scale - Not very much, Ok, Too much. b. The properties used for the measurement of the object employees are gender, ethnicity, work experience, age, salary package, performance rating etc. For instance: Gender can be measured using nominal scale (i.e. categorical variables) such as male and female. c. The object factory output can be measured in terms of production numbers, defect rate, capacity utilization, profit earned etc. For instance: All the data for these variables are numerical. d. The properties used for the measurement of the object job satisfaction are work flexibility, company atmosphere, company benefits, challenging work, compensation etc. For instance: compensation can be measured using likert - scale. If the employees strongly like the compensation, then it affects his / her like towards job satisfaction.
Measurement is process of assigning numbers to the empirical objects or events based on the set of rules. These objects or events are measured using variables. The characteristics of the data or variables are classified as order, distance and origin, which combination provides the measurement scale. Measurement scale is used to categorize the values assigned to the variables. The four scales of measurements are as follows: • The nominal scale information can be grouped into two or more categories, which is mutually exclusive and collectively exhaustive. • The ordinal scale includes the characteristics of nominal scale along with indication of the order. It provides the ranking of the data. • The interval scale includes the characteristics of nominal and ordinal along with concept of equality of interval (i.e. distance). • The ratio scale includes the characteristics of all nominal, ordinal and interval scale along with absolute zero or origin. This scale represents the actual amounts of variable. The differences in these scales are that all scales do not combination of all three characteristics such as order, distance and origin, except ratio scale. The impact of these differences of the scales in the statistical analysis techniques is as follows: • As nominal scale doesn't have all the three characteristics, it provides only frequency or count of the data, which is mode of the central tendency. • As ordinal scale has the characteristics of "order", it provides the mean of the central tendency and percentile, quartile as measure of dispersion. • Interval scale has both order and distance characteristics, which can able to provide mean of the central tendency, standard deviation as a measure of dispersion. Also, this scale is usable to perform statistical procedures such as co-relation analysis, t -test, f-test and other parametric tests. • The ratio scale includes all the characteristics such as order, distance and origin, which is usable to perform all statistical techniques.
In a research, it is hoped that any measurement represents a true portrayal of the property being measured, but we recognize that it often does not. Errors can be of either a random or a systematic type, but it is the latter, which represent the greater threat to the measurement process. Most systematic error in a survey, for example, flows either from the following sources: • Respondent: Respondent refers to individuals on whom a interview or survey is conducted. Some of the factor such as hunger, mood, boredom, ignorance, and no subject knowledge and so on limits the ability of a respondent to respond truly. For example, an interview about national income is conducted where a question about person's annual income is included. The respondents themselves may be a source of error, as they may not actually think of every source of income that is included in the research definition. Respondents may also bias results if they tend to exaggerate or understate their income. • Situational factors: The situation under which the interview takes place may distort the measurement and responses, especially if the study is conducted in the presence of someone else, even a family member. Other situational biases arise if the interview conditions do not foster carefully reasoned answers or provide access to records. For example, an interview about sex education is conducted on youth in presence of their parents. Respondents might not feel comfortable while expressing their views in front of their parents. This could distort findings. • The researcher/interviewer: The interviewer can affect results, if he/she fails to achieve rapport with the respondent and therefore appears as a threat. Careless recording, coding, and tabulating are some of the sources of researcher error. For example, if a manager himself conducts a survey to access employees attitude and thought about him/her. Employees may not feel free to express their view. Therefore, his presence will distort the findings. • The instrument used: Instrument itself can be a source of error. Use of complex word, leading questions, ambiguous meaning, poor printing, and lengthy question can even distort the true meaning of question. Such ambiguous question may confuse the respondent. For example, in one major national survey more than a dozen questions were used to establish a family's total annual income but don't cover all of the possible income sources. This will leave respondent clueless that will result in wrong response.