Variables and Measurements<br />Communication Research<br />Syed Muhammad Jamal<br />MS in Media and Communication Studies<br />International Islamic University<br />Islamabad<br />firstname.lastname@example.org<br />
Concepts and Constructs<br />Concepts (formally developed ideas that a researcher may seek to operationalize)<br />They are the abstract terms we employ to explain or make sense of our experience.<br />Definition:<br />A concept is a term that expresses an abstract idea formed by generalizing from particulars and summarizing related observations.<br /> Our understanding of the concept develops as we improve our ability to relate it to particular phenomena.<br />Parts of Concepts:<br />Symbol (word or term)<br /> Definition<br />Example<br />Height(symbol) is a simple concept from everyday experience. It is a characteristic of a physical object, the distance from top to bottom(definition). We can measure height or compare it. A height of zero is possible and height can increase or decrease over time. As with many words we use the word in several ways. Height is used in the expressions the height of battle, the height of the summer, and the height of fashion.<br />
Concepts and Constructs<br />Importance<br /><ul><li>Concepts simplify the research process by combining particular; characteristics, objects, or people into more general categories.
Concepts simplify communication among those who have a shared understanding of them.
People must share an understanding of a concept in order for the concept to be useful.</li></ul>Example<br />Authoritarianism<br /> It is a concept that was developed by social scientists to explain a phenomenon which came to be recognized after World War II: the state of mind that disposed individuals to accept the kind of authoritarian regime that appeared most dramatically in Nazi Germany.<br /> In this case, a number of beliefs and opinions, which appeared to be logically connected to one another and to the kind of behavior the social scientists were trying to explain, were drawn together to form a single concept—authoritarianism.<br />
Concepts and Constructs<br /><ul><li>Concept involves logical relation
In developing the concept of authoritarianism, social scientists were not just drawing their ideas out of thin air. They were reflecting on other related concepts and on the applicability of these concepts to particular phenomena in everyday life in order to develop a new concept which would more accurately capture the complex nature of the appeal of antidemocratic ideology.</li></li></ul><li>Concepts and Constructs<br />Construct<br /> A construct is a concept that has three distinct characteristics.<br /><ul><li>It is an abstract idea that is usually broken down into dimensions represented by lower-level concepts. In other words, a concept is a combination of concepts.
Because of its abstraction, a construct usually cannot be observed directly.
A construct is usually designed for some particular research purpose so that its exact meaning relates only to the context in which it is found.
Concepts and Construct are valuable tools in theoretical research.</li></ul>Example<br /> The term authoritarianism represents a construct defined to describe a certain type of personality; it comprises nine different concepts, including conventionalism, submission, superstition, and cynicism. Authoritarianism itself cannot be seen; some type of questionnaire or standardized test must determine its presence. <br /> The results of such tests indicate what authoritarianism might be and whether it is present under given conditions, but the tests do not provide exact definitions for the construct itself.<br />
Variables<br />Variable (A measure on which differences in response can be established)<br />Definition <br />The empirical counterpart of a construct or concept is called a variable. Variables are important because they link the empirical world with the theoretical; they are the phenomena and events that can be measured or manipulated in research.<br /> Variables take on two or more values.<br /> The values or the categories of a variable are its attributes.<br />Example<br />Gender is a variable; it can take on one of two values: male or female.<br />Male is not a variable; it describes a category of gender and is an attribute of the variable gender. <br />
Types of Variables<br />Types of Variables<br />Researchers usually begin with an effect, then search for its causes while focusing on the relationship between two things in which one causes other. Variables are classified depending on their location in a causal relationship.<br /><ul><li>Independent Variable</li></ul> The cause variable, or the one that identifies forces or conditions that act on something else, is the independent variable.<br /><ul><li>Dependent Variable</li></ul> The variable that is the effect or is the result or outcome of another variable is the dependent variable.<br /><ul><li>Intervening Variable</li></ul>A basic causal relationship requires only an independent and a dependent variable. A third type of variable, the intervening variable, appears in more complex causal relations. It comes between the independent and dependent variable and shows the link or mechanism between them.<br />Example<br />French Sociologist Emile Durkheim developed a theory of suicide that specified a causal relationship between marital status and suicide rates. Durkheim found evidence that married people are less likely to commit suicide than single people. He believed that married people have greater social integration (i.e. feelings of belonging to a group or family). He thought that a major cause of one type of suicide was that people lacked a sense of belonging to a group. Thus, his theory can be restated as a three-variable relationship: Marital Status( independent variable) causes the degree of social integration(intervening variable), which affects suicide(dependent variable)<br />
Identification of VariablesNumerical<br />Numerical Variables<br /> These variables are broken down into units in which the numbers used to present each unit of the variable carry mathematical meaning.<br />Discreet Variable<br /> A discreet variable includes only a finite set of values; it cannot be divided into subparts. <br />Continuous Variable<br /> A continuous variable can take on any value (including fraction) and can be meaningfully broken into smaller subsections.<br />Examples<br /><ul><li>Number of children in a family is a discreet variable because the unit is a person. It does not make much sense to talk about a family size of 2.24, Political affiliation, population, and gender are other discreet variables.
The average number of children in a family is a continuous variable; thus, in this context, it may be perfectly meaningful to refer to 0.24 of a person. Height is also a continuous variable.</li></li></ul><li>Identification of VariablesCategorical<br />Categorical Variables<br /> A categorical variable is made up of a set of attributes that form a category but do not represent a numerical measure or scale. These variables are made up of set of categories(or attributes) which must follow two rules.<br />Mutually Exclusive<br /> They must be distinct from one another; this is, they must be mutually exclusive. This means that no respondent should be able to place him/herself in more than one category.<br />Exhaustive<br />They should cover all of the potential range of variation in a variable. In other words, even respondents with a very extreme position on one variable should be able to place themselves comfortably within one of the categories.<br />Example<br /><ul><li>Occupation: teacher, plumber, salesperson, etc.
Religion: Catholic, Jew, Protestant, Muslim etc.
In the religion example you would need to add the categories of “other” and “none” to Catholic, Protestant, and Jewish in order to have an exhaustive list appropriate to American sample.</li></li></ul><li>Conceptual Definition<br /><ul><li>Conceptual definition defines constructs or variables by relating them to other constructs.
Researchers use two categories to generate conceptual definitions.
“Learning”(construct) acquisition of information; two concepts: ‘acquisition’ & ‘information’
“communication” verbal and nonverbal interaction between two or more persons</li></ul> To fully define construct communication, definition of verbal and non verbal are needed.<br /> ‘Verbal’ written or spoken words<br /> ‘Non-Verbal’ gesture or other nonlinguistic devices<br /> ‘Interaction’ the exchange of messages<br /> Conceptual definition: verbal and non verbal exchange of messages between two or more persons.<br />
Primitive <br />Primitive term is a concept with an agreed upon meaning.<br />Example<br />Communication between two or more persons between two or more persons is generally agreed upon primitive term and no formal definitions are needed to explain them.<br />
Derived<br /><ul><li>Derived terms are constructs and concepts which are required to be defined by primitive terms.
The researcher must provide conceptual definition to all concepts using in problem statement to avoid confusion and misunderstanding
Nonverbal gesture or other nonlinguistic devices
Written or spoken language is primitive terms having agreed upon meaning.</li></li></ul><li>Operational Definition<br />An operational definition assigns meaning to a construct or variable by specifying the activities to measure it.<br />Example<br />The construct ‘learning’ might be operationally defined measuring the ‘information acquired from educational TV programs’.<br />
Proposition<br /><ul><li>A proposition is simply a statement about one or more concepts or variables. It is a statement not for testing; .e.g. increase in population will increase illiteracy and crime rate.</li></ul> It is good proposition not hypotheses. This single proposition make two hypotheses,<br /> 1. increase in population, increase in crime rate.<br /> Every hypotheses is a proposition but every proposition is not hypotheses<br /><ul><li>Univariate: A proposition that discusses a single variable is called ‘Univariate’.
Bivariate: A proposition that discusses the relationship between two variables is called ‘bivariate’.
Multivariate: A proposition that discusses relationships between more than two variables is called multivariate. I</li></ul> Multivariate are usually written as two or more bivariate propositions.<br /><ul><li>Examples
There is a continuous increase in the urban population of Pakistan.
Increase in the urban population will increase illiteracy higher the increase in population—higher the illiteracy in urban area.
Increase in population will increase illiteracy and crime rate</li></ul> Increase in population—increase in illiteracy.<br /> Increase in the population– increase in the crime rate.<br />
Relationship b/w VariablesCorrelation, Positive, Negative <br />Correlation: when a proposition says that two or more variables e.g. X and Y are related, it means that they vary together. It means that X is accompanied by a change in Y and vice versa. Such variation is called correlation.<br />Example<br />When the population increases, illiteracy and/or crime rate also increase. The proposition can say that increase in the population, illiteracy and crime rate are correlated. If increase in the population does not influence the literacy ratio or the crime rate, these variables are uncorrelated.<br />
Correlation, Positive, Negative <br /> Positive or Direct relationship: if an increase in the value of one variable is accompanied by an increase in the value of the second variables, the relationship is called positive. <br /> Similarly, if decrease in on variable is accompanied by decrease in other variable, the relationship is called positive or direct.<br /> Negative or Inverse relationship: If increase in one variable is accompanied by decrease in the second variable, the relationship is called a negative or inverse.<br />Example<br />Increase in education level is accompanied by an increase in the income, the relationship is positive. If increase in the education level is accompanied by a decrease in the crime rate, the relationship between the two variables is negative or inverse.<br />
Scales of Measurement<br />Most studies in the Social and behavioral sciences collect data that are in the form of numbers. Usually, it is not the actual numbers that are of interest, but what the numbers represent. Measurement is the assignment of numbers to objects or events according to predetermined rules. Because there are different rules for assigning numbers, the same number have a different meaning, depending on the rules used to assign the number. S.S. Stevens(1951) proposed four levels of measurement that differ in the assumptions that are made regarding the underlying characteristic dimensions to which the numbers apply. <br />
Measurement<br />It indicates that measurement is a doing activity—assigning numbers—which involves performing operations sequentially(following a particular order).<br />It specifies that what you are doing must follow certain rules or a model which lays out the principles of the measurement system.<br />Example<br /> Suppose you are measuring the variable educational attainment. You decide to use “number of years of education completed” to represent the variable of educational attainment. A person finishing elementary school would be assigned an 8 (for eight years of schooling); a person who dropped out of high school in tenth grade would be assigned a 9; a graduate of a two-year college would be assigned a 14; a holder of a master’s degree would be assigned an 18.<br />
Measurement<br /><ul><li>The numbers that are assigned in the process of measurement can represent two different types of attributes: Qualitative or Quantitative.
Qualitative: Reflecting the presence of a quality or attribute.
Quantitative: Reflecting the amount or extent of an attribute.</li></ul>We can measure gender, male or female, by assigning the number 1 to males and the number 2 to females. This is qualitative measurement as the numbers reflect the underlying dimension of gender. These numbers reflect a quality (maleness or femaleness) and don’t provide any quantitative information. We can’t say that women possess more of the quality of gender than do men because they are given a score of 2 while men are only given a score of 1.<br />
Measurement<br />With qualitative measurement the numbers are arbitrary and don’t provide any quantitative information such as rank or distance.<br />If, in measuring gender, we arbitrarily assign the number 37 to men and the number 3.21 to women, we are still performing measurement, since different numbers still accurately reflect the different qualities. The number here function as labels.<br />
Levels or Scales of Measurement<br />Levels of measurements<br />Categorical Numerical<br /> Nominal Ordinal Interval Ratio<br /> Scale Scale Scale Scale <br />Researchers rely on scales of measurement to select statistical techniques.<br />In the social sciences four types of scales for measuring a variable (two types for categorical variables, two types for numerical variables) have been delineated. These scale type ( or levels of measurement, as they are usually called) are useful in helping to classify and catalog variables in a study, as well as in designing questions to measure variables.<br />
Levels of Measurements<br />Nominal Measurement<br /> The assignment of numbers using a nominal scale is a labeling activity. When using a nominal scale, one cannot interpret the numbers as anything other than the names of things. <br /> A variable with a nominal level of measurement consists of a set of distinctive categories that imply no specific order.<br />It is the weakest form of measurement. In it numerals or other symbols are used to classify persons, objects, or characteristics.<br />Example<br /> In physical sciences rocks, can generally be classified into three categories: igneous, sedimentary and metamorphic. A geologist who assigns a 1 to igneous, a 2 to sedimentary, and a 3 to metamorphic has formed a nominal scale. Note that the numerals are simply labels that stand for the respective categories; they have no mathematical significance. A rock that is placed in Category3 does not have more “rockness” than those in Categories 2 and 1.<br /> Gender: Male or Female<br />
Levels of Measurements<br />An essential requirement of nominal scaling is that subjects be classified into mutually exclusive and exhaustive categories. In other words, each subject or observation is assigned to one and only one category, and all observations or subjects are classified into the specific categories.<br />The use of a nominal scale requires a consistent application of an assignment rule.<br />Suppose a researcher wants to indentify which set of variables predicts response to treatment. All subjects are assigned to one of to categories: responder or non-responder. According to the rule of mutual exclusivity and exhaustiveness, subjects are assigned to one or the other category, not both, and all subjects are assigned to one or the other category.<br />The researcher must clearly define what is meant by responder versus non-responder. In this instance, a responder might be defined as someone who falls within the “normal” range of anxiety and a non-responder as someone who falls outside of this range.<br />
Levels of Measurements<br />Ordinal Measurement<br /> Variables that have two or more categories with an inherent order among them are measured at an ordinal level of measurement. <br /> Nothing is specified with regard to the distance between any two rankings i.e., this level possesses the property of equivalence.<br /> It also possesses the property of order among the categories. Any given category can be defined as being higher or lower than any other category.<br />Example<br /> While measuring the variable “socioeconomic status” by categorizing families according to class: lower, lower middle, middle, upper middle, or upper. A rank of 1 is assigned to lower, 2 to lower middle, 3 to middle, and so forth. In this situation, the numbers have some mathematical meaning: families in category 3 have a higher socioeconomic status than families in category 2.<br /> All families placed in a category are treated equally, even though some might have greater incomes than others. <br />
Levels of Measurements<br />Interval Measurement<br /> An interval scale is a created scale that has clearly defined intervals between the points on the scale, and it has order; but it has no true zero point.<br /> It assist us in ordering things quite precisely. <br /> An interval level of measurement has separate categories, like nominal scales, and also has ordered categories, like ordinal scales; but in addition, the distance between the points on an interval scale can be determined mathematically and precisely.<br /> These are used for continuous variables that can register very small differences between categories.<br />Example<br /> Think of a thermometer. It has lines marking off points on the scale to register the changing temperatures. However, there is no true zero point—no point at which there would be no temperature. If it is a Fahrenheit thermometer, the zero on the scale will be at 32 degrees below the freezing point of water; if it is a centigrade thermometer, the zero on the scale will be at the freezing point of water. In neither case, however, will the zero refer to a point where there is no temperature.<br />
Levels of Measurements<br />Ratio Measurement<br /> A ratio scale encompasses all the qualities of the earlier forms of scale: it must have more than one category; it must have an implicit order; it must be able to determine the exact distance between the intervals. In addition, however, it must have a true zero point.<br />Example<br /> Income, age, number of children, or cost of housing. Your income could be zero; you might not have children; your rent could be free.<br />
References<br />Mass Media Research- Wimmer, Dominick<br />Social Research Methods- Neuman<br />Contemporary Communication Research Methods- Smith<br />http://www.socialresearchmethods.net/kb/index.php<br />http://www.une.edu.au/WebStat/unit_materials/index.htm<br />