Measurement

Measurement

The way to link and analyze abstract concepts to observable events. "What do I need to understand?" "What data will I collect?" The measurement process: Identify and define the concept of interest Specify an observable event Evaluate and revise the observable event It is a quickly moving field  

Dated Measures  Caused by fast paced industry Things change all the time changes what is measured Accurate Measurement units Getting harder to measure time and frecuency Consistency in measurements  

Stage 1: Identify and Define the Concept

This stage is when you determine the general area to be explored and end it with a detailed, explicit definition of what is measured.  

Identify the Concept

All research starts with a need for informat Concept: An invented name for a property of an object, person, state, or event. There are explicit and abstract concepts  

Develop a Definition

A conceptual definition expresses a concept's central idea. Researchers need to be sensitive to the fact that for any particular concept there is a bunch of definitions. The process requires that all involved in the research agree on what is the conceptual definition.

Develop a Definition

Develop a Defintion

Once a concept has been identified and defined, then it is operationally defined. An operational definition translates the concept into an observable event Similar to the different operational definitions of affection, differences in perspective, among researchers often lead to differences in conceptual definitions. Process: Explicitly Specify the concept of interest Explore different aspects of the concepts meaning Explicitly specify what can be observed Evaluate and select one or more of the alternatives identified  Keep in mind the the operational definition states what will be observed.

Stage 2: Specify an Observable Event

Determines what specific questions will be asked to collect information

Identify Level of Measurement

There are four levels of measurement: nominal ordinal interval ratio

Nominal

occurs when the goal is the classification of a characteristic/attribute common in observational or behavioral data collection Categories in nominal measurement are mutually exclusive and collectively exhaustive. This means that every level of the characteristic fits into only one category and each characteristic fits somewhere numbers are assigned to categories/responses, but they are simply labels all objects/people are assumed to be the same   

Ordinal

arranges characteristics or attributes in an ordered relationship to some degree of magnitude a number represents an object's ranking among other objects relative standing - total extent of an ordinal measure's interpretive value relative distance - the degree to which one object ranks from the other on the ordinal level of measurement numbers only represent a place in an ordered array  

Interval

Advanced form of ordinal measurement allows you to determine both the relative ranking of objects and the distance between the objects interval scales do not have an absolute 0 point and as a result you cannot possibly make a statement about how many times higher one score is from another  

Ratio

the highest level of measurement have equal distances and a meaningful 0 point a common form of ratio measurement is the constant sum scale (a respondent distributes an allotted amount of points across a set of objects) ratio level measurements are common for research that focuses on respondent behaviors  

Identify Question Type

once concepts are identified and level of measurement determined, specific questions are written to collect the information open-ended questions:  respondents answer in their own words without influence provide context to interpreting an individual's responses to closed-end questions a good way to introduce a topic area tend to be easier to write than closed-end questions unaided responses Closed-end questions: produce less variability in range of response easier for respondent to answer quicker to administer easier to analyze only should be used when all options are present aided responses

Stage 3: Evaluate and Revise the Event

pilot research should be conducted in order to assess the reliability and validity of any untested measures Reliability: the extent to which the survey produces the same results consistently Validity: the extent to which the survey actually measures what it is intended to measure both are independent but related aspects of measurement      

Assessing Reliability

Test-retest reliability is a common way to assess reliability obtained by surveying the same group of people under equivalent conditions second alternative is to administer a second but similar test if internal consistency is examined, reliability is valid

Assessing Validity

validity only occurs when there is a high degree of correspondence between a concept's operational definition and the specific observable event used to record the concept. Face Validity is the most basic form, means all parties are in consensus Concurrent validity is assessed by comparing a new measurement with an alternative at the same point in time Predictive validity estimated by determining the extent to which performance on one variable predicts performance on another variable (take the SAT for example)

In Conclusion

Measurement is the way the abstract is linked to the observable  Three stages: Identify and define the concept of interest specify an observable event evaluate and revise Level of measurement is determined by the level of detail needed data at higher levels is less limiting data collected at higher levels can be grouped together and downgraded higher levels can be more thoroughly analyzed

Hope this Helps!

    :)