Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out

Introduction to measurement

VIEWS: 13 PAGES: 16

  • pg 1
									              Lecture 2:
         Types of measurement
•   Purposes of measurement
•   Types and sources of data
•   Reliability and validity
•   Levels of measurement
•   Types of scale


                                1
               Introduction
• Overview of measurement section of course
  – lecture 2: introduction
  – lectures 3-4: screening and diagnostic tests
  – lectures 5-6: scales
• Traditions of measurement:
  – “Clinimetric”: clinical, epidemiological (focus
    on screening and diagnostic tests)
  – “Psychometric”: psychology (focus on scales)
                                                      2
         Introduction (cont)
• Readings:
  – Streiner & Norman
  – Gordis
  – other




                               3
     Purposes of measurement
• Clinical
  – screening, diagnosis, monitoring in individuals
• Surveillance
  – planning and monitoring public health and
    health care in populations
• Research
  – measurement of study variables (determinants,
    outcomes, confounders/modifiers)
                                                      4
            Sources of data
• Primary vs secondary
• Clinical observations
• Questionnaires and interviews
• Reportable diseases and registries
• Health records
• Administrative databases (hospital
  discharges, claims, medication prescription)
• Vital statistics                            5
   Requirements by purpose of
          measurement
• Clinical
  – discrimination between health and disease
    relevant to management
• Surveillance
  – valid measurement of trends
• Research
  – maximize validity of study results

                                                6
         Selection of measures
•   Appropriateness to purpose
•   Feasibility
•   Acceptability
•   Cost
•   Validity
•   Reliability

                                 7
      Validity of measurements
• Synonym: accuracy
• Does it measure what it is intended to?
• Many types:
  – face, content
     • “eyeball” test
     • investigator, expert judgment
  – criterion (lecture 3)
  – construct (lecture 6)
                                            8
   Reliability of measurements
• Synonyms: reproducibility, precision,
  consistency
• Sources of variability?
• Random error vs systematic error (bias)




                                            9
10
       Types of reliability: rater
• Between- and within-rater/observer (inter-
  and intra-rater observer variation)
  –   At single point in time
  –   Independent ratings
  –   May be random or systematic
  –   Examples:
          – blood pressure
          – pathology slides

                                               11
   Reliability: stability over time
• Stability over time/ test-retest (T1 and T2
   – Same conditions and rater at T1 and T2. Why?
   – Random or systematic
   – Example of systematic error:
      • Regression toward the mean: when subjects initially have
        extreme values (more likely to be in error than typical values)
        E.g., population blood pressure screening
      • Questions about undesirable beahviour
      • Uncalibrated instruments



                                                                          12
                    Example
• Measurement of patient depression using
  clinical rating scale
  – Sources of variability?
  – Measurement of inter-rater reliability?
  – Results for 10 patients:
     • Rater 1: 3 patients are depressed
     • Rater 2: 5 patients are depressed
  – Conclusion?
                                              13
    Types of variables (level of
          measurement)
• Continuous (syn. dimensional, quantitative)
  – interval
  – ratio
• Categorical (discrete)
  – dichotomous, binary
  – polychotomous
     • nominal
     • ordinal
                                            14
             Response scales
• Categorical or continuous?
  – Continuous variables can be categorized
  – Categorical (ordinal) variables can be analyzed
    as continuous (pseudocontinuous)
• Example of scales
  – Visual analogue scale
  – Likert scale (agree/disgree)
  – Semantic differential scale

                                                  15
Examples of measures to be used
   in a study: for discussion
• Sources of data? Level of measurement?
  Type of response scale?
  – Use of health services during past year (doctor
    visits, hospitalization)
  – Satisfaction with medical care
  – Use of alcohol and drugs (current and lifetime)



                                                  16

								
To top