Bài giảng Chapter 6 Measurement

Chapter 6MeasurementAssociated with quantitative studiesNumbers used as a tool for identifying and presenting informationLinks the conceptual to the empiricalNecessary to conduct quantitative research1Measurement PrinciplesNumbers measure value, intensity, degree, depth, length, width, distanceDescriptive and evaluative deviceNumbers have no value until we provide meaning Includes everything the researcher does to arrive at a numberDetails the operationalization of the variable2Levels of MeasurementDiscrete or continuous Both representative of communication phenomenaEach produces different kind of dataHow data are collected determines how they can be used in statistical analyses3Discrete DataThe presence or absence of some characteristicAlso known as nominal or categorical dataCategories Reflect different types not differing amountsHave no inherent valueMust be mutually exclusive, exhaustive, and equivalent4Continuous Level DataReveals quantity, intensity, or magnitudeValues that differ in degree, amount, or frequency can be ordered on a continuumThree typesOrdinal dataInterval dataRatio data5Ordinal DataRanks elements in logical numerical orderSequence suggests value of dataRanking positions are relative Distance between ranks is unknownZero does not exist6Interval DataIdentifies highest, next highest, and so onIdentifies exact difference between and among scoresAcknowledges zeroAllows meaningful comparisonsLikert-type scalesSemantic differential scales7Ratio DataAll of the characteristics of interval data Zero is absoluteIndicates complete lack of the variable measuredProvides measure of degree to which something actually exists8ValidityExtent to which it measures what you want it to measure and not something elseValidity is a matter of degreeInternal validityFace validityContent validityCriterion-related validity: concurrent or predictiveConstruct validity9ReliabilityDegree of consistency among similar itemsReliability coefficient – 0.0 to 1.0Closer to 1.00, the greater the degree of reliabilityGenerally, above .70 is acceptableInternal reliabilityItems invoke same responseReliability between codersTest-retest reliabilitySplit-half reliability10Relationship between Validity and ReliabilityA measurement should be both valid and reliableValidity and reliability connected in fundamental waysReliable measurements can be obtained without validityWhen validity is achieved, reliability is presumed11Threats to Validity and ReliabilityIssues of data collectionInternal validityReliability over timeIssues of sample representativenessExternal validityEcological validityAlternative explanations12Issues of Data InterpretationResearchers responsible for Collecting data accurately and ethicallyInterpreting and reporting data responsiblyQuality of data interpretation cannot be better than quality of data collectedMeasurement is central to quality of outcomes and links to theory13