Topic3ValidityValidityDefinition:Atestissaidtobevalidifitmeasuresaccuratelywhatitisintendedtomeasure.Typesofvalidity:

ContentvalidityCriterion-relatedvalidityConstructvalidityFacevalidity3.1Contentvalidity

Atestissaidtohavecontentvalidityifitscontentconstitutesarepresentativesampleofthelanguageskills,

structures,etc.withwhichitismeanttobeconcerned.Inordertojudgewhetherornotthetesthasacontentvalidity,weneedaspecificationoftheskillsorstructures,ect.thatitismeanttocover.AcomparisonoftestspecificationandtestcontentisthebasisforjudgmentsastocontentvalidityTheimportanceofcontentvalidityFirstly,thegreateratest’scontentvalidity,themorelikelytobeanaccuratemeasureofwhatissupposedtomeasure.Secondly,suchatest,atest-specificationsnotrepresentedatall,islikelytohaveaharmfulbackwasheffect.Areaswhicharenottestedarelikelytobecomeignoredinteachingandleaning.3.2

Criterion—relatedvalidity

(效標關聯(lián)效度)

Definition:ClassificationDefinition:

Anotherapproachtotestvalidityistoseehowfarresultsonthetestagreewiththoseprovidedbysomeindependentandhighlydependableassessmentofthecandidatesability.(CET摸擬題與CET實考題的各項分數(shù)的相關性)效標就是考察檢定考試效度的參照標準，效標關聯(lián)效度也就是與該參照標準有關的效度。效標關聯(lián)效度就是以某一種測驗分數(shù)與其效標分數(shù)之間的相關來表示的效度。換言之，一組考生參加某種考試的分數(shù)與同一組考生參加另一公認、的有較高效度的測驗結果的一致性程度就是該考試的效標效度。Therearetwokindsofcriterion—relatedvalidity:

ConcurrentvalidityConcurrent(共時性)validityisestablishedwhenthetestandcriterionareadministeredataboutthesametime.(例如:設計一次口語測試，因為沒有充足的時間去測試課程標準要求的全部語言功能，所以，對每位測試者進行10鐘的口試。但我們不能確定10分鐘的口試能否反映被試著的實際口語能力。我們先設計了一次考試，考試時間45分鐘，并覆蓋較多的語言功能，以此，建立了一個口試標準；再將著10分鐘的口語考試結果與這個標準作相關分析，然后確定這十分鐘的效度。）Anoralachievementtest

Courseobjectives

Alargenumberof‘functions’

45minutes10minutes?FulltestscoresshortenedtestscoresEvidenceforconcurrentvalidity:Degreeofagreementcalculationofdegreeofagreement:

correlationcoefficient(possiblevalue:-1~1,thehigherthebetter)1:perfect(positive)agreement0:zeroagreement-1:negativeagreementAtestoforalabilityforahighleveldiplomaticpostNotrationaltoruntheriskofappointingsomeonewithinsufficientoralabilityLoweragreementNOTallowedAbriefinterviewformingpartofaplacementtestLoweragreementallowedCalculatingcorrelationcoefficientinSPSS.PredictiveValidityThisconcernsthedegreetowhichatestcanpredictcandidates’futureperformance.e.g.Couldaproficiencytestpredictastudent’sabilitytocopewithagraduatecourseataBritishuniversity.Discussion:Validatingaplacementtest.How?3.3ConstructValidity(結構效度):Atest,partofatest,oratestingtechniqueissaidtohaveconstructvalidityifitcanbedemonstratedthatitmeasuresjusttheabilitywhichitissupposedtomeasure.Theword‘construct’referstoanyunderlyingability(ortrait)whichishypothesizedinatheoryoflanguageability.Forexample,whentestingreadingability,wehypothesizethattheabilitytoreadinvolvesasetofsub-abilities,suchasguessingunknownwordmeaningsfromthecontext.Thenweshouldrefertoresearchliteraturetoestablishwhetherornotsuchanabilityexistedandcouldbemeasuredinapartofthetest.

Ifwewereabletodemonstratethatwewereindeedmeasuringthatability,thenwesaythetesthasconstructvalidity.結構效度研究是一個搜集證據(jù)證實某一考試的確在測試我們想測試的那種心理屬性的過程。一個考試如果被證實是在測量我們想測量的那種能力，那么這個考試就具有較好的結構效度，它是一個編制假設再證實假設的過程。Howdoweknowthatthetestismeasuringwritingability?

Hypothesis:Ourtheoryofwritingtellsusthatunderlyingwritingabilityareanumberofsub-abilities:controlofpunctuationsensitivitytostylesstructurescoherenceandcohesion….Inordertomeasurethesesub-abilities,weadministerapilottestwhichinvolvestwosteps:用選擇題測試寫作能力（間接考試）步驟：找一組被試（第一次參加考試），對他們的寫作能力全面抽樣（讓他們做包含標點、語法、體裁等的選擇題）；把測試結果與范文相比較：如果結果與范文相關系數(shù)高，說明可以用選擇題測試寫作能力1.Weobtainextensivesamplesofthewritingabilityofthegrouptowhomtestisfirstadministered,andhavethesereliablyscored.2.Wethencomparescoresonthepilottestwiththescoresgivenforthesamplesofwriting.Iftheagreementishigh,wearemeasuringwritingabilitywiththetest.按照假設分別測量各項能力：假設：標點語法體裁寫作能力↙↓↘／∣﹨／∣﹨／∣﹨單選填空改錯單選填空改錯單選填空改錯7066685560628580823.4FaceValiditye.g.Anoraltestdoesn’trequirecandidatestospeakwouldnothavefacevalidity.Atestissaidtohavefacevalidityifitlook