Xpieaisiiossmoijgsjg|T\g/tt9!A9HBTIBAyM9TA\^T[VIt0：-61SI：-lJiCauaim，tuq^djMioaj乘uJ律H73呻彳時?g呷聊j*61.WTHLd[ULUliJcJIJ—I~ponreQ_..—[口口J虺-MllPfMieuraco3icliaiiaum^*)區(qū)國sdKj3[qvuej3[60$89TIO|-,JguoHE8_|:juaTiirtf-，!nceyalIamsira網(wǎng)suuiqo^)^UISSIJ^ggnpej\R^ISpeUTOQQWM邱1gmejQI顧回圄顧回屋昆阿PH電西網(wǎng)7而I割面方dpHmopiq甌S9RHBn理di&瑚aziCfeuyuuojsureiJ即可水晦jipgapg10FPH歌日QSSJS-P哂皿卻X主洽MWajSfmss990id踞d$|、<?1/MB!A咽丹戲AyMKA珥下。\|?1*-5c帆61SI：-lGlOO'Ol0100'6600'88OOZL00'9900'fS00't00工￡007Iari\—I8AIJ8^Ile/^I6AX6AIBAIEAI6AT^KJEAuosraduosjM:[[|n|U|T|L!|_J*網(wǎng)同日|E||5匾|?|B|^d[通mepu而S9RHnn謎血康9zK[6uyuuojsurejJ即6丘小吭;ipg弗百工叩啊BflBQSSdS-panBnfflgYBWW鳥&基哥：irrssds魏瞰遍中ValueLabelsEEValueLabelsValue:[TOValueLabel:[JOK|注杖I~1-oo5311:personFileEditViewDataTransformAnalyzeGraphsUtilitiesWindowHelp|田Untitled?|。昌度宜|］亶信度的^新.qc-Micg.|［藺IMW-SPSSDaiaEditor翟］未命名j?小,牙家|詢?nèi)蚨噗?|S33T^F04:00ta^iewXVanableVie^v)EaCtdA|pason|■SPSSDataEditor聲§普聲|SFSSProcessorisready§聲§聲§聲§聲Il>一踣精Piocessoi一踣精PiocessoiisreadyReliabilityAnalysis本次課題包含：※如何進行各種信度分析再測信度(Test-RetestReliability)復(fù)本信度(Alternate-formReliability)折半信度(Split-halfReliability)內(nèi)部一致性(InternalConsistencyCoefficient)【計算a系數(shù)】KR20【壹】再測信度(Test-RetestReliability)某空間性向測驗有20題單選題，分別在十月與第二年四月施測同一組10名學(xué)生，以下是測驗結(jié)果，請計算信度。PersonABCDEFGHIJOct18165131516125810Apr18186161716145711步驟一按【Analyze】—【Correlate】—【Bivariate?】?liiilUntitled-SPSSDataEditoiFileEditViewDataTransfoimHGraphsUtilitiesWindowHelpog|2|昌|回Rejiorts?DescriptiveStatistics卜CiistomTables：?CompareMeajis?jJenerslLmeaiModel卜MixedModels??匚1磷雁1廠personoctapr,varvarvarvarvarvar1A18.018.02B16.018.01!3C5.006.C0CorrehteBivaiiaie...Re.airesaon卜L-o.gliiieai?L1-.-Uh-Partial...4D13.016.05E15.017.06F16.016.0DataRed.u.ctionRc：；=di=!7G12.014.0?8H5.005.00Nmnp：3i：3Jjie1xicTestsTmeSeries9I8.007.C0?10J10.011.0Siir-.dval?MultipleFlespcoiseMissingY^lueAnalysis...131924254303132I?|\DataViewXVariableView/Bivaiiaie|SPSSProcessorisready步驟二會出現(xiàn)下面的對話框，將左邊兩變項選入右邊「Variables」內(nèi)，在「CorrelationCoefficientsJ方盒內(nèi)選取「□Pearson」；在「TestofSignificanceJ方盒內(nèi)選取「□Two-tailed」；勾選最下面的「□FlagsignificantcorrelationsJ，之后按口式|鍵。補充若想呈現(xiàn)平均及標準差則可在按。k鍵前按①和眼..進入下個對話框，在Statistics的方盒內(nèi)選取「□Meansandstandarddeviations」，按繼續(xù)。DescriptiveStatisticsMeanStd.DeviationNOCT11.80004.6619010APR12.80005.1380910

CorrelationsOCTAPROCTPearsonCorrelation1<972*nSig.(2-tailed)..000N1010APRPearsonCorrelation.972**1Sig.(2-tailed).000.N1010**.Correlationissignificantatthe0.01level(2-tailed).紙筆計算結(jié)果:(1720/10)-(1720/10)-(11.8)(12.8)r=xx(4.4226)(4.8744)172-151.04=.97228=.9721.557521PersonABCDEFGHIJXoOct1816513151612581011.84.4226Apr1818616171614571112.84.8744X1X2324288302082552561682556110￡X1X2=1720N=10LMJ復(fù)本信度(Alternate-formReliability)某自我概念量表，有20題是非題，題本A與B分別施測同一組10個人，分數(shù)愈高表示愈具有正向自我概念，以下是施測結(jié)果，請計算信度。PersonABCDEFGHIJFormA161214109111391612FormB1512151010121491613做法與再測信度相同，請參照。CorrelationsDescriptiveStatisticsMeanStd.DeviationNFORM_A12.20002.5733710FORM_B12.60002.4129310

CorrelationsFORM_AFORM_BFORMAPearson1Correlation<963*:Sig.(2-tailed)..000N1010FORM_BPearsonCorrelation.963**1Sig.(2-tailed).000.N1010**.Correlationissignificantatthe0.01level(2-tailed紙筆計算結(jié)果:PersonABCDEFGHIJXoFormA16121410911139161212.22.4413FormB151215101012149161312.62.289XAXB2401442101009013218281256156牧aXb=1591N=10159.1-153.725.5883798=.9627=.96159.1-153.725.5883798=.9627=.96XX(2.4413)(2.2891)【參】折半信度(Split-halfReliability)6題的隨堂測驗施測5位學(xué)生，Y表示答對，N表示答錯，以下是測驗結(jié)果，請計算信度。PersonItem1Item2Item3Item4Item5Item6JoeYYYYNYSamYNNYNYSueYYNYYYPegNYNNYNGilNYNNYY

步驟一輸入數(shù)據(jù)步驟二轉(zhuǎn)換數(shù)據(jù)為數(shù)字按【Transform】—【Recode】—【IntoSameVariable】…出現(xiàn)下面的對話框后將左邊方格內(nèi)iteml?item6選至右邊StringVariables內(nèi)后點選OldNev/V^s...鍵按|SPSSProcessorisready按【Transform】—〕Compute…】即出現(xiàn)下面的對話框步驟三將string的屬性改為numericIDnSPSS-SPSSDataEditor|c?XIFileEditViewDateTransformAnalyzeGraphs|SPSSProcessorisready按【Transform】—〕Compute…】即出現(xiàn)下面的對話框NameTyieWidthDecdmalsLabelValuesMissingColmnnS心Measme1jzftSOnString80■foneNone8LeftNorninal2item」Numeric80■JoneNone8RightNorinirial3itern_2Numeric80■IoneNone8RightNorLiiriAlNorLiirial4itercij!Numeric80■IoneNone8Right5item_4Nuiii&ic80■fonBNew8RightNciiiinal6item_5Nuiii&ic80■fonBNew8RightNciiiinal7item_6'i-irliLgI...80■foneNone8LeftNorininal910fVATfolkloTvno國歸二Numeric；rCommaWidth:|iT...,rDot-'匚皿iT.*DecimalPlaces:UTTnCScientificnotationHelpCDaterDollarCCustomcuneiucy1516212627步驟ft./pn、:四計算奇數(shù)題與偶N漢題的、f—vjij，V\\4P宜9|昌I國|丑J』?I.虹圜蟲|氏|T@||口、DataViewXV&nableView/

（此為奇數(shù)題的范例，偶數(shù)題亦然）結(jié)束后便會在spssDataEditor對話框中出現(xiàn)奇數(shù)題和偶數(shù)題的和

步驟四執(zhí)行BivariateFlagsi^xiificaiLtcorrelationsTestofSigmiicanjcef*Twci-tailedCorrelationsDescriptiveStatisticsMeanStd.DeviationNODDS1.4000.547725EVENS2.2000.836665ODDSEVENSODDSPearsonCorrelatioi1U37T】Sig.(2-tailed)..053N55EVENSPearsonCorrelatioi.8731Sig.(2-tailed).053.N55Correlations紙筆計算結(jié)果StudentJoeSamSuePegGilXoOdds212111.4.4899Evens323122.2.7483X1X262612X1X2=17I.計算兩個”半測驗”的相關(guān)N=5rXX(17/5)-(1.4)(2.2)=="29=幻.3665922(.4899)(.7483)II校正相關(guān)系數(shù)為折半信度Spearmen-Brownprophesyformula史比校正公式（當兩個半測驗變異數(shù)相等時使用）NEWXX(N)(curren[)1+(N-1)(currenty)NEW=(2)(0.8729)=_17451=.9321=.931+(2-1)(.8729)1+.8729Guttmanprophesyformula哥德曼校正公式（當兩個半測驗變異數(shù)不等時使用）r=2(1-Co2+Ce2)OEb2XNEW=2(1-(0.5477)2+(0.8367)2)=2(1-0^)=0.888oe(13416)21.8(1.3416)2步驟一步驟二ItelieitelieiteIteIte*折半信度*折半信度也可直接使用SPSS計算步驟一步驟二ItelieitelieiteIteIte輸入數(shù)據(jù)按【Analyze】一【Scale】一【ReliabilityAnalysis］將左邊方格內(nèi)的變項依所需次序分前后半選入右邊items的方格內(nèi)，在左下角的Model框中選取Split-half后按泗瞄..鍵，再按QE.|■ReUabilityAnalysisReliability******Method1(spacesaver)willbeusedforthisanalysis******RELIABILITYANALYSIS-SCALE(SPLIT)ReliabilityCoefficientsNofCases=5.0NofItems=6Correlationbetweenforms=.8729Equal-lengthSpearman-Brown=.9321GuttmanSplit-half=.8889Unequal-lengthSpearman-Brown=.93213Itemsinpart13Itemsinpart2Alphaforpart1=-2.5000Alphaforpart2=.0000標準的折半信度算法：先求各題的難度，再依難度加以排序，再施行上述的方法DescriptiveStatisticsNSumNSumMeanITEM_153.60ITEM_351.20ITEM_553.60ITEM_254.80ITEM_453.60ITEM_654.80ValidN(listwise)5透過平均值可看出其難度平均高的難度低平均低的難度高排序由難度低到高261453,在丟入變項時依單偶分為：215、643兩組，排列數(shù)據(jù)時前后排列。CorrelationsCorrelationsCorrelationsSFSPearsonCorrelation1.000Sig.(2-tailed).1.000N55FPearsonAAACorrelation.0001Sig.(2-tailed)1.000.N55Correlations******Method1(spacesaver)willbeusedforthisanalysis******RELIABILITYANALYSIS-SCALE(SPLIT)ReliabilityCoefficientsNofCases=5.0NofItems=6Correlationbetweenforms=.0000Equal-lengthSpearman-Brown=.0000GuttmanSplit-half=.0000Unequal-lengthSpearman-Brown=.0000>Note#11999>Thecorrelationbetweenforms(halves)ofthetestisnegative.This>violatesreliabilitymodelassumptions.Statisticswhicharefunctionsof>thisvaluemayhaveestimatesoutsidetheoreticallypossibleranges.3Itemsinpart13Itemsinpart2Alphaforpart1=-.9000Alphaforpart2=.6923

【肆】內(nèi)部一致性(InternalConsistencyCoefficient)【計算a系數(shù)】有5題問答題的隨測驗施測5名學(xué)生，每題問答題配分是5分，以下是施測結(jié)果，請計算信度。personItem1Item2Item3Item4Item5Joe34435Sam43433Sue23323Peg44534Gil32433Dot32323步驟一輸入數(shù)據(jù)國SPSS-SPSSDataEditor13回區(qū)］FileEditViewDataTransfoimAnalyzeGraphsUtilitiesWindowHelp洛舊I昌I期6E回七IM創(chuàng)用隹I耋I垂席I劍|1:personpersonitem_2item_3itern_4item_5vanwwww1Jce344352Sam434333Sue233234Peg445345Gil32-4336Dot32.3238y141520212t527r步驟一按【Analyz:e】—-【Scale】-—【ReliabilityAnalysiis】正將左邊方格內(nèi)的變項全選入右邊items的方格內(nèi)，在左下角的Model框中選取Alpha后按-tetistk：s-鍵。

國SPSS-SPSSDataEditorFileEditViewDataTransform崔GraphsUtilitiesWindowHelp咨R膺|國|gJ國|ReEoiis?DescriptiveStatistics?Customlables?CompareMeans?jJeneralL-meaiModel，MixedModels卜correlate?Re.sxesiion?■Lo.gliiieai卜-Classify?-DataReduction卜|也|匚|@1刨▲1:personJcpersonitem.:!itan_4item_5varvai1Jce4352:血「i4333Sue32.34Peg5345Gil4336Dot323IScale■Rpli：=t?nli+,.rAriab.rs-i-?Nonparajji&tricTests?Tim匚Series?MiiltidDjienaonslScaliis：...MdlidmieaaonalScslii.gJPROXSCAL)...Sur-Zivsl?r11MultipleR嘴ptmse?_MissingValueAnalysis...hi17222_:：28約31_■ReliabilityAnalysis11Itemitem_3OKPaste□item_3Resetitem_41也］715Cancel|Help摩偵：|Alpha三］I-ListitemlabelsStatistics...I步驟三出現(xiàn)下列對話框候選取下列勾選處，后按。皿血址|鍵

Reliability******Method2(covariancematrix)willbeusedforthisanalysis******RELIABILITYANALYSIS-SCALE(ALPHA)CorrelationMatrixITEM_1ITEM_2ITEM_3ITEM_4ITEM_5ITEM_11.0000ITEM_2.29701.0000ITEM_3.7647.59411.0000ITEM_4.6860.4330.85751.0000ITEM_5.1588.8018.4763.46291.0000NofCases=6.0Item-totalStatisticsScaleScaleCorrectedMeanVarianceItem-SquaredAlphaifItemifItemTotalMultipleifItemDeletedDeletedCorrelationCorrelationDeletedITEM_113.00006.4000.5251.6471/.8472\ITEM_213.16675.3667.6757.7500.8116\ITEM_312.33335.4667.8333.8588.7642ITEM_413.50006.7000.7481.7857\.8093}ITEM_512.66675.8667.5922.7143\.8333/ReliabilityCoefficients5itemsAlpha=Q8457yStandardizeditemalpha=.8609紙筆計算結(jié)果Question12345TotalScoreJoe3443519Sam4343317Sue2332313Peg4453420Gil3243315Dot3232313.4722.6667.4722.2222.5833ai2=X=16.1667,a2=7.4722a=2.7335k=5N=6

a=—(1一4.4722+.6667+.4722+.2222+.58337.4722)=(1.25)(1-2.4166a=—(1一4.4722+.6667+.4722+.2222+.58337.4722)=(1.25)(1-2.41667.4722)=(1.25)(1-.3234)=.84575=.85有6題的隨測驗施測5名學(xué)生，Y表示答對，N表示答錯，以下是施測結(jié)果，請計算信度。Question123456AYYYYNYBYNNYNYCYNNYNNDNYNNNNENYNNYY步驟一轉(zhuǎn)換數(shù)據(jù)Y正確為1，N不正確為0步驟二與題四求內(nèi)部〕致性的步驟相同Reliability******Method2(covariancematrix)willbeusedforthisanalysis*