第十一章：连续变量的统计推断（一）------------T检验练习一，样本均数与总体均数的比较从一批木头里抽取5根，测得直径如下，是否认为这批木头的平均直径是12.3cmOne-SampleStatisticsNMeanStd.DeviationStd.ErrorMean直径512.460.2881.1288One-SampleTestTestValue=12.3tdfSig.(2-tailed)MeanDifference95%ConfidenceIntervaloftheDifferenceLowerUpper直径1.2424.282.1600-.198.518由上表知总体均数为12.3，T=1.242，P值为0.282可知,由于P值元大于检验水平0.05，因此不拒绝H0，还不能认为样本所在的均数与假设的总体均数不同，即可认为这批木头的直径为12.3.。二，成组设计两样本均数的比较如比较两电子器材的电阻，随机出去的样本测量电阻，是比较两电子器材的电阻是否相同。A批0.1400.1380.1430.1420.1440.1480.137B批0.1350.1400.1420.1360.1380.1400.141GroupStatistics批类NMeanStd.DeviationStd.ErrorMean电阻A批7.14171.003773.001426B批7.13886.002610.000986IdependentSamplesTest直径12.312.812.412.112.7Levene'sTestforEqualityofVariancesFSig.电阻.652568.435IIdependentSamplesTestt-testforEqualityofMeanSigMeanStd.errorTdf(2-tailed)DifferenceDifference电阻Equalvariancesassumed1.64812.1250.0028570.001734Equalvariancesnotassumed1.64810.67110.6710.0028570.001734由上表知：t=1.648，v=12，P=0.125.则P>0.05,故接受H0，认为两者的电阻值是相同的三，适应条件与方差齐性检验批类CaseProcessingSummary批类CasesValidMissingTotalNPercentNPercentNPercent电阻A批7100.0%0.0%7100.0%B批7100.0%0.0%7100.0%Descriptives批类StatisticStd.Error电阻A批Mean.14171.00142695%ConfidenceIntervalforMeanLowerBound.13822UpperBound.145205%TrimmedMean.14163Median.14200Variance.000Std.Deviation.003773Minimum.137Maximum.148Range.011InterquartileRange.006Skewness.440.794Kurtosis-.1421.587B批Mean.13886.00098695%ConfidenceIntervalforMeanLowerBound.13644UpperBound.141275%TrimmedMean.13890Median.14000Variance.000Std.Deviation.002610Minimum.135Maximum.142Range.007InterquartileRange.005Skewness-.492.794Kurtosis-1.2681.587电阻Stem-and-LeafPlots电阻Stem-and-LeafPlotfor批类=A批FrequencyStem&Leaf2.0013.784.0014.02341.0014.8Stemwidth:.010Eachleaf:1case(s)电阻Stem-and-LeafPlotfor批类=B批FrequencyStem&Leaf.0013.3.0013.5684.0014.0012Stemwidth:.010Eachleaf:1case(s)四，配对设计样本均数的比较如为研究女性服用某避孕新药后是否影响其血清总胆固醇，讲20名女性按年龄配成10对。每对中随机抽取一人服用新药，另一人服用安慰剂。经过一定时间后，测得血清总胆固醇含量，问该新药是否影响女性血清总胆固醇？配对号12345678910新药组4.455.84.64.94.865.94.35.1安慰剂组6.25.25.554.45.456.45.86.2PairedSamplesStatisticsMeanNStd.DeviationStd.ErrorMeanPair1新药组5.08010.6197.1960安慰剂组5.51010.6402.2025PairedSamplesCorrelationsNCorrelationSig.Pair1新药组&安慰剂组10.020.956PairedSamplesTestPairedDifferencestdfSig.(2-tailed)MeanStd.DeviationStd.ErrorMean95%ConfidenceIntervaloftheDifferenceLowerUpperPair1新药组-安慰剂组-.4300.8820.2789-1.0609.2009-1.5429.158由上表知P=0.158，故可以认为该新药影响女性总胆固醇《由于样本中新药组与安慰剂组的差值为正数，因此该新药对女性的总胆固醇有影响。
