为什么t.test返回R的最小p值2.2e – 16？
当我们在R中进行测试时，两组之间的差异非常大，则测试的p值将打印为2.2e–16，这是假设检验程序R的打印行为。可以通过使用t检验函数作为t.test（“Var1”，“Var2”，var.equal=FALSE）$p.value来提取实际的p值。该p值不可能与2.2e–16相同。
例1
> x1<-1:100 > y1<-100001:110000 > t.test(x1,y1,var.equal=FALSE)
输出结果
Welch Two Sample t-test data: x1 and y1 t = -3617.2, df = 10098, p-value < 2.2e-16 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -105006.9 -104893.1 sample estimates: mean of x mean of y 50.5 105000.5
> t.test(x1,y1,var.equal=FALSE)$p.value [1] 0
例2
> x2<-sample(1:10,50,replace=TRUE) > y2<-sample(500:510,replace=TRUE) > t.test(x2,y2,var.equal=FALSE)
输出结果
Welch Two Sample t-test data: x2 and y2 t = -427.61, df = 12.789, p-value < 2.2e-16 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -500.4179 -495.3785 sample estimates: mean of x mean of y 5.9200 503.8182
> t.test(x2,y2,var.equal=FALSE)$p.value [1] 5.881324e-28
范例3
> x3<-sample(101:110,50,replace=TRUE) > y3<-sample(1001:1010,50,replace=TRUE) > t.test(x3,y3,var.equal=FALSE)
输出结果
Welch Two Sample t-test data: x3 and y3 t = -1730.7, df = 97.907, p-value < 2.2e-16 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -901.4725 -899.4075 sample estimates: mean of x mean of y 105.38 1005.82
> t.test(x3,y3,var.equal=FALSE)$p.value [1] 2.07048e-221
例子4
> x4<-sample(1001:1010,50,replace=TRUE) > y4<-sample(100001:1000010,50) > t.test(x4,y4,var.equal=FALSE)
输出结果
Welch Two Sample t-test data: x4 and y4 t = -14.798, df = 49, p-value < 2.2e-16 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -620129.5 -471835.6 sample estimates: mean of x mean of y 1005.6 546988.1
> t.test(x4,y4,var.equal=FALSE)$p.value [1] 1.043251e-19