統(tǒng)計描述Stata實現(xiàn)

1、第二章統(tǒng)計描述的Stata實現(xiàn)本章主要用到的Stata主要命令*描述性統(tǒng)計：summary 變量名 weight = ,detail頻數(shù)表：tabulate 變量名weight =計算均數(shù)：means 變量名 weight =頻數(shù)分布圖：histogram 變量名,bin() start()*Stata的命令可以用前幾個字母代替，如 summary可以由su代替，tabulate可以由tab代替。下面就直接用縮寫，不再贅述。例2-1某市1982年調(diào)查120名20歲男子的身高（cm）資料如下，試編制 頻數(shù)表和頻數(shù)圖164.4169.2174.7175.0165.0162.8170.2160.11

2、70.9170.1175.5170.3172.3168.2166.7171.7166.8171.6165.2172.0171.7168.8171.8174.5171.7172.7166.3174.0169.0174.5171.8174.9180.0173.5178.1169.7176.1181.3173.8164.7172.2172.8178.6172.0182.5172.0173.5166.3176.1169.2176.4168.4171.0169.1166.9181.1170.4165.2168.0172.4164.3166.9176.4164.2177.2168.3177.8173.01

3、67.4173.2169.9172.1170.4174.8172.1164.4170.5165.0172.8166.8175.8171.1174.8172.7169.4178.2174.1177.2170.0172.7168.3177.1172.5166.3175.1174.4162.3171.3177.0163.5168.8177.6175.2171.5172.5179.1172.6184.8168.3182.8170.3167.5171.2170.4166.9178.5164.1169.5173.4170.3數(shù)據(jù)格式如下:x1164.42175.53171.74171.85172.2617

4、6.47164.38169.99175.810168.311168.812170.313169.214170.315168.816174.917172.818168.419166.920172.121171.122177.123177.624167.525174.726172.327171.82818029178.63017131176.432170.433174.834172.535175.236171.23717538168.239174.540173.54117242169.143164.244174.845172.746166.347171.548170.44916550166.751

5、171.752178.153182.554166.955177.256172.157169.458175.159172.560166.961162.862171.763172.764169.76517266181.167168.368164.469178.270174.471179.172178.573170.274166.875166.376176.177173.578170.479177.880170.581174.182162.383172.684164.185160.186171.68717488181.389166.390165.2911739216593177.294171.395

6、184.896169.597170.998165.299169100173.8101176.1102168103167.4104172.8105170106177107168.3108173.4109170.1110172111174.5112164.7113169.2114172.4115173.2116166.8117172.7118163.5119182.8120170.3為制作頻數(shù)表，鍵入Stata命令：.gen f=int(x-160)/2)*2+160產(chǎn)生用以作頻數(shù)表的新變量“ f”.tab f對變量“ f ”作頻數(shù)表“gen”命令產(chǎn)生新變量“ f”，將各觀察值轉(zhuǎn)換成相應(yīng)該組的下限

7、值。int 為取整函數(shù)，結(jié)果為括號內(nèi)函數(shù)值的整數(shù)部分，如 int(3.24)=3 。 “160”為第 一組的下限，“ 2 ”為組距。以第一例觀察值160.1cm為例， f=int(160.1-160)*2+160=160,則它應(yīng)歸入 “ 160”組。結(jié)果如下:f | +_Freq.PercentCum160 |10.830.83162 |32.503.33164 |108.3311.67166 |119.1720.83168 |1613.3334.17170 |2218.3352.50172 |2218.3370.83174 |1411.6782.50176 |108.3390.83178 |

8、54.1795.00180 |32.5097.50182 |21.6799.17184 |10.83100.00+Total |120100.00作頻數(shù)圖hist f,start(160) width(2)結(jié)果如下:wwfsneD例2-2某醫(yī)生測定230名正常成年男子的空腹血清胰島素樣生長因子-1(IGF-1F)水平，整理后編制為頻數(shù)分布表(表 2-2),請根據(jù)該頻數(shù)分布表作 頻數(shù)圖。表2-2 230名正常人空腹血清胰島素樣生長因子-1水平中位數(shù)的計算IGF-1F人數(shù)f(2)頻率(%) (3)累計頻數(shù)三f(4)累計頻率()(5 = 4 /n503013.043013.01507130.8710

9、143.92504921.3015065.23502812.1717877.4450146.0919283.5550125.2220488.7650104.3521493.075083.4822296.585052.1722798.7950105031.30230100.0合計230100.00一一Stata數(shù)據(jù)格式如下:x f150302150713250494350285450146550127650108750898505109503tab x weight=f制作匯總表的頻數(shù)表在本傷J中，x的取值為各組的下限，而f是各組的頻數(shù)，所以需要用 Stata 的頻數(shù)選項weight=。結(jié)果：x

10、 | Freq. Percent Cum.+50 |3013.0413.04150 |7130.8743.91250 |4921.3065.22350 |2812.1777.39450 |146.0983.48550 |125.2288.70650 |104.3593.04750 |83.4896.52850 |52.1798.70950 |31.30100.00+-Total |230100.00hist x weight=f,start(50) width(100)制作頻數(shù)圖wvfsneD20040000 O例2-3 隨機測量某地10名2030歲健康男性居民血清鐵含量（mol/L）, 測量

11、值分別為 6.58 , 7.42 ,15.32 ,15.78 , 17.60 ,17.98 , 15.21 ,17.53 , 20.11 , 22.64 ,試求其平均血清鐵含量。Stata數(shù)據(jù)為1x6.5827.42315.32415.78517.6617.98715.21817.53920.111022.64su x計算均數(shù)標(biāo)準(zhǔn)差結(jié)果:Variable | +-ObsMeanStd. Dev. Min Maxx |1015.6175.0752546.5822.64所以平均平均血清鐵含量為15.617例2-4 計算例2-1的頻數(shù)表（表2-3）中120名男子的平均身高顯然，如果用Stata ,對

12、于例2-1的資料不需要事先轉(zhuǎn)化為頻數(shù)表，然后再su x weight=f計算均數(shù)。不妨用例2-2作為例子展示Stata計算頻數(shù)表資料的均數(shù)計算頻數(shù)表資料的均數(shù)結(jié)果:Variable | Obs +WeightMean Std. Dev.MinMaxx |10230290223.412350950例2-5 8名麻疹易感兒接種麻疹疫苗3周后，其血凝抑制抗體滴度分別為1:4, 1:8, 1:16, 1:32, 1:64, 1:128, 1:256, 1:512。試求其平均抗體滴度。數(shù)據(jù)格式:x4816326412825651212345678gen logx=log(x)計算x的自然對數(shù)su log

13、x計算自然對數(shù)的均數(shù)，得到 3.81231disp exp(3.81231)計算均數(shù)的反對數(shù)，即指數(shù)得到幾何均數(shù)為45.254857另外，Stata還有一個直接計算幾何均數(shù)的命令:means x結(jié)果為：Variable | Type+ObsMean95% Conf. Intervalx | Arithmetic8127.5-20.53203 275.532| Geometric845.2548310.94481 187.1206| Harmonic816.06275.Missing values in confidence interval(s) for harmonic mean indic

14、ate that confidence interval is undefined for corresponding variable(s). Consult Reference Manual for details.Means命令計算算術(shù)均數(shù)、幾何均數(shù)以及調(diào)和均數(shù)，其中幾何均數(shù)為45.25483,和前面的結(jié)果相同。例2-6 某地區(qū)50名麻疹易感兒童接種麻疹疫苗3周后，測其血凝抑制抗體滴度，如表2-4中第(1)欄和第(2)欄，求平均抗體滴度。表2-4 50名麻疹易感兒童平均抗體滴度計算表抗體滴度(1)人數(shù)f(2)滴度倒數(shù)X(3)lg X (4)f lg X(5) = (2) X (4)1:4

15、140.60210.60211:8280.90311.80621:166161.20417.22461:3210321.505115.05101:6416641.806228.89921:12881282.107216.85761:25652562.408212.04101:51225122.70935.4186合計50一一87.9003數(shù)據(jù)如下：xf14128231664321056416612887256585122gen logx=log(x)計算x的自然對數(shù)su logx weight=f計算自然對數(shù)的均數(shù)，得到4.04798disp exp(4.04798)計算均數(shù)的反對數(shù)，即指數(shù)得

16、到幾何均數(shù)為57.281631也可以用means x weight=f結(jié)果：(analytic weights assumed)Variable | Type Obs Mean 95% Conf. Interval +x | Arithmetic 895.76-2.30755 193.8275| Geometric 857.281622.59723 145.2029| Harmonic 8 32.82051.Missing values in confidence interval(s) for harmonic mean indicate that confidence interval i

17、s undefined for corresponding variable(s). Consult Reference Manual for details.例2-7 5名成年男子的體重(kg)為60, 70, 75, 80, 90,求中位數(shù)。例2-8某醫(yī)生測定了 6名正常成年男子的空腹血清胰島素樣生長因子-1(IGF-1F)水平為 150, 170, 185, 245, 265, 280,求中位數(shù)。注：這兩個例子用軟件算實在是浪費，可以用 |su x,detail |來處理。對于summary加上detail后可以輸出百分位數(shù)，其中的 P50就是中位數(shù)。xPercentilesSmallest1%60605%607010%6075Obs525%7080Sum of Wgt.550%75Mean75LargestStd. Dev.11.1803475%807090%9075Variance12595%9080Skewness099%9090Kurtosis2.05例2-9 8 名7歲男孩的身高(cm)分別為116, 118, 119, 120, 121, 123,125, 126。請計算這8名7歲男孩身高的第25百分位數(shù)和第90百分位數(shù)。Stata數(shù)據(jù)為：Stata命令