兩個(gè)群體的比較-Stata教學(xué)課件

兩個(gè)群體的比較-Stata教學(xué)因?yàn)橹形南嗳菪詥栴}有一些亂碼，辨識(shí)不易以j2、j3為例j2問受訪者「拾.2.通常您平均每週大約花多少時(shí)間做家務(wù)工作？_______小時(shí)」j3問受訪者「拾.3.通常您的配偶平均每週大約花多少時(shí)間做家務(wù)工作？_______小時(shí)」?Ming-chiChen2社會(huì)統(tǒng)計(jì)我們的資料裡有變數(shù)標(biāo)籤，但是因?yàn)橄嗳菪缘年P(guān)係會(huì)有亂碼查看是否有亂碼？Data-dataeditor在j2這個(gè)變數(shù)名稱上click一下，下面一整欄的數(shù)值都反白了滑鼠右鍵-variable-properties-label出現(xiàn)的中文是「通常您平均牢週大約花多少時(shí)間做家務(wù)工作︺」把亂碼改好也將j3變數(shù)標(biāo)籤的亂碼改好?Ming-chiChen3社會(huì)統(tǒng)計(jì)查看變數(shù)有無異常值關(guān)掉Dataeditor視窗用boxplot來看有無極端值Graphics-easygraphs-boxplot-main-在variable的空格裡鍵入j2?Ming-chiChen4社會(huì)統(tǒng)計(jì)用boxplot來看有無極端值?Ming-chiChen5社會(huì)統(tǒng)計(jì)同樣方法也可以查看j3的極端值也可以直接在指令欄?Ming-chiChen6社會(huì)統(tǒng)計(jì)這就是指令欄?Ming-chiChen7社會(huì)統(tǒng)計(jì)在指令欄裡直接鍵入Graphboxj2然後按enter?Ming-chiChen8社會(huì)統(tǒng)計(jì)Summarizevarname,detail

指令欄鍵入summarizej2,detail或statistics-summaries,tables,&tests-summarystatistics-summarystatistics?Ming-chiChen9社會(huì)統(tǒng)計(jì).通常您平均每週大約花多少時(shí)間做家務(wù)工作？-------------------------------------------------------------PercentilesSmallest1%005%0010%00Obs192425%20SumofWgt.192475%20998太愛做家事了吧！高得不合理?Ming-chiChen10社會(huì)統(tǒng)計(jì)Recode極端值我們到85q1_format.txt去看，發(fā)現(xiàn)J2J3996"不知道"998"不適用"999"拒答"所以要把995以上定義為systemmissingRecodej2995/max=.這裡的句點(diǎn).就是Stata系統(tǒng)定義的缺失值。?Ming-chiChen11社會(huì)統(tǒng)計(jì).summarizej2,detail通常您平均每週大約花多少時(shí)間做家務(wù)工作？-------------------------------------------------------------PercentilesSmallest1%005%0010%00Obs184925%20SumofWgt.184975%15105一週只有168小時(shí)，所以應(yīng)該合理?yè)Q算，以一天16小時(shí)算，一週112小時(shí)12.inspectj2j2:通常您平均每週大約花多少時(shí)間做家務(wù)工作NumberofObservations-------------------------------------------------------------------------TotalIntegersNonintegers|#Negative---|#Zero305305-|#Positive15441544-|#---------------|#Total18491849-|#....Missing75+---------------------------01681924(47uniquevalues)用inspect來看大致分佈以及缺失個(gè)案數(shù)Data-describedata-inspectvariables13Recodej2168=112?Ming-chiChen14社會(huì)統(tǒng)計(jì).inspectj2j2:通常您平均每週大約花多少時(shí)間做家務(wù)工作NumberofObservations-------------------------------------------------------------------------TotalIntegersNonintegers|#Negative---|#Zero305305-|#Positive15441544-|#---------------|#Total18491849-|#....Missing75+---------------------------01121924(46uniquevalues)15.sumj2,detail通常您平均每週大約花多少時(shí)間做家務(wù)工作？-------------------------------------------------------------PercentilesSmallest1%005%0010%00Obs184925%20SumofWgt.184975%1510516.inspectj3j3:通常您的配偶平均每週大約花多少時(shí)間做家NumberofObservations-------------------------------------------------------------------------TotalIntegersNonintegers|#Negative---|#Zero263263-|#Positive16611661-|#---------------|##Total19241924-|#...#Missing-+---------------------------09991924(54uniquevalues)17.summarizej3,detail通常您的配偶平均每週大約花多少時(shí)間做家務(wù)工作？PercentilesSmallest1%005%0010%00Obs192425%40SumofWgt.192475%99699818Missingvalue&recodeRecodej3990/max=.Recodej3168=112?Ming-chiChen19社會(huì)統(tǒng)計(jì).recodej3168=112(j3:4changesmade).inspectj3j3:通常您的配偶平均每週大約花多少時(shí)間做家NumberofObservations-------------------------------------------------------------------------TotalIntegersNonintegers|#Negative---|#Zero263263-|#Positive11441144-|#---------------|#Total14071407-|#....Missing517+---------------------------01501924(50uniquevalues)20.summarizej3,detail通常您的配偶平均每週大約花多少時(shí)間做家務(wù)工作？PercentilesSmallest1%005%0010%00Obs140725%20SumofWgt.140775%2111221Recodej3112/max=112Tabulatej3?Ming-chiChen22社會(huì)統(tǒng)計(jì)------------+-----------------------------------?Ming-chiChen23社會(huì)統(tǒng)計(jì)來看看男女的差別A1.這題是性別，男是1，女是2。Data-dataeditor-找的A1這個(gè)變數(shù)-滑鼠右鍵Variable-properties-label改成性別Valuelabel-define/modify-define-labelname輸入gender-OK-value鍵入1-text鍵入男-OKvalue鍵入1-text鍵入男-OK-cancel-close-valuelabel選擇gender-OK關(guān)掉Dataeditor視窗?Ming-chiChen24社會(huì)統(tǒng)計(jì)男女的家務(wù)分擔(dān)是否有不同？Statistics-Summaries,tables,&tests-tables-One/Two-waytableofsummarystatistics自變數(shù)依變數(shù)?Ming-chiChen25社會(huì)統(tǒng)計(jì)差別很大嗎？|Summaryof|通常您平均每週大約花多少時(shí)間做家務(wù)工作|性別|MeanStd.Dev.Freq.------------+------------------------------------男|6.048553710.23684968女|18.33030616.287017881------------+------------------------------------Total|11.90048714.7918771849?Ming-chiChen26社會(huì)統(tǒng)計(jì)母體變異數(shù)未知但已知相等Statistics-Summaries,tables,&tests-Classicaltestsofhypotheses-Groupmeancomparisontests依變數(shù)自變數(shù)信賴水準(zhǔn)?Ming-chiChen27社會(huì)統(tǒng)計(jì).ttestj2,by(a1)level(99)Two-samplettestwithequalvariances------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ho:diff=0degreesoffreedom=1847Ha:diff<0Ha:diff!=0Ha:diff>028母體變異數(shù)未知但已知不相等以上的方法是假設(shè)母體變異數(shù)未知但已知相等。不管樣本大小，統(tǒng)計(jì)軟體一般用t檢定那如果母體變異數(shù)未知但已知不相等，怎麼辦？?Ming-chiChen29社會(huì)統(tǒng)計(jì)母體變異數(shù)未知但已知不相等Statistics-Summaries,tables,&tests-Classicaltestsofhypotheses-Groupmeancomparisontests變異數(shù)不相等自由度需要比較複雜，由Welch提出的運(yùn)算方式?Ming-chiChen30社會(huì)統(tǒng)計(jì)男女性負(fù)擔(dān)家務(wù)工作時(shí)數(shù)的差異，在母體變異數(shù)未知但已知不等的情況下.ttestj2,by(a1)unequalwelchlevel(99)Two-samplettestwithunequalvariances------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ha:diff<0Ha:diff!=0Ha:diff>0?Ming-chiChen31社會(huì)統(tǒng)計(jì)變異數(shù)相等與否的Levene檢定Statistics-Summaries,tables,&tests-Classicaltestsofhypotheses-Groupvariancecomparisontests依變數(shù)自變數(shù)?Ming-chiChen32社會(huì)統(tǒng)計(jì)變異數(shù)相等與否的Levene檢定.sdtestj2,by(a1)level(99)Varianceratiotest------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ho:ratio=1degreesoffreedom=967,880Ha:ratio<1Ha:ratio!=1Ha:ratio>1sd(男)/sd(女)不等於一，p值顯示可以拒斥變異數(shù)相等的虛無假設(shè)?Ming-chiChen33社會(huì)統(tǒng)計(jì)根據(jù)Levene檢定的結(jié)果，選擇變異數(shù)不相等的假設(shè)比較正確。也就是男性分擔(dān)家務(wù)的時(shí)數(shù)顯著地少於女性。?Ming-chiChen34社會(huì)統(tǒng)計(jì)已婚未婚者的家務(wù)工作負(fù)擔(dān)的比較A5為受訪者的婚姻狀況1為未婚，2為已婚，3為其他已婚者家務(wù)負(fù)擔(dān)比較大嗎？?Ming-chiChen35社會(huì)統(tǒng)計(jì)已婚未婚者的家務(wù)工作負(fù)擔(dān)的比較仿照男女的比較得到如下的錯(cuò)誤回報(bào).ttestj2,by(a5)level(99)morethan2groupsfound,only2allowedr(420);這是因?yàn)閍5這個(gè)變數(shù)有三個(gè)變數(shù)值：未婚、已婚和其他要用條件是來限制，僅比較未婚者和已婚者?Ming-chiChen36社會(huì)統(tǒng)計(jì)Statistics-Summaries,tables,&tests-Classicaltestsofhypotheses-Groupmeancomparisontests?Ming-chiChen37社會(huì)統(tǒng)計(jì)變異數(shù)相等.ttestj2ifa5!=3,by(a5)level(99)Two-samplettestwithequalvariances------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ho:diff=0degreesoffreedom=1835Ha:diff<0Ha:diff!=0Ha:diff>0?Ming-chiChen38社會(huì)統(tǒng)計(jì)變異數(shù)不相等.ttestj2ifa5!=3,by(a5)unequalwelchlevel(99)Two-samplettestwithunequalvariances------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ha:diff<0Ha:diff!=0Ha:diff>0?Ming-chiChen39社會(huì)統(tǒng)計(jì)Levene檢定.sdtestj2ifa5!=3,by(a5)level(99)Varianceratiotest------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ho:ratio=1degreesoffreedom=305,1530Ha:ratio<1Ha:ratio!=1Ha:ratio>1無法拒斥變異數(shù)相等的虛無假設(shè)?Ming-chiChen40社會(huì)統(tǒng)計(jì)兩層群體的比較已婚男女間，未婚男女間是否有差異？婚姻是否不利於女性（至少就花在家務(wù)勞動(dòng)上的時(shí)間而言）？?Ming-chiChen41社會(huì)統(tǒng)計(jì)變異數(shù)相等Statistics-Summaries,tables,&tests-Classicaltestsofhypotheses-Groupmeancomparisontests?Ming-chiChen42社會(huì)統(tǒng)計(jì).bya5,sort:ttestj2ifa5!=3,by(a1)level(99)-------------------------------------------------------------------------------------------->a5=未婚Two-samplettestwithequalvariances------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ho:diff=0degreesoffreedom=304Ha:diff<0Ha:diff!=0Ha:diff>0多重比較變異數(shù)相等43多重比較變異數(shù)相等->a5=已婚Two-samplettestwithequalvariances------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ho:diff=0degreesoffreedom=1529Ha:diff<0Ha:diff!=0Ha:diff>044多重比較變異數(shù)不相等.bya5,sort:ttestj2ifa5!=3,by(a1)unequalwelchlevel(99)-------------------------------------------------------------------------------------------->a5=未婚Two-samplettestwithunequalvariances------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ha:diff<0Ha:diff!=0Ha:diff>045多重比較變異數(shù)不相等->a5=已婚Two-samplettestwithunequalvariances------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ha:diff<0Ha:diff!=0Ha:diff>046多層次比較變異數(shù)相等檢定.bya5,sort:sdtestj2ifa5!=3,by(a1)level(99)-------------------------------------------------------------------------------------------->a5=未婚Varianceratiotest------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ho:ratio=1degreesoffreedom=176,128Ha:ratio<1Ha:ratio!=1Ha:ratio>147多層次比較變異數(shù)相等檢定->a5=已婚Varianceratiotest------------------------------------------------------------------------------Group|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------Ho:ratio=1degreesoffreedom=783,746Ha:ratio<1Ha:ratio!=1Ha:ratio>148BoxPlot箱型圖的比較?Ming-chiChen49社會(huì)統(tǒng)計(jì)單身男性和已婚男性是否有差別？單身女性和已婚女性是否有差別？?Ming-chiChen50社會(huì)統(tǒng)計(jì)配對(duì)樣本結(jié)婚對(duì)女性不利？前例的分析中，我們比較已婚者與未婚者從事家務(wù)時(shí)間的差異，由此來推論婚前婚後可能產(chǎn)生的變化。但「婚前組」與「婚後組」是由不同受訪者所構(gòu)成的獨(dú)立樣本。如果「是否結(jié)婚」與某些人格特質(zhì)有關(guān)，則我們不知道是因?yàn)椤富橐觥贡旧碓斐尚袨樯系母淖?，還是具有某種行為傾向的人比較容易選擇婚姻。即我們的分析可能隱藏「自我選擇」self-selection的問題?Ming-chiChen51社會(huì)統(tǒng)計(jì)配對(duì)樣本為了證明婚姻對(duì)從事家務(wù)時(shí)間的影響不是來自於自我選擇，更好的分析樣本為長(zhǎng)期追蹤資料（longitudinaldata），即能追蹤同一個(gè)受訪者，在婚前及婚後所產(chǎn)生行為上的變化。但這種樣本資料的蒐集十分費(fèi)時(shí)費(fèi)力。?Ming-chiChen52社會(huì)統(tǒng)計(jì)配對(duì)樣本夫妻之間從事家務(wù)的時(shí)間是否有顯著的差異？我們可以用兩種方式來分析：將已婚男性與已婚女性當(dāng)作兩獨(dú)立樣本，比較所有先生的平均值與太太的平均值是否有差異？?Ming-chiChen53社會(huì)統(tǒng)計(jì)配對(duì)樣本但夫妻從事家務(wù)的時(shí)間不是「獨(dú)立事件」，先生多分擔(dān)，太太自然可以少做。因此應(yīng)該比較同一家庭中，夫與妻從事家務(wù)的時(shí)間是否有差異，而不是比較所有的「夫」的平均值與所有「妻」的平均值。?Ming-chiChen54社會(huì)統(tǒng)計(jì)Statistics-Summaries,tables,&tests-Classicaltestsofhypotheses-Meancomparisontests,paireddata1st-2nd?Ming-chiChen55社會(huì)統(tǒng)計(jì)夫妻之間的家務(wù)分工Pairedttest------------------------------------------------------------------------------Variable|ObsMeanStd.Err.Std.Dev.[99%Conf.Interval]---------+-----------------------------------------------------------------------------+--------------------------------------------------------------------diff|1380-1.517391.657830424.43732-3.214199.1794161------------------------------------------------------------------------------Ho:mean(diff)=0degreesoffreedom=1379Ha:mean(diff)<0Ha:mean(diff)!=0Ha:mean(diff)>0配偶間相減，但是是妻減夫還是夫減妻？?jī)H知夫妻間有差異比配偶少，且達(dá)顯著水準(zhǔn)?Ming-chiChen56社會(huì)統(tǒng)計(jì)配對(duì)樣本如果要比較先生與太太從事家務(wù)時(shí)間的多寡，則應(yīng)該如何分析？男女分開分析?Ming-chiChen57社會(huì)統(tǒng)計(jì)產(chǎn)生新的變數(shù)並定義其