... and Hispanic origin15–19 years:Hispanic 1,150 55.0 1,078 52.0Non-Hispanic white 5,962 49.5 5,693 47.1Non-Hispanic black 1,392 59.5 1,351 58.315–17 years:Hispanic 688 50.0 673 48.8Non-Hispanic ... 2.8Race and Hispanic originHispanic 3,524 100.0 41.2 45.7 6.9 2.5 3.6Non-Hispanic white 19,445 100.0 35.1 46.5 11.1 3.9 3.4Non-Hispanic black 6,141 100.0 23.7 52.5 14.7 6.0 3.1Non-Hispanic other ... thepopulation.Thenumberofwomensherepresentsinthepopulationiscalledher‘‘samplingweight.’’Samplingweightsmayvaryconsiderablyfromthisaveragevaluedependingontherespondent’srace,theresponserateforsimilarwomen,andotherfactors.Aswithanysamplesurvey,theestimatesinthisreportaresubjecttosamplingvariability.SignificancetestsonNSFGdatashouldbedonetakingthesamplingdesignintoaccount.Nonsamplingerrorswereminimizedbystringentquality-controlproceduresthatincludedthoroughinterviewertraining,checkingtheconsistencyofanswersduringandaftertheinterview,imputingmissingdata,andadjustingthesamplingweightsfornonresponseandundercoveragetomatchnationaltotals.Estimatesofsamplingerrorsandotherstatisticalaspectsofthesurveyaredescribedinmoredetailinanotherseparatereport(13).Thisreportshowsfindingsbycharacteristicsofthewomaninterviewed,includingherage,maritalstatus,education,parity,householdincomedividedbythepovertylevel,andraceandHispanicorigin.IthasbeenshownthatblackandHispanicwomenhavemarkedlylowerlevelsofincome,education,andaccesstohealthcareandhealthinsurance,thanwhitewomen(14).Theseandotherfactors,ratherthanraceororiginperse,probablyaccountfordifferencesinthebehaviorsandoutcomesstudiedinthisreportamongwhite,black,andHispanicwomen(15).TableBshowsafactorthatshouldbeconsideredininterpretingtrendsinpregnancy-relatedbehaviorintheUnitedStates:thechangingagecompositionofthereproductive-agepopulation.In1982,therewere54.1millionwomenofreproductiveageintheUnitedStates;in1988,57.9million;andin1995,60.2million(16).Thelargebabyboomcohort,bornbetween1946and1964,was18–34yearsofagein1982,24–42yearsofagein1988,and31–49yearsofagein1995.Theselargebirthcohortswerepreceded(upto1945)andfollowed(1965–80)bysmallercohorts.Whiletheoverallnumberofwomen15–44yearsofageroseby6million,or11percentbetween1982and1995,thenumberofteenagewomendroppedbyabout6percent,thenumberofwomen20–24yearsofagedroppedby15percent,andthenumberofwomen25–29droppedby6percent(tableB).Incontrast,thenumberofwomen30–44yearsofageincreasedsharply—forexample,thenumberofwomen40–44yearsofageincreasedby59percentbetween1982and1995.Also,women30–44yearsofageaccountedfor54percentofwomen15–44yearsofagein1995comparedwith44percentin1982.Thesedifferencesinagecompositionmayberelevantwhenevertimetrendsamongwomen15–44yearsofagearebeingdiscussed.Publicusefilesbasedonthe1995NSFGareavailableoncomputertape.TheywillalsobeavailableonCompactDiscRead-OnlyMemory(CD-ROM).QuestionsaboutthecostandavailabilityofthecomputertapesshouldbedirectedtotheNationalTechnicalInformationService(NTIS),5285PortRoyalRoad,Springfield,VA22161,703–487-4650,or1–800-553-NTIS.QuestionsregardingtheCD-ROMfilesshouldbedirectedtoNCHS’DataDisseminationBranchat301–436-8500.ResultsTables1–17containmeasuresofpregnancyandbirthintheUnitedStates.ChildrenEverBornandTotalBirthsExpectedIn1995,women15–44yearsofageintheUnitedStateshadhadanaverageof1.2birthsperwoman(table1).Thiscompareswith1.2in1988and1.3in1982(17).In1995,women15–44yearsofageexpectedtofinishtheirchildbearingwithanaverageof2.2childrenperwoman(table1)comparedwith2.2in1988and2.4in1982(17).Theproportionwhoreportthattheyhaveneverbeenpregnantwasmarkedlyhigherforcollegegraduatesthanforthosewhodidnotcompletehighschool(table3).Thissamepatternbyeducationisalsoseenwhendataforlivebirthsareexamined(tables4–5):about49percentofwomen22–44yearsofagewhohadgraduatedfromcollegehadhadnolivebirthsasofthedateofinterviewcomparedwithjust8percentofwomen22–44yearsofagewithoutahighschooldiploma(table4).WithinraceandHispanicorigingroups,thepatternwasthesame:collegegraduateshadmarkedlyhigherpercentschildlessthanwomenwithlesseducation(table5).Table6showsacomparisonbetweenlivebirthsreportedintheNSFGandlivebirthsregisteredonbirthcertificatesintheyears1991–94.Ineachindividualcalendaryearandforthesumoftheyears1991–94,theNSFGestimateofthenumberofbirthsisveryclosetothebirthcertificatetotalanddiffersfromitbylessthantheNSFG’ssamplingerror.TheNSFGestimateisalsoverycloseforwhitewomen.TheNSFGestimateforblackwomenisslightlylower,andtheestimateforotherracessomewhathigherthanthebirthcertificatedata.Adiscussionofthisdifferenceisgiveninthedefinitionof‘‘RaceandHispanicorigin’’inthe‘‘DefinitionsofTerms.’’Overall,andbycharacteristicsotherthanrace,however,table6showsthatTableB.Numberofwomen,byage:UnitedStates,1982,1988,and1995Age¬...