University of California College of Agriculture Agricultural Experiment Station Berkeley, California Estimating Production of Natural Raisins in California by Use of List Sampling by George M. Kuznets and George Harvey August 19$2 Contribution from the Giannini Foundation of Agricultural Economics Mimeographed Report No. 13& LIBRARY university of california Davis • Estimating Production of Natural Raisins in California by Use of List Sampling* by George M. Kuznetsl/and George Harvey^/ During 19U9, 1950, and 1951, the California Crop and Livestock Reporting Service, a cooperative office of the California State Department of Agriculture and the Bureau of Agricultural Economics, U. S. Department of Agriculture, pre- pared and published special estimates of production of raisins in California. These estimates were based on information provided by specially designed samples drawn from a listing of raisin-type grape units located in the major producing counties of the state. Table 1 provides in summary form a schedule of the ex- perimental surveys during the three seasons. It will be noted that four surveys were carried out each season, namely, a preharvest survey, two surveys during the harvesting period (hereafter referred to as within-season surveys) one week apart, and a final survey.^/ In each of the preharvest surveys, the operators of sample units were interviewed by specially hired and trained enumerators, who had a background of experience in the local grape industry. The final sur- veys were carried out by mail with personal interview follow-up of all nonre- spondents. In 19h9 and 1950, the within-season surveys were by mail only and in 1951 by interview of one-half of the total sample. The schedules employed in the 1951 surveys are shown in the Appendix. In addition to raisin production by main varieties, information was obtained from operators on bearing acreage in sample units and acreage harvested for production of raisins. In 195lj in- formation was also obtained prior to harvest on acreage to be harvested for * The activities discussed herein were conducted under the general direction of G. A. Scott, State Statistician in charge of the Federal-State Crop and Livestock Reporting Service in California. 1/ Professor of Agricultural Economics, Economist in the Experiment Station and on the Giannini Foundation, University of California. 2/ Agricultural Statistician, Bureau of Agricultural Economics, Sacramento. 3/ The following persons supervised the field work for these surveys: 19U9, Irwin Rust, Extension Service, University of Hawaii; 1950, James H. Swedberg, Bureau of Agricultural Economics, Sacramento,- 1951, Robert D. Parr, Bureau of Agricultural Economics, Sacramento. •xn-xolxXeO at anxeisH Xstx/cteH lo xoifruhoi? *nltsa±#e& *SnlLqmZ tztd lo qeU yyf vd rem ion §fi.c.i?oqefl jfoectaavxJ bite qo?0 fllrrrolxXcO »ri.t 5<>X t ^i!RX gnxoiKI ewJXuox*rgfi lo rfnaratosqaCI octecJS cxniolxXsQ eric)- lo eoXTto srvicteiaqooo e t aoxvr93 -eiq t 9iud-.Coo.t:igA lo ^rrdKvHsqsd ,V t 3oxjrronoo3 Is icily o±-r§A lo usaiuS 90 bns 0Btn1ollX.nO nx onxaxei lo rro±froi/bc"Tq lo a^ercMao Xsioaqa barfexXdi/q fans baiBq SQlqmse fcengxeafe YXXsloaqe yd bebxvo-xq noxcbwntolni no baaed ©taw aotemxctea aaari? 3nxoi/bo*rq -loten? ga\t nl. befool a^xni/ eqsrrg sqxJ-riiats-f lo nnxcfaxX s moil nwstb -xs aritf lo aXtrborioe b /rrtol yis/TEB/a rtx zobtvoiq X aXdsT .a^e-fre axW lo aaitfni/oo mot ferlt beion 9d IXiw dT .arioscae asuii odd- 'sniiyt? p.yevrt/s Xfitaemx'saq grrinyb eyoVTUB owi ,ysv~us d-awiExio-rq g t Y.Xsman t nos<39B Aob9 duo botTtnc> araw jj9©w eno i&t&ri'jz noasaa-nxrfi^ bb od bsnel3i lacHcanari) bsxiaq anxJaavmri otit sioJsisqo erfd- ..eygviua iasvtsriaiq erf* lo doss nl XL^vtx/s Xsrtll s bns t cHieqs ,8i0oGtamun9 banxEiJ ban botxri yXXexooqa yd b9w9xvxe.tnx si9w etf.cnif aXqntes lo -* 3 Xsnxl arfT .\-r*60bnx 9qBt§ Xsool sd& nt aonexmqxo lo bntioigdoscf s bed odw -omon XXs lo qu-woXXol wexv-tatfni Isaoaisq dJiv Xxam ^d Jx/o b^xnao siaw gvsv ni bns xlno XxUci yjo* 9-row ay^nrn/a no as aa-nxrictxw otW t 0 XBiutXyoxigA lo itesnjH t ni5loiuRl'tBd8 XBW^'Xyor'rgA. \S t W?X :ay,9viifB eaori^ xol >('io>f ^bXoJrl 9ri.t basxviaqwc anoe-x^q gnxvroIXol 9riT \£ tS'^sdbava ,H aemsT, t O^X iliB^H lo McHatavlnU t 9oivoaa nol8nocb:3 tl$k 2. TABLE 1 Schedule of Raisin Production Surveys, California 19U9, 1950, and 1951 Preharvest survey First within- season survey Second within- season survey Final survey Field work Reference^ Closing date Report date Field work Referenced/ Closing date Report date Field work Reference^/ Closing date Report date August 5-31 Not issued August 9-23 MM August 23 Not issued August 21-25 August 25 August 29 19JU9 Season (four- county sample) September is/ September 5 September 12 September 13 September September 12 September 19 September 20 1950 Season (seven- county sample) August 283^ August 31 September k September 6 September September 7 September 11 September 12 1951 Season (seven-county sample) September b-5 September k September 5 September 7 September 11-12 September 11 September 12 September lU September 23^ October 11 November 3 September 2 October 13 October 19 October 2 October 6 October 10 -62/ 2/ Mailing date. ]>/ Date to which observations pertained. SJ Interview follcw-upj mailing date one week earlier t AT 1 v orTV ix>«ir*i 1 1 j • ■ " '* j ,. „,. \?".ait: i • • i Si " r 3'Siii9& r v r irroqafl | ■ CO O 4<3fmriA \cf * „ •£X nsdoteO XI isdmatqea 4I nsdneJasE <) 'rectos d-qaS ( eXqmsn x$mi po -novas) SX-IX ■tedraetfqeC &I ■x*fsw*qeE j Vara** d xodoctoG 1 SI i8-ihrs*qaa 5S iex/goA 1 gdab g/tfcaatO sdeb itoqgJI 1 01 istfoioO L 1 4X I9tfnr9*q93 . V -rstfwedqeS • s-Aab gafXisM \s * bsnisdisi"' s>nn ^J^^^P'ia dirriv r *vt i+a(T-\cT .'I9.CX-IR& ?I39W 9TI0" 9«tftb ^flxXXfiP 1 ! ^ jqU-WOXXol WSXVSd&tl \2. 3. raisins and expected total production of raisin grapes and, in the final survey, on tons of raisin- type grapes harvested for all uses. Funds for this work were made available jointly by the State Department of Agriculture and the Production and Marketing Administration under the Research and Marketing Act of 19U6. Mem- bers of the staff of the Giannini Foundation of the University of California participated in the planning of these studies, the design of samples, and the analysis of data collected in the surveys. The purpose of this report is to provide a description and appraisal of the procedures employed, with emphasis on the 1951 surveys. Particularly for de- scriptive material, this report has drawn freely on a memorandum regarding the several RMA projects on methods of estimating California raisin and grape crops prepared for administrative use in May, 1951. Background Information Of the total production of raisins in California, including both natural and bleached, over 90 per cent is made from Thompson Seedless variety grapes and most of the remainder from Muscat, Black Corinth (Zante currant), and Sultana varieties. In 1951, the bearing acreage of raisin grape varieties in California was 230,379 acres, of which the Thompson Seedless variety accounted for 187,1*71* or 8l.it per cent of the total. The other two major varieties, Muscat and Sultana, accounted only for 36,360 (l5.8 per cent) and h,0k3 (1.8 per cent) bearing acres, respectively. The remainder of the bearing acreage in 1951, 2,502 acres, included 2,UlU acres of Zante currant. The seven counties of Fresno, Kern, Kings, Madera, Merced, Stanislaus, and Tulare, in which all but a few tons of the raisin crop is produced, account for 9l*.l per cent of the total, and Fresno County alone for 57.0 per cent. Slightly under 60 per cent of total fresh weight of raisin grapes sold during the five-year period, I9I46-I95Q, was made into raisins. The other major outlet for raisin grapes is wine, brandy, or juice. This outlet accounted for almost 30 per cent of all sales during this period. Of greater relevance to the problem of efficient estimation of raisin production is the marked year- to-year variation in utilization of raisin grapes. During the years 19U6, 19U7, and 19U8, total sales of raisin grapes varied only slightly. However, in 19U6 raisins accounted for hi per cent of total use; in ±9h7 this percentage jumped to 72; and in 19U8 it dropped to 55. The possibility of abrupt shifts in uti- lization apart from annual variation in the yield of raisin grapes complicates the problem of designing an efficient sample for the estimation of raisin pro- duction. ttwuip. Izail arid nx t bcw sonata nXaXai lo noxdouboaq Xsdod b9doeqx© bras snxaxci 9-r?w yhon, sxitd tol abnul .eeatr XXb tot bsrfewisri saqBig aqxd-Aiais'x lo axtod no noxdo.oboTrS add bns ©ludJxroxTsA lo dngmdicqod adsdo arid vrf TjXdnXop oXdsXisvB ebs/u -maK ..c>4q aaxdnuoo navaa srfT .Josiioo o-triES lo aeioc 4x4 t S bgbyloni: taeios SC5 t S \l>Sl ftt XXs rioxriw nX t 9is]l;T bns ,?«BXaj:n[s*8 tbeo-te?! t si8bBM ^B§n±X t nio.l ^onegi^ lo 9rW lo &aeo 'iaq X.jl^ 'xol .tnuonop, t baoxrt>cviq bx aura nieisi ertt lo enoJ wol s Jxrd .cbioo i9q 0,V?. tol onols yinuoO onaortl bne t Xscfoi bXos a9qsi§ niaist lo .trisXow rise-jl Is&of lo 4nso teq Od iobn« xlSd%lIZ lot am isri^o oriT .snxBisi ofci .ebBra enw t 059X-dd?X t bori9q issi-ovn ed? ^nixib «>1 boinuooos tel-fuc BidT *&o«;t io t^biBid t onxw a.t esqats nlaXfii iol d-sXduo orf 9otcbv9X9'i -J9i69T9 10 .boxTOq t atrii ssnxtub eaXsB XXs lo itneo i©q OC d-aoklB •iB?iY bsjiism arid ex nol^ouboiq nXaisT lo ncxdsmx^Bs JnoXolllo lo maXcfotq 3dJ t \»JW tdii'eX ats^ atfd pnx*oKI .33qBi§ nisxF/i 1c noidssriXtcto nX nol*B.r£6v ibs^-o* o4oi nx t -r&v9MoH .yXd-riaiXa ylno beiiev ssqsis nxexsi lo sgXbb Xsdo* ,3j1?X b.T6 baqnurt agsfflao'roq eXrid Vjj^X nX joatr Xstd 9oxytb- neboow to taqsq no beo'slq beta mt\ 9rfd motx b9vom9t eis eadoosfl aeqiitg xo dd§iaw 'rfeetx ssstevfe orfT ' .awcrx saiv nasvded fcial ata rfordw (ifx *S) •od one add at dlsX 3tB asqeig arfT .abrweq SS duode di eystd §oiyib ori.t no fti'?I to grrlXXot ssdxA .ddg/9K ox drtoo isq £f duods gnraol ,aiteow wax s tox ytb dtorfe s d29l brm* aaxod'bXeit to eaxoddsaws nx baoslq ere soiaxst t eyatJ soisJosda i»Xii res tol- botoda ad ysm aoiatst t axdd gncWoXXo"! ,atua od eaxod add ru araxd tox toXboGri s od yXeterbapiffii b9tevxl9h to mrsx arid no eacd xo boitaq 9dxni'l9fr qoto Xswdso arid xo dnso tsq Oil di/6d<-> ,0^1 nl .sojbbsq bos <§oidtoa <^xnc9Xo % I ytBiirtfiT, yd 4-090 teq IT t X tadnrooatl yd taeo toq Qb e I tedhsavoM yd betevxXob esw qoto Xati/dso X&X erfd xo aaitavxlaa jC&I Ls9td od ocli neblcO etc BiitolxXsO fix abun ylxionnnoo daora asqyd eriT ►atodatbyrieb to oetr odd yd diJQX moti ^toridegoi enaxtt bos t anraist bsqqJtS-sbdS fans t bario39XH ti/lXo8 ^ftsdoaeXa to 39qy.t badosold XXA ».obsm ehiBxst XXe xo dooa teq 6 duodd od befouoirts s 0?.QX od oadio ,bXei:'r edd n± ncrii teridet etotetbyrfsb ds to abtsy \tb nl ob&M ate anxa-trt ate a.t-3Wota t riB3tx bios ei qoto eqa-ig n28iBt £ II .tsyxrd sritf ridiw dostdnoo tohoif -efc ed XXx;; qoto txedd teridariw worrti doo ysm to'^sm fiofi dd§xew nod riaot'l oo- bif.q -eb yXXsxoxxxdta) bsdofioXS O3bXo0 lo oo tdoi/botq arid tot nocsee otfy ,b*b--tbY.i. . ♦tadodoO odni Haw nut ysm \absm soyd borfosaXd Xjpqionxtq odd ^enxedst (bataibyn Obsm eta eeqets 9qy*-oJtsxGt 'lo roidoubotq Xedod 10 aedBioiuao boe ad^oeto'-T -xm&d t X yluL ddi'w grti:ooi39d yjjidnom ooJ^teS soidtoqeR qotO sdndB-XBtsb^ jrit yd xo sdoinxdse oA •.tsdm&ooCI tit ytcrac/S XfiiuonA qotO dtrb bOB dxi/t? erid ridxw §ox^bh ytenx.-nxXotq 6 fws- ooxrtiqo ytdai/btx 00 baaed enieiat lo ooidoiibotq XeooBBae odd Xst/noA todmaosa a ' dna^^tBqsXI 9rld of bsdaxlddq dstxx ax Moado ooiixaoqeib qoio atsay add to* 1 ! - .ariifL ^of woXXol odi 'noiaxvot od ioafcdiia a± bne ytsmmuB qotO teq C»£I od d.0 BRWft bsgoBt earf 'odafcUdae ytsoxmils'ra 3xd«/ ni tott'j arid t ?d^X od -Xda9 baa-.tvat ?xfX. .dnsa teq 5 narld tsdp.ots esw tl ©taay ; ^Jt odd lo 8 oi bo& dnoo adaXqsnoo yXXsxdnBdsdwa batebrahoo ets rioxriw fedsb jloario eXdsXxsvB no- bessd 3i odsm- n/sxst xo adBmldaa ds aa>!sm (»a.A.^JI) sintolxXsO- "io ooxdc-tooaaA diirfi bsctO odT b de daavtad- snxwsXXot tsoy arid lo aoitqa odd oi barisiXdoq z£ doxdw oexdoubotq oaXe ai ajsmxiaa aidT .sisXboad yd bavieoat nasd esri qoto' 3dd lo i ^om noriw eiitld -' »t9ieX tsay ano ooisxvst od doafcdua 5. Sample Design For all surveys during the three seasons, 19h9, 1950, and 1951 > estimates were based on information provided by a sample of units drawn in a random-strati- fied manner from a listing of all raisin grape units. Such a listing was avail- able from records accumulated in the Crop Reporting Service periodic enumerative surveys of all fruit acreage by varieties in the state. A vineyard unit is con- sidered in these records as a block of grapes consisting of one-tenth or more acres of a single variety. Usually, a unit will contain more than one variety of grapes. An individual may operate several units, and a single vineyard or farm may consist of more than one unit, although the single unit vineyard is most common. Stratification was by county and by bearing acreage of all raisin grape varieties in unit. Sampling units were allotted by strata for the 19^9 prehar- vest survey on the basis of total number of units in stratum and the unit standard deviation of stratum bearing acreage. For later surveys, allocation was made in proportion to size of stratum and the sample standard deviation of stratum unit production of raisins. The project was on a trial basis in X9k9 with sampling re- stricted to the four major producing counties of Fresno, Kings, Madera, and Tulare. Estimates of raisins made in other producing counties were obtained from members of the industry. This was not satisfactory from a research standpoint and in 1950 and 1951, units located in Kern, Merced, and Stanislaus counties were added to the sample, thus providing in the latter two seasons a complete coverage of the producing areas of sun-dried raisins made from the varieties samples — Thompson Seedless, Muscat, and Sultana. The estimates and their (estimated) standard errors shown in Table h for 1950 and 195lj based on the 1951 survey data, were computed with reference to four broad acreage strata, as shown in the tabulation below, disregarding county breakdowns. Reference Strata, 1951 Surveys Number of Bearing acres units in Number of Strata in units surveyed counties units in sample N h nh I Below k0 10,067 578 II 1*0-99.9 780 liiO III 100-399.9 161 60 IV U00 and over 8 J3 Total 11,016 786 n^ 1 iC9Q sJLom.so J.9Xd;t srtt gniUJb ?,yevii;e XIs -i^B'xcftj-.Tohflus'? s ni irorsib atinu lo s/qmfia s yd bebivciq noMennolni no b -Iibvb Bsw ^nxJail b rfoaa .eJinx' sqsig nxalBi XIb lo gnxtfail £ /noil 10 sv Wfii^njn^ oiboxigq sonneS snxtfionpH qoil) 9itt ni bsJrlumi/oofi abnoooi -noo sx Jim/ bi/?Y9fliv A . t» tsia 9rii nx asiiaiiBv yd o^s^tob Jiircl XIb 1 anom 10 riin&cNenb lo sniJaxsnoo 39CjBtsj lo toold s «b abico?* -939ritf 0 yteiisv 9no oBiii emu nis.jnoo Iliw .rinx; s »yIX?u;8U .yJoiiBv ©I^cjs s 10 bisxjerrxv elsnxs e bn£ t ztimt Isi9V9B eiBi^qo yam Xswbivibni tiA » teom at biB\?nxv Jint/ olgnje srict rfguorfi.-s ,.txf!i/ 9no nsrii ?«iom In dte-xanoo oqeig ntilBi XIb 'lo 9g£97on gnlissd yd bns yj-ncoo. yd asw nox^BoilxiBiiS .noimnos -anrisiq ©rid - *tol eisita yd Be^oils s-isw atinu 3nxIq.tiB2 .tisw ni aaidsiisv rahnsda titty srid bne flH/iaiia ni silnt; lo todmi/n Isiotf lo eiaad 9/W no ysnna taev ni ©bBin as* noxixooXIs ^sysvxya 19+sI la's ;9gs9ioB gniiced muiBUe lo neiixivec 1 elds yeviua s-iebia 3 91 OB 319 lO in rrnfi'i BtniJea .txnu mi/jsiJ-3 lo noidfiivab b-iBbnsda 9iq;pB3 eiii bnx; ro.r.jsiv'S lo esis od n •x .^nilqmsB dthr S>UVI ni acasd L&iii & no sew ioetoiq 9riT .sn/axBi lo n sI;. ■ Oti woX99 TOY. ' l.^li.S r & +rtT X>J 'J X • • oXG t XX 6. The effective stratification comprised many more strata. Not only were units drawn from size strata of every county included in the survey, but the size strata were less broad. Within reference Strata I and II, sample units were drawn in a systematic random manner from size classes of 10 acres and, within Stratum III, largely from 2$-acre classes (see Table 2). Standard errors computed on the basis of reference strata would, therefore, tend to overestimate somewhat the sampling variability of estimates. It should be noted that Stratum IV, consisting of only 8 units, was enumerated. Stratum III was enumerated in the 1950 surveys. In 1951 all of the units of this stratum were contacted for the final survey. It will be noted from the various entries in Table h, which show for several items, estimates and their estimated standard errors separately for sampling and enumeration of Stratum III, that the enumeration of this stratum (which required the addition of some 100 units to the sample) had neither changed significantly the magnitudes of the estimates for 1951, nor materially decreased the (estimated) standard errors. The data shown in Table 7 for the wi thin-season surveys are based on the finer size stratification as given in Table 2. Several types of estimates of population totals are shown in Table U: 1. Linear (stratified) . This estimate is of the form R n, (1) T. - 2 N, (ETJ, ,/n, ) L , h . hr h h x in which X hi is the value of some item for the ith unit in the sample drawn from stratum h, N h is the total number of units, n h is the number of units in the sample drawn from stratum h, and R, the number of strata. The variance of T^ is easily shown to be ' w A - 1 - V/SA - u L h in which is the population variance in stratum h. An unbiased estimate of o-j is given by L h with s, defined as h m -J = 2 h (x hi - r.) Z /\ - i) i The estimated standard errors of the linear estimates shown in Table h were cal- culated using expression (3). eSinii stow yino ioW .b*s-i.13 ^toffl vnam basrsqmoo noiieoXlXJaite avxioalle 9dT scfaicN asia ad* dwd ^yftvtua pri.* nx bsbulocfi •fttaooo viava 10. stfaitfe asxs csoil nwxb s nr nwBifc siaw adxntr alqmaa t II bna I b&bi& aonaialsi ninJxW .bsoid 33 9X trt0& till aiuJsi^a nidcfiw ,bns aaios 01 lo easaalo esia moil lannara arofcwn 0 .c;J aerate \a aiaed adj. M bsiuqaoo- siotip M«bc8*8 ,(S aldsT aaa) eac-aaXo eioa-5S moil yXaaial snxlqsiee MM Jsdwamoa e.t£ntxdr.9X9vo bnai ,9iol9isd.t r>;?a ©d ysc «b lo abyjinaBre ©rut nx no£tobfe9*' {slirtsiadifa s ^inoiJBOoils "mxfflu-fqo'' nem^sK orii anieu \d \Iebiyr \rrsv < jo .seansiisv | \ ht Mm - °a (5) • \o • r; : . oa Is x ism.+srt>lBa \6 bsbivo-rq s>on©bxv© Isinsm i"ieqx3 ,( l - u)\,Jl V .© * doxiiw 1U nox«jsIi/qoq sri* 3nioslq9i \d be-mnez ©d xsoi <5iIxfE6t mumiiqo issn isrfi esixotbni erti ai b©Jlo eilx/eea ©dT . rf a aotfsmiiae ©Icpuia ^d (5) nx l .x> ano ri&iveb bisbfie-ta 9on6i5>'isi aetrfj Jaixl ©rfi ioi assia edqmse LzuSoa ©ri* iscii worfe woI©ci 'n.oijBLu^sS boaxGico aTedraxm ©Iqnuaa mxnniiqo -xegn ©d* fno*ri YLbeifrsw i©llib ion ob sis-iie cii y^niia Isnll ©ricf *c asasKti insj-foqrnj: ©rii lo ©ffroe ioi anoxisooll A muraxtfqO b^emxj's.I (stall ynvw3 Ifini-? £&1) | XSSI ■ siqmsa •anxiRSU » 60TB13 nxa re/; oraxsF. ©oim©l©5I ar*.>±?< . ,.*•:!£, ..^.j — : ,. / ryx^fco-fq^ 1 sU^ 1 bid ; * M d8 • 89 OSI OS ■ db m ■ bTT 8TV ©Jemite© oliBi ©Hi r io amtol owi t noxi£oi:liJs'ii3 dJxW . ( b^xlxia^i ^a) :.'I\tgH ,S ©Xdxaaoq oi& AW < d > • * r * rf 1 bne ■ .boriJsm ©vi jsinssViq&i ©rii xo aioe-qas irtsiexlifo owd-erii nO : .1, t xieary,9V! .4^1 t5Sd-8^:t£ .o?a -ifi^a .boriiera' sviaajnoseiq©-! ©di lo yrioerfi ©ri^ oi.Bn^i-^ydxiifioO .1 t ©«Kh;n'jiua 9. in which Y^ is the current year sample mean in stratum h for some item, X^ the corresponding mean for the preceding year, £ h is the population total for the item in stratum h f or^preceding year, and £ the over-all population total for preceding year ( £ = 2 g )» Estimate T^ 1 ) is to be preferred if ratios differ h h r substantially from stratum to stratum; however, its use presupposes knowledge of strata population totals, Since the latter is not known (except for bearing acreage, 1950), the second form of the ratio estimate was employed in all cases for which the ratio estimate was computed, X representing the 1950 and Y the 1951 data. As is well known, the ratio estimate is biased unless the regression (of Y on X) is linear and through the origin.^ The bias, however, is presumably neg- ligible for large sample sizes. The mean square error of that is, E(T^-t]) 2 , is given approximately by in which r\ is the current year population total (which is being estimated) and ^"xyh is t ^ ie P°P u l a tion correlation coefficient between X and Y in stratum h. The estimated standard errors of the ratio estimates shown in Table k were obtained by substututing in (8) the sample equivalents of the population parameters t) , £, ct 2 , a 2 , and , . xtf yh' ' xyh 3. Difference ( stratified ) ♦ This estimate, also linear, is of the form (9) + ~\h h and is, obviously, unbiased. Its variance is (10) ^ = 2 II 2 (N h - % )(a 2 h ♦ a 2 yh - 2a xh a yh p^)/^\ - 1% 2 An unbiased estimate of a 7. is given by (U) V " t V N h - \><4 * 4 - 25 xyh )/n h d h 1/ Cochran, W. G. Recent developments in sampling theory in the United States. International Statistical Institute. Proceedings of the International Statistical Conferences. September 6-18, 19^7. Vol. Ill, Part A, pp. kO-66. edj- n -X t nrsii emoa lol rf fficreixa nj nean MflMM tb&\ inantfo ?dx aX J rising' ac e>ti& "iol Xsioi nolisl'jqoq $ti3 8i 2' %nib?n?aq ed.+ ioi neeffl yjibnoqesnoo io2 Xexoi noiJ-^Iuqoq. XXb-ipvo 9df 3 bns ^b^y snxbeosiq _iol A nw delete nx msix •tex'Ub sox.tfii ac bensls-xq ed a{ ^ifiauiaS »{ 1 X » }' ) "see^ gnibswiq xo sgb&Xwoa^ eesaqqusenq eeu axx ,7sv¥W0fi t inuiBTt8 o3 nrt/Xs'iJa itio'i'?;- ^Xisi^niJvtadtia linsd tox .tq^oxa) nwojxtf ion ex isiisi erii eoniu . .5 ^aXs-toi nciisXi/qcq sie.iia aeas') XXs nx bsY'^Xqns a«w oisnijae oXXst: erftf lo miol bnonea ettt ,(0dJ ,rworni 1I9W zl aA -■Sen Y-C^sfflu^^q a* t iev9v/ori ,?»nid s»riT -.ni^xto sricr rtyjotrii bns rjswuX si (X nc (jt- T)3 t ei Jarii t 'T V> ion-is ot&vpz neera eri'I .resxa *>XqniAja syifiX iox d£dx$xi >jd TfX'e'dsmxxo'tqqa nsvrg ax bns (bsirjnxia© goled ax rioxdw) Xsjo.j nor.jBD.rqoq ise^ inn-rajo edi ai jt rioiriw nx eril ,d mniciJa nx Y bns X n»ew^od ins»xoix1?o3 norisXenoo noxioXx-qoq srii ax j„J°S , n- ate.i9rairxsq noxj-sXiiqoq E»di lo a.+naievii-fpa eXqiriBS vdj (ft) nX gaiii'di/isdija ^d :>da£x r -(m- -.x bnfi Yd n«»vxg ei lo 9ifiBiLf35» beafiidnt; nA (XI) ,35^a?8 bsv-JxnU srii nx xtosdj ^nil'jmsa ni a^nanqoldveb inpo^Ji .0 .H ,fi9idao*3 \X XBolisxifii2 XfirraiiBnisinl frix la snn.tbasooil .siydx-tanl leox.tsxisia lenox-tsm^ 10. R in which s , - z(X - X, )Y /(n - 1). The estimated standard errors of differ- xyh ? v hi h' hi h ence estimates shown in Table h were computed using (11). Estimates of Bearing Acreage Estimates of bearing acreage of raisin- type grapes (Thompson Seedless, Mus- cat, and Sultana) for 1950 and 1951 are shown in Table h> Orchard records place the 1950 bearing acreage of these varieties in the seven surveyed counties at 217,7^8 acres as compared with the sample expansion of 215, Ul7. The ratio of 1950 bearing acreage of Thompson Seedless, Muscat, and Sultana varieties of grapes in the state to that in the seven sample counties is 1.0558. Expanding the sample estimate by this factor and adding the 3,990 acres of Zante currant and other raisin varieties yields 231,1*27 acres as an estimate of the 1950 bearing acreage of all raisin-type grapes in the state. This indication compares favorably with the official estimate which is now placed at 233,888 bearing acres for 1950. It will be noted that the computed standard errors of the seven-county estimates of bearing acreage for 1950 and 1951 are a little over 2 per cent. However, the standard errors for Muscat and Sultana varieties are relatively much larger and, particularly for the latter, are so large as to make it inadvisable to prepare separate estimates for this variety without increasing substantially the size of the sample or changing radically the sample design. Estimates of Acreage Harvested for Raisins The relation between bearing acreage per unit and acreage harvested for raisins (natural dried and bleached) for 1951 is shown in Table 3« These data are intended to provide an indication of this relation for the population of raisin- type grape units which was sampled. Hence, the entries (percentage frequencies) in this two-way table have been adjusted for differences in sampling rates of ref- erence Strata I and II (Stratum III was enumerated for the final survey which pro- vided these data). Some interesting facts emerge from a consideration of this table. Almost h0 per cent of all raisin-type grape units did not produce raisins in 1951. Nonproducing units ranged in size from smallest to largest. Thus, U9 per cent of units containing 100 or more bearing acres and U6 per cent of units containing fewer than 10 bearing acres did not harvest for raisins in 1951. Of the units producing raisins, about 85 per cent produced to full capacity, that is, harvested all of the bearing acreage in their units for raisins. Linear estimates of acreage harvested for naturally dried raisins in 1950 and 1951 are shown in Table U. These estimates are subject to a relatively larger sampling error (about h per cent for 1951) than those of bearing acreage. Acreage harvested for natural raisins in 1951 was indicated to be about lh per cent larger -rollxb lo gsorxe biebneJe bsisraxJas 9riT . (I - _it)V ,Y( .5 - , X)s « . s rfoxrirr ni n id d i.i !J rfipc .(IX.) sfiian be^uqpraoD siew il eicfeT ni nwona eeismijae 9on9 egas-ioA 3niiae8 1© se^emld-ea •ax/M ,339lb96>3 noeqfloriT) aweiy sq^-niEisi lo sgeeios aniiaad in 39tannxd-e3' M&Iq ab-iooei h-isdo-sO .it 9ldnT ni nwptfa sis XSSI bns C^-i. ?ol (snscMu'3 bns ,crso i& aaxJm-oo b9^£wxua nevea erii ni aaxieiisv 939rfd jo ggse-toa ^ni-ised 05?I s»ri^ 0i£LU oi 2G 93SSI oa 9is t ieiin£ arid: lol ylis LuoxdiBq to esia ad* Y^.lBxin«d/ea'x/3 snxaaa-ionx i^od^xw Yiaxiav axricf *xol egismi.taa e^-raq^ .ngi39b eXqmsa sri* xil&'>ibBi gnignsrio io aXqmaa 9rii 3nx3.cfiH lol bPisavisH 9 , 989 , xoA lo sgtfami^aS fol baH&rt&d 9^fi9ioB boa &xwj agaeioa gniiasd n99«»ied nft&teie* edT •ta fitfao 93f dT .C eldaT rtx nworis =?1 X?9X r xol (benoeeld bna baiib I&taian) znittisi -ntexBT lo xtox^BXcjqoq ari* tol noi-tals-i sidi lo nox^eolbni ne abXvaiq o* b9hr..ectax (asionsufp9il e%stci9oi$q) aehtine »rii .sroneH .boXqras ; asw doxriw sjx/h; aqaig &qv,3 ■l«-i la a?ici gxixXqmss ni S93n*i9'llib lo'i beiautba n99d evsri eldsJ ybw-ow* sldi nx ■oiq doidw i{9irtif3 XsnXl 9rii. sol b&iBtemno saw III nu.^6fja) 71 brts I BtriBd ►nfi 0^ C I ni sniaxBt beiib ^IlBiuisn lol bsisoviBd ygcoioa lo aoJemid-?.? isaniJ •sa^isl ^Ievxd-Bl9M 6 oi d-oe^daa 9-is aeiBmitee eaodT. ^Xo'aT nx nwoda 97P 1^1 3B9TOA .9^39-106 anrjB^d lo OBorfi nadi (X^X sol d-rrao t9q J cfoodB) toits sailqitrae :9giBi inao ieq iil iuods sd ot bsisoxbni bbw i3^X nx eniaiB-i i&m&eo tot bod-spvipd TABLE 3 Bearing Acreage and Acreage Harvested for Raisins, Three Main Raisin Varieties, Per Cent cf Units in Specified Acreage Classes — Seven Counties Surveyed, 1951 Acres harvested for raisins, 1951 0.1- 9.9 10- 19.9 20- 29.9 30- 39.9 40- 49.9 50- 59.9 60- 69.9 70- 79.9 80- 89.9 90- 99.9 100-124.9 125-149.9 150-174.9 175-199.0 200-299.0 300-399.9 400-499.9 Total 16.60 0.1- 9.9 14.02 10- 20- 19.9 29.9 13.86 2.20 21.92 30.62 37.98 4.49 0.83 1.87 5.91 30- 39.9 2.81 0.15 0.69 0.84 40- 49.9 50- 59.9 4.43 1.24 0.16 0.06 0.26 0.38 0.78 13.10 8.92 0.74 0.06 0.32 0.06 0.22 0.06 2.88 60- 69.9 70- 79.9 0.77 0.51 0.06 0.10 0.01 0.01 2.23 0.21 0.01 0.06 0.11 0.11 0.19 0,88 Be aring acreage, 1951 80- 89.9 0.26 0.06 0.06 0.01 0.06 0.44 0.94 0.07 90- 99.9 100- 125- 150- 124.9 0.18 0.01 0.02 0.54 0.21 149.9 0.23 0.02 0.01 0.01 0.02 0.01 0.02 0.08 0.08 0.02 0.20 174.9 0.70 0.20 0.02 0.02 0.01 0.01 0.01 0.03 175- 199.9 0.10 0.01 0.01 0.02 0.01 0.07 0.30 0.22 0.06 0.01 0.01 200- 299.9 0.02 300- 399.9 0.16 0.01 0.10 0.01 0.02 0.01 0.02 0.23 0.02 400- 499.9 0.04 0.02 0.01 500 and over 0.02 0.02 0.07 0.01 0.05 Total 0.01 0.03 39.17 20.08 24.99 7.24 5.18 l.OS 0.82 0.3O 0.52 0.16 0.04 0.24 0.06 0.08 0.05 0.02 0.02 0.02 100.00 NOTE: Percentage frequencies are adjusted for differences in sampling rates of reference strata. HOLE* £ex,ceijcff&! ixinSneircjGii no rg^nz^g tot g7tt.01.om:© QJTJJE. J.U. ■■•-'ii i r~ ro i:-o- TOO-. - d : j 90- 83 *3 1 iO- ia*a ■{ eo- etra ao- e&*a so-* the percentage frequencies were adjusted for differences in sampling rate of the reference strata. The slight differences in the marginal frequencies of harvested acreage in Tables 3 and 5 are accounted for by the fact that Stratum III was not enumerated in the preharvest survey so that, unlike Table 3, the rate at which Stratum III was sampled had to be taken account of in the construction of Table 5. It will be noted that for almost 25 per cent of raisin- type units, har- vest intentions were not definite enough to be specified numerically at the time of the preharvest survey. For other units, the agreement between harvest inten- tions and harvest performance appears to be quite good. On the assumption that an entry in a diagonal cell of Table 5 indicates equality of intentions and perfor- mance, Table 5 shows that for approximately 86 per cent of respondent units, the agreement between intentions and harvest performance was exact. Similar impression is derived from Table 6 in which it is possible to compare by reference strata the acres intended to harvest and the acres actually harvested per respondent units. The agreement between these averages is excellent. Table 6 also makes it possible to compare, by reference strata, the harvest Performance per unit, for both 1950 and 1951, of the 1951 units with uncertain in- tentions (hereafter termed D.K. units) and the remainder of the sample. For both seasons, acreage harvested for raisins per unit was substantially smaller for D. K. M m aniaiei XxvrudBn lo ' holiou:ho~q at ?>3B©-t?ni Aisfenrsqmo') sriT, .p5SX ni dead nsrfd nworta sr dn9o t©q ili dxmdn brut xi ©XdsT n r. adcb tafcfdo td nwoiia si dn©3 wg dcods nsv.d "isridai bi&tn b©aB©*3fU -terid edaassjua airiT - .BdBb Moerio noiiojj&mq dnwxtra y;d lo noidoxJboiq ni: ©acsrisni 9%t&i Srid iol ©Xdiano^s©'! ^XxTBrti"iq 3i ©^sanoB baaeemoni -©no dxrccte x rf bsasMpni bX©X\ ifirid .dosl iri <£?dBoxbrsr &J»b ©Xqr.Bil ,X2?1 ni anxaiBi baiii? yii Bidden) eniciBt Xia iol br^ta9v"iad ©jssiob lo aedsmid39 ifi?nx£ ©riT .btirid ,X$P1 Yo'i a«o« UtM)' ^ fans' 'idl mos O$$»x0 0%?.^ &i& (bsdoaeXd fans ,8/xgtx© bisbfis ja bajciil^a'i? enea Vjadc/nlda© ©rid- jnlwoXiol sde^oBid ni a&uj^ll ©riT ©rid tfioil inol ria9i1'ni driji-xitio bPSBrimxiq' qoio s lo noidi3oqaib Xsnil ©rid esni«3 lo'i b^daevnsd e^Bs-ioa leioi lo 2edsfofrt39 ( ied£»X end- od mrofiJi ?d don x&m iswottg noid:>-rv?v ■taaviBria'xq ©d? n99wdsd noxdfiX9t ©rid aworte £ aitfiiT iCtoz'nte X5£X arid, gnitx/b sniais-i -itl bPdeavtBri .I + d^iXa a«dT i-edaide esnaiPlei ©rid lo edsi att'ds-xdo dsrid dosl edd ^d to'i feMnx/oooB 1 916 < bns £ aeXtfeT nl -e^esiq* b^deovisd to ade-i erid t £ aXdeT 9?iXJny ,dsdi* oa xevtoa deevosdeiq sxid nx •h©d-at9!Tix/n9 don acv III lo rtoid^xndsnoo arid 'nx dnxioooa n^Jlsd 9d od bBft b^XqiriBa. esw III raxrdsida rioxdw i& -isn «^.dinu e ixd-nxaist "''la' iiieb 'i9q daomXa lol derid bedon od XX xw dl ,5 ©XdeT amid add dl \llx-6xi«nx/n beilxoaqa ©d od d3xrona ©dXnildb doa 9iew anoidnstni tasv- -nsdnX daeviBd nes^deu-f dnpaKJ^i^fl arid t adlaa -xaddo 7o"$ . {y.^vixra dsev-iBfi^iq 9rij- 1c rta dadd noitqfnx(88B dxid rtO .booa 9dix/p'9d 04 eiaaqqs aonsrraol'teq da?viBri bns anoXd -tclipq bnc anoidnsdtix lo ^dXiBijpe' ^ddfinibnx 5 ©.xdsT-lo Xiao iano^sib b rtl ^rdn© ©rid f 3dXnx; dnebnoqeei Jm-?' tsq xXodarixoiq'TB tol dadd-aworia 3 aXdsT .eansm noxae.eiqnti TsXivrra .dosxe asw 90J T Bitiio'ii9q daavian bne encxdnednx n?9wd9d dnsmeet^s ©dd Bd£id3 sonoial^T V ©^aqiEoo od ©Xdxaaoq ai dX rloXriw ni 5 .©X r,: sT mo*rl beviiab 8i .adinx; dnobnoqaei ts-q b^daaviari -iilBXtdda e.etnc pdd- bh^ daartsri od babnednx senor. .dripXigoV© aX a«%6't©vs <>e©dd wwfed dn«m9fi?B w^T d39vi.3fi ©rid t .adSTia ©onsT^ls-i^d teicqrao? od ©idiaaog dX 3e?tem oai© b ©XdsT -ni n^d-X90rtb lidiw adiriu X5eX edd lo J^X be* 05^X ddodf lol. ,di«iu l©q e>mettnnj'i*q adod -xol .©Iq/nsa odd lo •i©bniBir,©ic ©rid bns (ai&ru -.3.0 bpir^ad, ladlBei^r.) srtoidiw? .2 .a id ifitXisms \XXaidnad3du8 a«w dinu ^c»q aniaisi -xol b©Jaovicri ©ib©i>b ^^noaBes TABLE 5 Acreage Intentions and Acreage Harvested for Raisins, Three Main Raisin Varieties; Per Cent of Units in Specified Acreage Classes — Seven Counties Surveyed, 1951 | Acreage harvested for : raisins, Acreage intentions, 1351 > ! ~ Un-" certain 0 ' o.I- 9.3 io- " 19. 9 20- 29.9 50- 39.9 40- Cn 50- 53.9 60- 69.9 DTI on n on yu- no o yy.y. l nn "1 O /I C 125- i en J.OU— xi 4_.y J- 1 O— TOO O 200- 2^9 9 >~ ouu— ^OO Q 400 499.9 Total j 1951 ■ 0 (0.1- 9.9 : 10- 19.9 14. S2 3.68 3.83 1.18 1.23 1.75 13/75 0.36 0.38 1.00 18.93 0.86 0.56 0.59 0.11 0.05 0.05 39.24 20.02 j 24.99 { 20- 23.9 1.42 0.18 0.09 0.23 5.21 0.05 0.05 7.23 j i 30- 39.9 0.69 0.18 0.05 0.12 4.02 0.05 0.05 5.16 i 40- 49.9 | 50- 59.9 0.05 0.11 0.03 J. 05 0.64 0.u3 0.05 0.03 I-m!' U04 i 0.80 i 60- 69.9 ! 70- 79.9 • 80- 89.3 0.05 0.05 0.05 0.05 0.03 0.05 a. 44" 0.08 0.28 0.49 0.16 i 90- 99.9 100-124.9 125-143.9 150-174.9 175-199.9 I200-2S9.9 |300-399.9 |400-499.9 0.03 J. 03 0._06. 0.05 oTIs" [ 0.08 j 0.27 j 0.03 0.03 0.06, j 0.03 0.03 0.01 0.03 j 0.04 i I 0.03 0.05 0.06 1 0.03 0.05 1 1 0.01 . 0.02. ! Total 1 24.53 23.76 15.93 21.19 7.17 t - 4.02 1.17 0.74 0.31 0.54 0.11 0.13 u . .. 0.16 — — 0.06 0.06 0.03 0.07 0.01 0.01 100.00 I j NOTE; Percentage frequencies are adjusted for differences in sampling rates of reference strata. H Cn L r 1 ■ ■ Q"02 L" -1 ! SO- 8c>'3 0*02 1 ! 1 jo- cva'a €0- 6c» k d O'Oy a* it so-- TO- Td'd n j- .9* j o- I8 ;•• Mb m O'Oe O'TI xt\i j 0 T fcj 2 I fliT* - 1 ' OT-j rOJJ? 3T7CJ \ TABLE 6 Bearing Acreage and Acreage Harvested for Raisins Respondents and D. K.'s, Seven Counties Surveyed, 1951 Proharvest and Final Surveys 1950 Bttt s , J D. K .'s __ 1951 Bearing. Acrep_ Per JJnit ' a/ 14.7 54.7 113.4 15.3 54.8 158.9 Acres Harvested for Raisin .Per, IJn it*/ 4.? 18.4 20.8 9.7 27. 2 55.8 14.7 52.8 115.8 4.9 22.0 20.8 Acres Intended to H arvest P er Unitg/ 9.9 28.9 55.5 Per Cen t of Bearing Acr es Harvested for Ra isins 62.9 8 35.1 a/ Preharvest Survey, August, 1951. b/ Final Survey, Sopteiiber-October, 1951. ■r • 8.*S5 . 1 NS [ 1 i ! s.^x II III 8.33 I II \ 1 313* II III 3 fix " x aT I i 0,31 I I S J..UG j 1 17. units than for respondents. One possible explanation is that on the average the bearing acreage in D. K. units is smaller than among respondents, although some control over size of unit is imposed by the use of reference strata. Table 6 in- dicates that the average size of unit among respondents is slightly larger than for D.K. units in Reference Strata I and II and substantially so in Stratum III. These differences, however, are not large enough to account for the markedly larger acreage harvested for raisins by respondents. Differences between percentages of bearing' acreage harvested for raisins in the two groups of units parallel, in fact, the differences in means. The inference seems unavoidable, at least with reference to 1951 data, that operators responding as uncertain to questions regarding inten- tions will on the average allocate a smaller acreage to production of raisins than operators whose preseason plans are definite. Production of Raisin-Type Grapes Linear, ratio, and difference estimates of 1951 total production of raisin- type grapes in the seven surveyed counties together with estimated standard errors are shown in Table U. It will be noted that the standard errors of both the ratio and difference estimates are about half the standard errors of linear estimates, even with Stratum III enumerated for the latter, but not the former. The standard errors are relatively quite small, amounting in the case of linear estimates to about 2.5 per cent and in the case of ratio and difference estimates to about 1.2 per cent. Linear estimate for 1950 is also shown in this table. Multiplying this estimate by 1.07^1, the ratio of 1950 bearing, acreage of all varieties of raisin- type grapes in the state to the bearing acreage of the three varieties sampled in the seven surveyed counties gives approximately 1,386,000 tons fresh weight as an estimate of the state production of raisin-type grapes. This estimate compares favorably with the latest check figure of 1,326,000 tons for 1950 as given in the 1951 December California Fruit and Nut Annual Summary. Applying the same expansion factor to the 1951 sample estimate gives an indication of l,8ll;,000 tons which can be compared with the current official estimate of 1,805,000 tons.i^ Official esti- mates of production by varieties are not made. In the preharvest survey, the operators were also asked to indicate the pro- duction of raisin- type grapes expected on their units in 1951. As can be seen from 1/ This expansion method assumes that the yield per acre in the counties surveyed is the same as in the nonsurveyed counties and that the yield of nonsurveyed varie- ties is the same as of the varieties surveyed. Data from other grape samples obtained at the same time as the raisin production survey data indicate a lower yield in the latter counties. ori.t erii no 3. •iiR03 rtgoorl.t.ls t a^ni< -fii d slieT .s .texts 3i noicfwrjlqxs :or lo segednsoi.^q neevttd aaona lolled .adrtebnoqa-.n .ani a is $ iol b«>i? ttafll nr tlellKisq aJxat/ lo aqi r OT;j owi «rtl nl anisisi *tol b*te?Yi6n &* 9Mtt&9 4 i&t tiivi ip.c&l is t ©idsbxo vsnif amours coasts!. il iq 9? " * _ aoqstO dC^X-fijgxai lo rsoiS-nsfynl -nxalsi lo r.olio'fbo iq Into* 12? L lo e^tsmii *.? aoiWiecliB bus .oiisi sxlrtoo »*soxi> tfl not oi.+e'i | 3ioi'is» b-rsbnoja o>ii dsrid bsion il lo ctotno r ^'tEbiiK^a (?ti$ llsrf p?. co sitiJt nx ^n£ctnirc3B '».t I ETC iffa o»,If cop .nxl p/fd nx bsiuoss fctfiiTixdae noxdoi.'boiq srid nsd) is-rol XbIoxIIo X iecfcsdqo3 arid woled dnao isq l.i asw dl ,Xsiod 9ds.t8 s oi t st»od*£ as 9onei9llxb arid lo eonsoilingxa 9rid noidasi/p at noasei sXddiX el 9ir*riT .9j«mii39 ,e»^soib«i fcoisd no Idsludsd ?rid sA .sd'-rnxdss Isriil hns dgsvisrisiq odd neswdec sonsislei srid .lo rises nx .y.sviwe dasnsrieiq sfij nx bei-smlvtapiebnu e-.-v noxdoi/boiq .sdsida asqsiO sq^T-nxaififl 'to noldooboiSi b9SiIeeH bns bsiosqxH V3lneisllxb bnfi oldBi lo \i ttc l*i$quz dneusqua sri? .anol^ruerrxs isfrul snibnoq don lud) aedsnixiso i^anxl erft lo 210119 bxebnela er!d 11 neve bsfllBdniBm ax asdsm >lom noiicoxlidBida ft gnxaii bedcLi/oIeoai sis (eedBmid89 sonswllrb bne o.fdsi ©rij lo biBbnsia 9/tt t aariT .slqmee erfi iniWBib ni bs^oloms yXXBudoB d^rid gniidm&asi x£t&»n *Sie I'^nxl srfd-sfixan t 0^I ni noidonboiq nXaiBi ifiiodsn lb edstnxdas iBsnil lo ions C8d t d 1c ions fnwbnada b ri*iw b&iBqmoo as anc* el t S sIiteT lo noidsollidsida ei noiiMjbsi axrlT ,(6 sIo'sT 9^3) Bcteitfe scnsielsi lo aiaBd etii no bftai/qpoo anoA nislsH) enoJ OdO^I od muz X^665 tons dried weight which represents deliveries to handlers to July 5, 1952. The linear estimate based on 1951 survey data is 135,589 tons (see Table h) and is within 1 per cent of the check figure. Further comparisons based on various estimates from the October, 1950, final survey, estimate released on October 19, 1950, and their difference in per cent from check data now avail- able are as follows: Estimates of production of natural raisins, 1950 Thompson Total*/ Differences Type of expansion Seedless Thompson Seedless Total*/ dried tons — ■ 1 1 per cent Linear 122,111 127,776 -U.3 -5.1 Ratio**/ 12li,713 130,ii6l -2.3 -3.1 Ratio, revisedly£/ 123,32ii 128,936 -3.U -U.2 Survey estimate as published^' 125,000 130,700 -2.1 -2.9 Recorded deliveries^/ to July 5, 1952 127, 65U 13U,557 0 0 a/ Includes Thompson Seedless, Muscat, and Sultana varieties. b/ Computed using ratios of sample indications of production of natural raisins in 1950 to 19i;9 weighted by estimated 19ii9 production within reference strata. c/ Based on revised estimate of production for 19lt9. d/ October 19, 1950. e/ Raisin Administrative Committee, Weekly Report of Deliveries to Handlers. f S4&I tot avpvtoa nojc^ifbo'tq ntzini s>At lo ach:oq3i en"* nJ bptetlduq noxtoirbonq -s*xi3Yrit i t> nx noidiubam lo ?.na fSE"i9mL'Cts> &i£i&qs>z cioil b&at£idd ©*i&w I$9X box t C?^X '0'$ .BbisY one ainsi'q id ?nox*nL'fanq nxaisi x&ixi&n lo as^fimrtea sriof nxmS ©eutt arii ^i&l nl bsoubo'jq anxexert Isiuififl iol noianfiqxe -tsonxl y,9vdjs iaevtsri aoiioubmq bailoailo ^XirteTn/i) t?rf* : riixw bcieqp'oo anoi 4dl\il£S qbw beysviaa -.lexiPxifiT -is*! ictujeri Q?QX no a^tn}/! >'oarfo ©idsiisvfi isoJxi sriT .anoj oY£ t $€S lo PtetrJtiw it»£ba&A o& aexisvxlefc ajneaeaqtrt rioxilw irf§iew bex*ib anod 5dd f 4fX ax noxjouboiq nx3 anoi $8&k«!£f ai sJeb fPfMtt X59X no beaed eisnii^aa isynil eriT .S5QI t 5 tXtfl anoaxrtflqfioo tttfNpt « 9*1/3.1 1 jlocrt'o t>ri«jr lo inao icq I fliri.ixv a± bns (il oXdsT ©£»a) baas^Xst aisnriias vYsvxt/a Iar.il ,.0$?! ^©cfotfoO s»dx wotl adtemxtfaa at/oiiev no beasd -Ixb/b won Bifib jfoerio moil dnao neq ni sonet?- 11 ib tisdt bns ,05?X t 9X isrfotoC no ■ •,' :awo.Ciol S£ ©TB sXds - r— : — • — — ^ i \_ 1 iffif 1 jT 1 aa?XbsaS i £'iW*h*hb 1 f 1 { El' tXT (1 i5X >ib=»^3 noeqnori? asfw/XonJ \b iq lo anoxisoxbni; olqmse lo aoiisi ^nxau be«t»/qmoO \d .9i5?X nol rroxdonboiq lo f.jsmxd'ae bo«».xve»i no D9Bb8 \a 20. Present indications are that the October final survey estimates for the 1951 season are not quite as satisfactory as those above for 1950. Seasonal deliveries through July 5, 1952, of natural Thompson Seedless, Muscat, and Sultana varieties as reported by the Raisin Advisory Committee amount to 220, k93 tons. Thus, the high- est estimate for 1951 shown in Table i; is about 7 per cent and the published survey estimate of 209,600 tons (October 10, 1951) about 5 per cent low. Of some inter- est also is a comparison of the published estimate based on the October sample and the estimate common in the trade at about the time the former was released. This is shown below. It should be added that a measure is not available at present of the accuracy of the check data used in appraising the 1950 and 1951 sample esti- mates. The check data are still subject to revision for both years but even in final form may not be completely accurate. Figures on seasonal deliveries are based on reoorts from a score or more firms, all of which may or may not have the exact data available at the time required or be able to distinguish receipts by croo years in which produced. Comparison of Estimates of Production of Natural Raisins From Thompson Seedless, Muscat, and Sultana Variety Grapes Season Published survey estimate^/ Estimate common in traded/ Present check data indicationc/ Per cent difference Survey Trade dried tons 19h9 226,300 259,000 235,576 -3-9 9.9 1950 130,700 135,000 13ii,665 -2.9 0.2 1951 209,600 235,100 220, U93 -h.9 6.6 a/ California Crop and livestock Reporting Service, Dried. Raisin Production Survey (November 3, 19l9)j Estimated Raisin Production— 1950 (October 19, 1950); Esti- mated Raisin Production— 1951 (October 10, 1951). b/ California Fruit News, November 5, 19u9 (p. 6); October lU, 1950 (p. 6); October 6, ~ 1951 (p. 6). c/ Raisin Administrative Committee. Weekly Report of Deliveries to Handlers. The check figures shown for 1950 and 195l represent seasonal deliveries (sweatbox weight) through July 5, 1 Q 52 of natural Thompson Seedless, Muscat, and Sultana raisins. 1 ea^BratJse xovius Isnil *iadoj-©0 9d-t .t^rlt 9ifl rirroi.+Grixbnx irjc z^iiovlLob XsnoasaS .08 fctBa r>.)xup foz ste nraeec 83x«tsi*irv £nsv+fuS bfus „tB©euM ,8e©Xbs©2 noaqraoriT Lenutza lo tS'^-X t ?. vXt'!. I'T-zotrlt ri^td ariJ t ai/rfT ..anoJ CSiUOSS at JauomBeeSSlmmo'J v-ioajrvbA niaxef artf yd b©cHoq-;<'t as \»riL>fj bsriaxidiiq edd bne ea.-.iX©* sew famo'i 9iU' ©mx* prid JucdF, tB ©bait add nx nomnroo 3+ftmxdaa exit lo in-3B3iq da 9XdfiX ifiVs . ion si ©ttuasom s itaxit babbs ©d bloods dl .woi9d r^worfa ?.x -.ides EfXqmefi XiiSX bcis 05^1 add ^nxaxsf-tqB nx ba'air B,tsb Jtoado ©rid lo yob-uxoob sdd lit n9V8 dud ■ aiaay dxpd. toi nnbaxvan od-'doatdcra Xfitfa atr.' sdsb Soedo 9dT .Badem sis 33x^-vil9b Xanose^e no eati^r*? ods 'iX^rxTqrooo ad ton yon mtol Isnil Id avfiri .ion iu ysm rioxxfw lo U.B t 8nnxx siom to atooa b moil edioo-a-x no n^aad ^d adqiaoaij daiirgnxdaxb od rfde 9d in baiiopet sjnii ©dt .+b ©XdsXieva fidgb tiBXv : ' Jhaot/botq rioxriw nr atsoy rcno moiT snie jfifl Xfix'xJfcV! to noxdo.fbor*! lo BPtBrnxta'" lo nosxT.fiqmoD esqstO Y.d9lisV Bn.otXi/3 bna tdaosifM ^eeXbnaS noaqmcdT ' 000 , , . . 5dd.xl£X 000 ?fX COX ,viuB owi f»riT . .eveisotq ni ssw iaavisri alrriw aaJsb boilxosqa oJ" aniaxai ©ii..f lo R\8fa £ niii.tbf bafesi <**rsw ateoqoT X5^£ ni .iieqs Jlsew eno snow noasaa doss x ^eXrra.n oriJ 1 lo Had-eno t ed-8or> aocbai oJ iab*to ni .v^vrx/a d*>B9 "io aieh gnL+To^s '#%re noasaa-nirtixw. artt Io dosa tdl hasrlliu bsw sqsig 9cn&r~n ;a tsi lo slnmBa b^dcum-navo brte ^Mftua .ft/iil orii to! sXqawa eftt babivoiq elLnu b(Vi9drarn-bbo jcy.av anxccsi Ifitttfsn lo agaanaf o* noxiibbB ni ,y9vxua bfioosa a At nx bear; sisw ctim/ _ *Jjw- IbiwJ on' Id nortfo ubmq tot bfWTeofrsrf 9gB^ioB brts s^ai* no soqsia mo-il bsiaaqxa lis to! b9.te9v1c.rf 9§b9'job arix no aiuxsiuqri moil baniasWo oaX* asw noxJ-Bcnolni «enia eniBisi lo e9q\\f lis lo noxioyboiq io! kt^Qcnr&ri ad ot bs.ioaoxe 5»^s9*iob ^aseoqii/q . bscfa 'Vijjrf aYfiid" lo vsdjuN 9rt\t bns t noes98 srii gnxTX'h ni baiosXXoo nox.tsnrB isenxj t d§s?ia6 f -tf.xxiBsd X&X lc esdajnxda;? -V aXdsT ni rtworia ais evavir/e no^ssB-nxridrw ?rid so £.a is': Isuxtfyi 1© nohtoubo-cq bns t 0^9I ni enxaxBix Xsmdsn no! badtfflttrif agsenoB •arBmiJea ad.f v^racfa'xvvTq bodcoxbni bs v v aA .aroo^i Xotdnoo sb badxo osXa 9i£ 0^|?I ni: 8ioii.> bishnsds Qrt't dsrfx baton ad flxvr d7 .S oXdcT 1c no&te'ottttaid-e. asie baXUdgb -bnsdc ':ix es :>buxin^Bnr 'to isbio o>mse ->r(J- jo vXri'axrot sib R^Jsmi^ag nocBae-niriixw lo 91 Io I YX-fBlifoWiBq { XBXirtcj8dxra lartfB-s 9ie aoisinxias noa£3B-nin*XK? ori* 'io swits -bisbna^a ^jgaa^B It totts hi&bo&iz BUMlilMC t a;/riT .Yt-'vnra no-tBga-nxriibT ^bix'i 9rii *rol iaix'l. *»rli tox (sj-Binli^s ad* lo) &nso ieq P ' i.-*vo ax e.txsiBi Xstu-ton iol t-^a^viBil Io enoJ- boxoaqxs bns bsi'esvijBd a^'tt tot &nao i9q ^ itrods bns Ys>v*ifJ8 noeBae-nxiitfiv •do 8BW a*Bb boillo-^qa or an±8:Bi lo noitouboiq no npWsrmo'ini ,bc>>tBi?. -XBT lo egsnnc^ ^hb b9.f g9V-ibi1 eAjsli Id ladmwn '3o..8!mDv+ ni:' a^svtxrc eaf)^ flc bWisi -o'o. abX-aiy nracfi T9ri*l9H .baj-a^VTuSri euitt a&qci^ irroil cio^B-rgqo y,d b/=».tonqx9 enxe ni noxJoxi'ooiq no noxieanolrti cvi&o^ijSO .bgojubcnq eniaim lo essnnotf 9dx yXeviiogc 9itt io. Jbqib bns bg^aiqmoo ai ^aavnan Xidxm bonxBTdo ad ^oxWbo t io6l at ,an6i baiib •tonx nx qX9rt araoa io sis smti 9z*rtt t iev9YJoK .eigXhriBri oi bo-ts-viXeb .9is snxaxBi' '.-ton o& ^89*iijXnx oufoa lb al il .'b^eovifid ^nisd sgennat- snc: *{X 9isorrxo^qqa gnitcc a- o* o*ni a^si* ^-lavrroo o* ypt& -igq aniaxAt lo gbniroq y/xboic+sxio 9d* gniai/ #sitt esttra rl nciJoubbtq nsrii ^gBr,noi lo B9im3stz9 -langid 'xitr.sf?. ianoo bdvxs anxci«t lo 9d>t icl snoi OOX t HOX ri^i-v t>9TBqflfi^o as O^j^XlVeinlsisqo mcrl yl^oaixb bgrrxe^do .(yavij/p bnocse %ritf-T.ol sno^ Ob5 t ^X tanxB^B es '06,^1 bns v^v*fua d-atil TABLE 7 Within-Season Acreage and Production Estimates, Raisin-Type Grapes and Natural Raisins, Seven Counties Surveyed, 1951 Type of estimate Acreage harvested for natural raisins i ut.ax trays har- , i -|&/ vesteds/ Tons" "of " natural raisins expected from grapes harvested ""Total acreage of raisin grapes harvested Acreage expected to harvest for raisins 1951 bearing .acreage _ Acres har- natural raisins. X X L/vXU.V' viUll of natural raisins, JL „ 2 3 . 5 7 8 ~_ .... acres 1,000 trays. tons, dried * weight acres tons, orisa weight , i ! First Withia-Season Survey t September 4 Linear estimate 58,829 41,285 108,135 78,812 111,928 213,545 91,821 135,455 Estimated standard error 4,839 3,698 •5,524 5,050 S,108 3,513 5,909 9,480 Socond Witliin-Season Survey, September 11 Linear estimate 93,362 70,040 135,432 117,206 111,569 216,091 96,586 I 140,303 Estimated standard error 5,258 4,297 11,431 5,324 5,914 3,527 6*555 j 8,804 a/ At 5.5 pounds of raisins per tray, these estimates correspond to the following dried weight tonnages of natural raisins: 113,534 (10,170) and 192,610 (11,817). 1 IPX. sruq ?, g 1 B*aaa | - 1 i in £*tos 2 v 2j2 ( &*60d 8TS TIT*oS8 6T*8ST ©&pcq ^,03. 1 j 23. Production of natural raisins to specified date may also be estimated by re- lating sample indications of number of trays or tonnage harvested to total produc- tion of natural raisins in the preceding season. Such ratio estimates based on reference acreage strata (competed using fojwula (7)) and their estimated standard errors are shown in the tabulation below for the two wi thin-season surveys in 1951. These estimates are somewhat lower than the corresponding linear expansions shown in Table 7, particularly for the secono wi thin-season survey, and have also some- what lower but still relatively large standard errcrs. Ratio Estimates of Natural Raisin Tonnage Expected from Graces Harvested to Specified Dates, 1951 Item First wj thi n- season, Sepremter k Second within- sea son, September 11 tons, dried weight Tonnage from trays harvested!' 112,775 183,611 Estimated^/ Standard error 9,058 9,803 Tonnage expected by operators 107,310 177,871 Estimated Standard error —————————— 7,806 9,552 a/ Reduced to tons at 5-5 pounds of raisins per tray. In units of 1,000 trays these estimates and their standard errors are hi, 009 O,20h) and 66,768 '3,565). Survey Costs The table below presents estimates of average miles of travel, average inter- view time (including time of travel), mileage cost, and total field cost per com- pleted schedule for the 1951 surveys. The latter average comprises all payments to enumerators for time ($l.h7 per hour) and mileage (graded from 5 to 7 cents per mile in calendar month). The average time and mileage cost for all surveys in 1951 was $1.82 per schedule; analogous average costs for 19h9 and 1950 were respec- tively $2.38 and $1.73. The relatively high average for 19U9 can be attributed in part to the fact that fewer enumerators were employed who resided in the locality of their work and in part to the unavoidably higher expenditures incurred in the location of sample units for the 19^9 preharvest survey. The 1950 and 1951 samples in the four main producing counties consisted very largely of the 19^9 sample units and, hence, these location costs were not reincurred in later surveys. Higher time u'ci M CS1 lB1U+£0 SW.CJfia txs 'ins aae&ai XXd.£6X ;J ad* ax r [v.amoa 9i£ TaX \Xavj:.J a MH &Mff!l agaiovB lo saXini agates 'la fcaJam-ttaa 8*naa$ iaq .taca Mail JC«fo.t 'brte eqcoLLn t (ieffftl$ 2o .sinl* gnihuloni) a.irii waxv e?fti*r\caq III «98i:i i.an$ emW'-.asrJeTjp edT ,(d*nom tabnaXsa fix aXxm -aaany: aiew (ftp I bna iol ai?c'o sa'a-r^va- aot>?oXwfc ;ftXubarf38 isq S8 .X$ -as* X3^X ni baWdi***** ad f»o'&&X'ioT aga-iava dgirf. ^XsvWaXo't adT .fc?.X$ tarn ffc.st v/evx* «lxJac ; oX ad* nx bebia a'X orfw ba^olqma -a-iaw .eiadx'tami'na -i«wal tadt &o&l ad* o*.vhtcq rjx bb*xuranx aanttfibnacpee 19^3 id . ?.5dabxoVamr ad* 0* toaq rix-bna tied* lo aalqaaa '^S- bna O&'X adT "."^avida aaav-iBdariq r ^lPI «d* rto'i atxrrt/ aXqaies' lo noiitsobX atfituf sXqsoae ' *^ yt — -<.-v c ..4 .. -«-. — IV. "■*• •■"»-■ • *.«S ' lit j'l .r.'JJt i .^e-njurtfrw bnodjS .-1*129 fwi* lo Ifisintaris nK o* YlsgtsI ha&ofrttzen It emldcnn wit) tm agA&toji to tftttuoir -lm eiffs men srro r,tlf no jtg.rsyb oIcrn<£:sj II.c.v Tjnxbsed Jain add- icbm' hs-bfencs sd II jw "tfr^nr^Tvssswr Pi3il9 eignoctssT no n^arcieft •I/B:~ib | 'ir b** *io*Bt<>qc «#$ tefttWjgi* nH^olrt* bite ixitm.aqqft MB') f ijttKf?)S-l' ! naoTt J-nsriWo iN# ^Hil-||j||^fiilfr' v •l^al 5 *** ^ «^ ^r-^^T A3lxf&&^oa 5o ffbxB'gseeoM rft iX^eusp, ^iif ^sfafrsrfi ia^- s>di. ^e/ro Ws^'p b od od - f •Vrf l*66 ..f-ssd .b/iis m» 25. reasonable approximation. This observation applies perhaps with greater force to production data than to acreage since the magnitude of the latter (e.g., acres harvested for production of natural raisins) is directly ascertainable by the op- erator. Ordinarily, only a small proportion of the raisin crop is delivered to handlers by October 13, the latest closing date of a final survey in the past three years. In most cases, therefore, the operator's response to questions in final survey on tons of raisins produced is inferential rather than directly factual, based as it is on approximate translation into tons of numbers of sweatboxes, field boxes, or other raisin containers. The same argument applies pari passu to operator's responses regarding total production (in fresh tons) of raisin-type graces although in cases of harvest completed by the time of the final survey for uses other than the production of raisins, the fresh tonnage disposed through commercial channels may be known directly to the operator. On the other hand, in a small proportion of units the crop may not even have been harvested by the time of the final survey. Response errors will naturally contribute to the total error to which esti- mates are subject. In addition, response errors may introduce a measurement bias. There are indications that estimates of production of natural raisins based on op- erator responses secured in the final surveys are, in fact, subject in some years at least to a downward response bias. This evidence will now be briefly reviewed. Information concerning production of natural raisins on sample units during the preceding season was secured from operators interviewed in the 1950 and 1951 preharvest surveys. It is, therefore, possible to compare, for identical units, production indications obtained in the final survey, at the time the actual weight of raisins produced was still not available to the majority of operators, with production figures provided some 11 months later, presumably based on all objective information regarding quantity produced which could possibly be known to the op- erator. Averages (per unit) for such comparisons are shown in Table 8 separately for the 1°U9 and 1950 seasons by reference strata. For the 19U9 season, the dif- ferences in production reported in the 19l9 final and the 1950 preharvest surveys, are, on the average, small and not consistent over strata. For the 1950 season the average differences are more substantial and are of the same sign in all strata. On the average, the 1950 production of natural raisins was underreported in the 1950 final survey by some 8 per cent in comparison with the 1951 survey. The analo- gous figure for 19ii9 is about 2 per cent. A further fact of interest is that underreporting in the 1950 final survey was not independent of the quantity of rai- sins produced on the unit. The data shown in Table 9 indicate that, on the average, oi &*rtx>\ itfg&T$ H&iyt aqsd-rjq aaiXqac nc-jriswaedo. e&fP .troiiBiasy^if^B sld&noBS&t 29*10*; ,.3.^ t9&&B£-&A& lo Abif&lftsh» add rocte z$Rtnon ' od nerid- s*«fa ftoidat/bOTrt ~qc> sif.t ^d-'snifd^oss ^X.*o9ii£j -sr'- (shiR-isi Xsiud^in lo no-rd-atrfcoTcr toI bade 9 v%3d od ber^VxXofc ex (pno'/ixeis* add nai^-jtjqwwf -Slam' b ^Xno-vtt&Bnttrtn ,*jct£io- *?iddd8sq add fit iga*n/a Xanxt « "to ed*b g«iao£&-.d»ed6l kid ,££ *id staXbnfri- XBntl ni snoids&xjp od aenonB-yi b •'locPst-aqo art* i9*jelaiadd ,398^0 daom nT- itressv jXsudoM \XdD93xb nend laddrn'JaB.xdrianalni.ai' hzoubonn snj:a.p?*i lo efrod no vavtwj t S3>:ocf;ts9Wa BisdJlfSB Id enod odnx ooidsXea^tvid sdfcmxXO > tqq£ no zi dl ae boBsd ys . g*H ffi 8 ** 'g^zXqqB d*?aiMn|ue osibb adT ;8tsrrJtei.->oo nrsis-x lerido 10 t 8axod bXatl sqTEt-nreJrjjrr lo (enod nearl • nx) noidoubotq X/idoi ?«±fnj?aat aoehoqesi a''iod.s~9Q-o of 'tol y i rtyff 'X snxl -add lo amid arid fcftdAXqjtwo; #S9TOsd lo b^sbo «r rigfloridXs $90^3 ri3foint* fosopaft ssmnod riaail 9-rfd K eK..eiz* 'lo- dol+onboia. add nor* J tarido aea,. ni t bnx>d t*rido arte) - nO" .tocfjwsqo arid oi ^Xdoo-jib n*orof -©d y;m *>f anrsria- Xfrioi^.w Mttd' eitd x'i bsdaWiBd- rood eved oava don \:sb"*|o*xo add stinut lo noidteqw? £.f"sme -xda^' r.'Aito o'i iotd laded add od ad#dx**dnoo YXXsi»dj?n 1 1 tit &swt& ^enocfesS " .a*rd draMtanriMr £ &$£fbo 4 Jdnx \£m -atoiia Bzaoqe^-i \ncirtbhB-ttX .deal-toa. «SB ec-dsw. <-qo no bated 'enteim •XeiodBn 'lo noxdouboiq lo* esdamxd:* . dsdd 6iroidMihfsi e*ts aiad?- sissv' 9ftoa nx daatdi/a t doal nx ^eis. b^vxmb ihnil add nx b-it&ooz ef-enoqedi •todsto *b3we±v9's v/lexdd yd won XI x* e&dobxva exrfT .Bcxd Gaftcqeai btfiwwf* « rod tze#& dp. fffiUrb Bdlflff \Co}3T53 no enieiat- /Bitfdan lo nofio/fboiq animvofloi nnxtnurtolhl I^-I 'bns t)^QI eaid rif k^tm&titk . V9o$v*aq&'Blvt\ beurs-st sb-t hobB'su •■snitso.i'jq erid. ,a.'inj.' Iteoxdnobi *w>l " 1 3*isq!tpr> oi aXdiaaoq »^"Solo79di ,zt dl .bvsvujS- i3. C| V*i r rjyto ddgi-aw Isadoc add aattd- art} jj^tc^a .Isiiil add- nx bsniftddo •8«oxd^^bM--' , '*J:toubwq dixv, , . t fcT>dei.jqo lo -tf.iru>l«;R arid od 'ildxl ffivra ■ df on ItidB. *»* baoobo^q eniB-im lo vtlittefcii Xlsno b^BBd xidsnttreaaq- v^9d/:I addnoir iX sojob babivctq z&iuaitl noliont'ciq -qo ad* 'od nwofijl •.•>d ^io.hes'oq bdifoo ribtdvr i>-: oi/bcrro Yd-x-t^Jaap gnxb^^s^'f oxdcrintolnc ■^Tsdfoidq&a' ? 9-idBT nx n.foris ei& en ca ii r.qnioo- rtocB-tol ''d.rn.cr i.^fr) B^ei^vJ; vioiT^e *llb sdd t no8s98 %1$X add* lof • -.ctstdB aonyir^is'; va' anoseae ^X'brw-.-^ox edd iol -ta^cdsiq O^JX 9rii -bns XjartJgt.^X 'add: tri bditoqat noXdot6p < iq ni e^ona-tal • »dd ffoer-se 05?X --and -io^ .rd^'iia-'tavo jwadaienoo don bfta 'JXBijw.^egB^B -ydd no -anc- .•edeitfa XXb • ei • ng i« 9 nxq .f. ajuoa \d T-^nua :X««fl-0^?I ffsdd ax da^tainx lo dosl «ihrid»««rl a. .drt&o TWqf S dwods ej? ^X -rol a-ifr^n auoa ia^ "to tdidnBf^>--add Id dn*ibtt$tifsbni don a/sw-xsvura X^nll- O^PX odd ni gwidaocpn 2bnn ssawsfTf 5 £>dd no- t dBdd edBoifotti ?.9£dst af nwoda ed«b odT .dihxr srii ito bsoxrfwm Bstia 26 TABLE 8 Comparison of Production of Natural Raisins Reported for Identical Units in Season Final and Subsequent Season Preharvest Surveys, 19U9 and 1950 Seasons Reference strata Reioorted production of natural raisins per unit Final as per cent of preharvest Number of units Current j Subsequent final ! preharvest tons 19U9 season^/ I if 0 96.7 II 100 62.1 61.1 101.6 III 111.2 112. U 98.9 IV h h 83. 5 U79.9 100.8 AT 1W 22. h 22.9 97.8 19^0 season I 502 ! 8.9 9.7 91.8 II 115 35.3 37.8 93. a III 60 57. U 63.6 90.3 IV 8 181- 2 193-3 93.7 All*/ 1 11.6 j 12.6 02.I a/ Sample units located in Fresno, Kings, Madera, and Tulare counties. b/ Weighted averages of stratum means; weights used are numbers of raisin ~ type grape units within the four acreage strata in seven surveyed counties. TABLE 9 Average Differences in Production Reports and Variances of These Differences Related to Quantity of Natural Raisins Produced, 1950 Season Reoorted production of , natural raisins-' Number of units Mean difference in production reports^/ Variance of differences tons tons per unit 0 351 + 0.3 7.1 0.1- 9.9 73 - 0.2 ' 3.U 10 - 19.9 66 - 0.3 23.9 20 - 29.9 S 58- - 2.0 39.7 30 - 39.9 hh - 3.7 93- b U0 - U9.9 21 - 6.1 11U.1 50 - 69.9 21 -10.6 271*. 7 70 - 99.9 15 + 2.9 66.3 100 -199.9 26 -10.7 735-9 200 -699.9 10 i . ■ — — -2U.9 2,2U0.1 a/ Based on reports in 1951 pretarvest survey. b/ Production of natural raisins reported in 1950 final survey ~~ minus production for identical units as reported in the 1951 preharvest survey. .7s 1 ■ S.O V, ■ •" P 0 - FY 1 - 01 O.S - . . j v> . *?S ~ OS ■ I Ju *i 1.1 XX " J 1 1 • • *.ox- i is - 05 I > V.OX- : ^ s . ] 9 .J^^X— oox J. -Uiii^i • j r ; § COS • )£9VTi/e j-asvtwisrr'- I ' : ;: . .eiim; *9XXxunB *iol nsrii aniaxs'T'lo aaxtxinst'c ~9t aonXs d^ttoq inn v.X'Sai+xmbs ax avods basxi nozrrzwnoo lo trcxr.bnsia axfT yicmam ox- ioatrit/a a-xs ncl-int/bom^e ■asay tjnxb'aoaiq an/Msas-* a'roisiaqo lo aaanoq'a ajatfoxi irfgJrew no ,9£qmsx9 lol ,nVrorf-e a;= airi^isw qoto iewto* jnoliioie ib bns asid &1a enoarrsqroo oayrfi tHeaWii^vaV .aisb- 3?oarfa yjoioslaiisa «toh sb rvoiq- hXuow fixaxisi Csnfi.sn lo eotemitea 'Haifa c-.t ions lo a-vinoa rtotam p. lo ov#j\*>xbni yiVfel (faoiieq ^n.tx r^siB arfi ion iud) noxioabo-cq arid- lo fens erii is" hsiuqma notiotrbotq n r qmsc grfi lo ssxe arfi gnxasstonx yd bar X<->tinco ed ionnsn tsia»rw doa8 .aldsxi sis to t»dm^cZ ax baisqoiq ad oi evsri .eaisisiiae xioiioafooic nxexsi isdi navxO -oalta gniXXot^noo lo yiiXxixaaoq yXno srli iarsi issqqs btcow ii r T3tfo-*o« nx YXisa otra atnafflstusfism aviioocdo «flfca lo nroiioobotinx adi ni eaxl eio«na eenoqasi yXsv& «MOBSon. grfi lo iifiq Xsisainx ne as Bt4Hiia#aoo n.iaxsi lo aiftuoo. bnc airigxoiw a I .yevxna Xsnxl srfi lo iaoo adi yXIaiinr>icdtfe aassTonl XXxw axdT .aiabarjorq inain baasd atubsoono yavit/a XboH s tsii^afftr snxwtdiab bbo noii-fisiiapvftx -ladiiirl y/nt) oxiexJco-i a t^bm sXdisssl XXs. is ei .Etnsaiawsaan :9viio3fcdo no i-xeq nx 10 yXXoriw- . J\ .iogbx/d isfenaiiix tn&texl s isrii b9i6tt r ?os vfaii&no ,a.t^Xrwo b^iqaooc 9d ieirci bu&sfid siinxlab e aXc.uss oi rfoxriTr matt ana**! s' ai ion ixid b^Xqwca 9d oi bobnsinx noiisfxrqnc 9ili nx ais rfoxdv? ain^n&Xd oXcnrsa oi ncxprfsqx9 bnc aifif/oo airorifv r*x eb^lDni bn^ ofqmsa oi ^ncvsr! xo bna araei?. ' srfi Xi/qoq grii nx i9§noX on iod ,3nii5if odi nx XXxia p.+osnrofg scroiixioix atoiaal -xb 8niXqas« t sisn/oos bos otvXqmoo ^Idsncosat . ax rfoinV aiax^s yrxisH a 1 r ,ioY aij;an©qitioo nx.di 9103? yXinairpqil doidw asgsinFVbs aioi^.o iaiX e rioirs .id-xx yXioat ' ~aX9 lo gnxiaxX a t amfXoJh:sB ni .ba-toiniroafra 9d-Y3m isrii ?9liXx»oijxib 9di 10I .a-ixny lo no-tis-xamuno aisd s isu^ nsrii 9*rom yXxisnibT". 3 x (aiex:ia ii: I£ , ! einim T D3xqr!oa 3a oj noxxcifjqoq Qtij gnxrrtocr 00 noiiannolllX &rooa aabcvom gniirxX 9dT 91 ni noijsoxlxiciia as rioas aaoxvab lo sax* ingloxlla ni5 sXdsna id^iffr do'inw sisb rii b»bbr, sd bXworia JI .eair^xxse xo yiHxcfR^sv ^niXoms?. >ti& y^Xexin/ktadya aoi/b ni yXai/oi:i38 ax §nxiaxX arii BRrrf bavftab as aioiosl noxsnsoxa. .10 yosiwoos 'aiJi li ae^di no oXi&otxb ep.aX bnaqgb doxdw b-.yoXqffra ad eomXJsmos ysin aaiswxiaa t idi>ob loi 9mi1 3ifi sb- b9V*»08 xioitfw t vXXat/nns baeivg«r ,eixnr oqsirj- lo gnstaxX adT yd aeqesig nx agsaits rfrii jxnu rfosy -10I babxvo^c ,ay3v*tu8 noxiot/hoT:q nxsxsi adi ass 10I baiaoibfix aeiisxl arii t eyqs-ia sa^i-rtiaxai iol ^bIuo x3*ieq ni .eaiiaxtev I .a^xiax-Jsv snc-tiuS bna t isoaaM t a8slba©8 nobqmoriT lo agsartOB gaxised edi ttno 29. will be recalled that bearing acreage of the three major raisin grape varieties was used as a stratifying variable. One would expect that stratification on the basis of bearing acreage would be efficient for the estimation of total production of raisin-type grapes, but not necessarily so for the estimation of raisin production. As is evident from Table 3, many units (an estimated hO per cent of units in 1951) do not produce raisins, and for units producing raisins, acreage harvested for raisins is far from perfectly correlated with total bearing acreage of raisin-type grapes . Some indication of the efficiency of stratification may be obtained by com- oaring the (estimated) sampling variance cf estimates provided by the stratified design with analogous (estimated) variance appropriate for unrestricted sampling. It is easily shown that an unbiased estimate of the over-all variance derived from data of a stratified sample is given by h N n in which | , h a h and the other symbols are as defined previously. A measure of relative precision of stratified as compared with unrestricted sampling is given for linear expansion by the ratio tj M(N - n)/n S a^U" 1 - £j which also supplies an approximate indication of the factor by which sample size must be increased if unrestricted sampling is to provide estimates matching the sampling accuracy of stratified data. This factor is shown in the tabulation below for four linear expansions based on the 1^1 final survey. As w a s expected, the efficiency of stratification ap- pears to be considerably greater for items pertaining to grapes than to raisins. Nevertheless, gains from stratification for all linear expansions are consider- able. The smallest gain is obtained in the estimation of production of natural raisins. Even in this case, calculations show that an unrestrictedly drawn sample of some 1,323 units would be needed to match the sampling accuracy of a stratified sample of 7?8 units. These remarks apply only to linear expansions. For other types of estimates shown in Table h, gains from stratification, if any, are likely to be more moderate. Km or 0-7 30. Efficiency of Stratification (Reference Strata I-IIl) Linear Expansions, 19!>1 Final Survey i Estimated sampling variance^/ Rplative Item Unrestricted Stratified precision Bearing acreage 8U.1 20.6 U.i Acreage harvested for natural raisins 37.1 19.2 1.9 Production of raisin- type grapes 5630 1823 3.1 Production of natural raisins 1U2.7 82.9 1.7 _ a/ Entries should be multiplied by 10° to obtain actual magnitudes. • A 31. APPENDIX Schedules of the 1951 Raisin Production Surveys I. Preharvest Survey Schedule California Crop and Livestock Reporting Service No. P. 0. Box 1258, Sacramento 6, Calif. 1951 Raisin Grape Preharvest Survey (Name of Operator) (Street, Route, Box) (City) (County) (Note Corrections) Phone No. Location of Vineyard Unit: Sec. : Twp. : Range : By Roads ; (Correct if necessary) Thompson Muscat Sultana 1. Bearing acres of raisin grapes in unit in 1950 Acres Acres Acres 2. Acres harvested for raisins in 1950 Acres Acres Acres 3. Tons of natural raisins Droduced in 1950 from Dry Dry Dry unit tons tons tons h. Additional bleached tons produced in 1950 from unit " " " 5- Bearing acres of raisin grapes in unit, 1951 Acres Acres Acres 6. As of today how many of the above bearing acres do you intend to harvest for raisins this year? . . . Acres Acres Acres 7. Estimated total oroduction of grapes on this unit: Fr. Fr. Fr. a) in fresh tons in 1951 » • tons tons tons b) in fresh tons in 1950 n ■ " 8. Has harvest of this unit been completed for this season or have all the vines been girdled so that no raisins will be produced this year? Interview made by Date: August , 1951 (Put comments on back) 301 vise Pat'et" Septembe r 1951 9l\ib?rfo2 bvsv-i/jS noas93-.':xri*iW Y.?v%u& noJth>ubcn1 season Thompson S.I Muscat Sultana acres acres acres Estimated number of tons of natural rai- sins produced in 1951 from the above acres harvested dry tons How many additional tons df any) of bleached raisins were made this season from above acreages har- vested dry tons Please estimate the total production of grapes on this unit in 1951, in fresh tons dry tons dry tons dry tons fr. tons fr. tons drv tons fr. tons Enumerator: Date : October , 1951 £»0