{"id":729924,"date":"2022-08-20T09:00:18","date_gmt":"2022-08-20T09:00:18","guid":{"rendered":"https:\/\/www.questionpro.com\/blog\/tabakali-ornekleme-bir-olasilik-orneklemesi-turu\/"},"modified":"2022-08-20T09:00:18","modified_gmt":"2022-08-20T09:00:18","slug":"tabakali-ornekleme-bir-olasilik-orneklemesi-turu","status":"publish","type":"post","link":"https:\/\/qa-release.questionpro.com\/blog\/tr\/tabakali-ornekleme-bir-olasilik-orneklemesi-turu\/","title":{"rendered":"Tabakal\u0131 \u00d6rnekleme: Bir Olas\u0131l\u0131k \u00d6rneklemesi T\u00fcr\u00fc"},"content":{"rendered":"

Tabakal\u0131 \u00f6rnekleme, hedef pop\u00fclasyonun benzersiz, homojen segmentlere (tabakalara) ayr\u0131ld\u0131\u011f\u0131 ve ard\u0131ndan her segmentten (tabaka) basit rastgele bir \u00f6rneklemin se\u00e7ildi\u011fi bir \u00f6rnekleme prosed\u00fcr\u00fcd\u00fcr. \u00c7e\u015fitli katmanlardan se\u00e7ilen \u00f6rnekler tek bir \u00f6rneklemde birle\u015ftirilir. Bu \u00f6rnekleme prosed\u00fcr\u00fc bazen “ara s\u0131ra \u00fccret \u00f6rneklemesi” olarak da adland\u0131r\u0131l\u0131r. En iyi yakalama i\u00e7in hat\u0131rlanmas\u0131 gereken baz\u0131 hususlar hakk\u0131nda a\u015fa\u011f\u0131da bilgi edinin.<\/p>\n

Tabakal\u0131 \u00f6rnekleme, kullanabilece\u011fimiz olas\u0131l\u0131kl\u0131 \u00f6rnekleme t\u00fcrlerinden biridir. Zay\u0131f ve g\u00fc\u00e7l\u00fc y\u00f6nleri hakk\u0131nda daha fazla bilgi edinmek i\u00e7in sizi okumaya devam etmeye davet ediyorum.<\/p>\n

Tabakal\u0131 bir anket i\u00e7in se\u00e7im ad\u0131mlar\u0131<\/h2>\n

Tabakal\u0131 rastgele \u00f6rneklem se\u00e7iminde sekiz ana ad\u0131m vard\u0131r:<\/p>\n

    \n
  1. Hedef kitleyi tan\u0131mlay\u0131n.<\/li>\n
  2. Tabakaland\u0131rma de\u011fi\u015fken(ler)ini tan\u0131mlay\u0131n ve kullan\u0131lacak tabaka say\u0131s\u0131n\u0131 belirleyin. Tabakaland\u0131rma de\u011fi\u015fkenleri \u00e7al\u0131\u015fman\u0131n amac\u0131 ile ilgili olmal\u0131d\u0131r. \u00c7al\u0131\u015fman\u0131n amac\u0131 alt gruplara ili\u015fkin tahminler yapmaksa, tabakaland\u0131rma de\u011fi\u015fkenleri bu alt gruplarla ba\u011flant\u0131l\u0131 olmal\u0131d\u0131r. Yard\u0131mc\u0131 bilgilerin mevcudiyeti genellikle kullan\u0131lan tabakaland\u0131rma de\u011fi\u015fkenlerini belirler. Birden fazla tabakaland\u0131rma de\u011fi\u015fkeni kullan\u0131labilir. Tabakaland\u0131rma de\u011fi\u015fkenlerinin say\u0131s\u0131 artt\u0131k\u00e7a, baz\u0131 de\u011fi\u015fkenlerin di\u011fer de\u011fi\u015fkenlerin etkilerini iptal etme olas\u0131l\u0131\u011f\u0131n\u0131n artt\u0131\u011f\u0131n\u0131 d\u00fc\u015f\u00fcn\u00fcn. \u00d6zellikle, en fazla d\u00f6rt ila alt\u0131 tabakaland\u0131rma de\u011fi\u015fkeni ve bir de\u011fi\u015fkenin en fazla alt\u0131 tabakas\u0131 kullan\u0131lmal\u0131d\u0131r.<\/li>\n
  3. Mevcut bir test \u00e7er\u00e7evesi belirleyin veya hedef pop\u00fclasyondaki her bir \u00f6\u011fe i\u00e7in tabakaland\u0131rma de\u011fi\u015fken(ler)i hakk\u0131nda bilgi i\u00e7eren bir \u00e7er\u00e7eve geli\u015ftirin. \u00d6rneklem \u00e7er\u00e7evesi tabakala\u015fma de\u011fi\u015fkenleri hakk\u0131nda bilgi i\u00e7ermiyorsa, tabakala\u015fma m\u00fcmk\u00fcn olmayacakt\u0131r.<\/li>\n
  4. \u00d6rnekleme \u00e7er\u00e7evesini gizli, a\u015f\u0131r\u0131 gizli, \u00e7oklu ve k\u00fcmeleme a\u00e7\u0131s\u0131ndan de\u011ferlendirin ve gerekti\u011finde ayarlamalar yap\u0131n.<\/li>\n
  5. \u00d6rnekleme \u00e7er\u00e7evesini tabakalara ve de\u011fi\u015fken(ler)in tabakala\u015fma kategorilerine ay\u0131rarak her bir tabaka i\u00e7in bir \u00f6rnekleme \u00e7er\u00e7evesi olu\u015fturun. Tabaka i\u00e7inde farkl\u0131l\u0131klar en aza indirilmeli ve tabakalar aras\u0131ndaki farkl\u0131l\u0131klar en \u00fcst d\u00fczeye \u00e7\u0131kar\u0131lmal\u0131d\u0131r. Tabakalar \u00fcst \u00fcste gelmemeli, birlikte t\u00fcm n\u00fcfusu olu\u015fturmal\u0131d\u0131r. Tabakalar ba\u011f\u0131ms\u0131z olmal\u0131 ve n\u00fcfusun alt k\u00fcmesini d\u0131\u015flamal\u0131d\u0131r. N\u00fcfusun her bir unsuru tek bir tabakada yer almal\u0131d\u0131r.<\/li>\n
  6. Her bir \u00f6\u011feye benzersiz bir numara atay\u0131n.<\/li>\n
  7. Her bir tabaka i\u00e7in \u00f6rneklem b\u00fcy\u00fckl\u00fc\u011f\u00fcn\u00fc belirleyin. \u00d6rnekleme dahil edilen \u00f6\u011felerin \u00e7e\u015fitli katmanlar aras\u0131ndaki say\u0131sal da\u011f\u0131l\u0131m\u0131, uygulanacak test t\u00fcr\u00fcn\u00fc belirler. Orant\u0131l\u0131 bir tabakal\u0131 g\u00f6sterim veya orant\u0131s\u0131z tabakal\u0131 g\u00f6sterimin \u00e7e\u015fitli t\u00fcrlerinden biri olabilir.<\/li>\n
  8. Her tabakadan belirtilen say\u0131da \u00f6\u011feyi rastgele se\u00e7er. \u00d6rneklemi temsil etmek \u00fczere her tabakadan en az bir eleman se\u00e7ilmeli ve toplanan verilerden hesaplanan tahminlerin hata pay\u0131n\u0131 hesaplamak i\u00e7in her tabakadan en az iki eleman se\u00e7ilmelidir.<\/li>\n<\/ol>\n

    Orant\u0131l\u0131 Tabakal\u0131 \u00d6rnekleme<\/h2>\n

    Tabakal\u0131 \u00f6rneklemenin iki ana alt t\u00fcr\u00fc vard\u0131r: orant\u0131l\u0131 ve orant\u0131s\u0131z \u00f6rnekleme. Orant\u0131l\u0131 tabakaland\u0131rmada, \u00e7e\u015fitli tabakalara atanan \u00f6\u011felerin say\u0131s\u0131, tabakalar\u0131n hedef n\u00fcfusu temsil etme oran\u0131yla orant\u0131l\u0131d\u0131r. Yani, her bir tabakadan al\u0131nan \u00f6rneklem b\u00fcy\u00fckl\u00fc\u011f\u00fc, hedef n\u00fcfusun o tabakas\u0131n\u0131n g\u00f6receli b\u00fcy\u00fckl\u00fc\u011f\u00fcyle orant\u0131l\u0131d\u0131r.<\/p>\n

    \u00d6rnekleme kesri her bir tabakaya uygulanarak her bir n\u00fcfus unsuruna e\u015fit se\u00e7ilme f\u0131rsat\u0131 verilir. Elde edilen \u00f6rneklem kendi kendine a\u011f\u0131rl\u0131kland\u0131r\u0131l\u0131r. Bu \u00f6rnekleme prosed\u00fcr\u00fc, ara\u015ft\u0131rma evren parametrelerini tahmin etmeyi ama\u00e7lad\u0131\u011f\u0131nda kullan\u0131l\u0131r.<\/p>\n

    Ara\u015ft\u0131rmac\u0131 genellikle sadece n\u00fcfus parametrelerini tahmin etmek de\u011fil, ayn\u0131 zamanda nispeten k\u00fc\u00e7\u00fck bir tabaka i\u00e7inde ayr\u0131nt\u0131l\u0131 analiz yapmak ve\/veya tabakalar\u0131 birbirleriyle kar\u015f\u0131la\u015ft\u0131rmak ister. Orant\u0131l\u0131 tabakal\u0131 \u00f6rnekleme, bu t\u00fcr bir analizin baz\u0131 tabakalar\u0131nda sonu\u00e7 vermeyebilir.<\/p>\n

    Tablomuzda a\u00e7\u0131klanan \u00f6rne\u011fi ele al\u0131rsak, 2. b\u00f6lgedeki elementlerin detayl\u0131 bir analizini yapmak m\u00fcmk\u00fcn olmayacakt\u0131r \u00e7\u00fcnk\u00fc \u00f6rnekte sadece 12 element bulunmaktad\u0131r. Ayr\u0131ca, 2. b\u00f6lge unsurlar\u0131n\u0131n di\u011fer b\u00f6lgelerle kar\u015f\u0131la\u015ft\u0131r\u0131lmas\u0131 da ku\u015fkulu olacakt\u0131r.<\/p>\n

    Orant\u0131l\u0131 tabakal\u0131 \u00f6rnekleme, bu t\u00fcr bir analizi ger\u00e7ekle\u015ftirmek i\u00e7in iyi bir \u00f6rnekleme se\u00e7imi de\u011fildir. Orant\u0131s\u0131z daha iyi bir se\u00e7im olabilir.<\/p>\n

    Orant\u0131s\u0131z tabakal\u0131 \u00f6rnekleme<\/h2>\n

    Orant\u0131s\u0131z \u00f6rnekleme, her bir tabakadan \u00f6rne\u011fe dahil edilen unsurlar\u0131n say\u0131s\u0131n\u0131n toplam n\u00fcfustaki temsilleriyle orant\u0131l\u0131 olmad\u0131\u011f\u0131 bir prosed\u00fcrd\u00fcr. Evreni olu\u015fturan unsurlar\u0131n \u00f6rnekleme dahil olma \u015fans\u0131 e\u015fit de\u011fildir. Her tabaka i\u00e7in ayn\u0131 \u00f6rnekleme kesri ge\u00e7erli de\u011fildir.<\/p>\n

    \u00d6te yandan, tabakalar farkl\u0131 \u00f6rnekleme oranlar\u0131na sahiptir ve bu nedenle bu \u00f6rnekleme prosed\u00fcr\u00fc e\u015fit olas\u0131l\u0131kl\u0131 bir se\u00e7im de\u011fildir. Pop\u00fclasyon parametrelerini tahmin etmek i\u00e7in pop\u00fclasyon bile\u015fimi, \u00f6rneklemin orant\u0131s\u0131zl\u0131\u011f\u0131n\u0131 telafi etmelidir. Ancak baz\u0131 ara\u015ft\u0131rma projeleri i\u00e7in orant\u0131s\u0131z tabakal\u0131 \u00f6rnekleme, orant\u0131l\u0131 \u00f6rneklemeye g\u00f6re daha uygun olabilir.<\/p>\n

    Orant\u0131s\u0131z \u00f6rnekleme, atama ama\u00e7lar\u0131m\u0131za ba\u011fl\u0131 olarak \u00fc\u00e7 alt t\u00fcre ayr\u0131labilir. \u00d6rne\u011fin, katmanlar i\u00e7inde analizi kolayla\u015ft\u0131rmak, maliyeti, do\u011frulu\u011fu veya hem do\u011frulu\u011fu hem de maliyetleri optimize etmeye odaklanmak olabilir.<\/p>\n

    Bir \u00e7al\u0131\u015fman\u0131n amac\u0131, ara\u015ft\u0131rmac\u0131n\u0131n \u00f6rneklem katmanlar\u0131n\u0131n ayr\u0131nt\u0131l\u0131 bir analizini yapmas\u0131n\u0131 gerektirebilir. Orant\u0131l\u0131 tabakaland\u0131rma kullan\u0131l\u0131yorsa, bir tabakan\u0131n \u00f6rneklem b\u00fcy\u00fckl\u00fc\u011f\u00fc \u00e7ok k\u00fc\u00e7\u00fckt\u00fcr; bu nedenle \u00e7al\u0131\u015fman\u0131n hedeflerine ula\u015fmak zor olabilir.<\/p>\n

    Orant\u0131l\u0131 da\u011f\u0131l\u0131m, bu t\u00fcr detayl\u0131 analizler i\u00e7in yeterli say\u0131da vaka \u00fcretmeyebilir. Bir se\u00e7enek, k\u00fc\u00e7\u00fck veya seyrek tabakalar\u0131 a\u015f\u0131r\u0131 \u00f6rneklemektir. Bu t\u00fcr bir a\u015f\u0131r\u0131 \u00f6rnekleme, pop\u00fclasyona k\u0131yasla \u00f6rnek katmanlar\u0131n\u0131n orant\u0131s\u0131z da\u011f\u0131l\u0131m\u0131na yol a\u00e7acakt\u0131r. Ancak, \u00e7al\u0131\u015fman\u0131n ama\u00e7lar\u0131 do\u011frultusunda gerekli olan tabaka analizini ger\u00e7ekle\u015ftirmek i\u00e7in yeterli say\u0131da vaka olabilir.<\/p>\n

    Tabakal\u0131 \u00f6rneklemenin g\u00fc\u00e7l\u00fc ve zay\u0131f y\u00f6nleri<\/h2>\n

    Tabakal\u0131 \u00f6rnekleme, olas\u0131l\u0131kl\u0131 olmayan \u00f6rnekleme prosed\u00fcrlerine k\u0131yasla \u00e7o\u011fu olas\u0131l\u0131kl\u0131 \u00f6rnekleme prosed\u00fcr\u00fcyle ili\u015fkili g\u00fc\u00e7l\u00fc ve zay\u0131f y\u00f6nlerin \u00e7o\u011funa sahiptir.<\/p>\n

    Basit rastgele \u00f6rnekleme ile kar\u015f\u0131la\u015ft\u0131r\u0131ld\u0131\u011f\u0131nda, tabakal\u0131 \u00f6rneklemenin g\u00fc\u00e7l\u00fc y\u00f6nleri \u015funlard\u0131r:<\/p>\n