Bireye uyarlanmış testlerde madde kullanım sıklığı kontrol yöntemlerinin farklı koşullarda ölçme duyarlılığına ve test güvenliğine etkisi
2017
Bu calismada, bilgisayar ortaminda bireye uyarlanmis test (BOBUT)uygulamalarinda iki kategorili (1-0) puanlanan maddelerde madde kullanim sikligikontrol yontemlerinin farkli olcme kosullarinda (farkli yetenek dagilimlarinda, testuzunluklarinda ve orneklem buyukluklerinde) olcme duyarliligina ve test guvenligineetkisinin incelenmesi amaclanmistir. Monte Carlo simulasyon calismasi olan buarastirma, kuramsal bir arastirma olup calisma kapsaminda kisa test 25 madde, uzun test50 madde; kucuk orneklem 250 birey, genis orneklem 1000 birey olarak belirlenmistir.Testi alan bireylerin yetenek parametreleri, her bir orneklem buyuklugu kosulu icin (θ)degeri -3 ve +3 araliginda sirasiyla saga carpik, normal, tek bicimli ve sola carpikdagilim gosterecek sekilde BOBUT orneklemleri olusturulmustur. Arastirmadakullanim sikliginin kontrol edilmedigi durum referans olarak alinmis ve maddekullanim sikligi kontrol yontemlerinden Sympson Hetter stratejisi ve Asamali Dusurmeyontemi tercih edilmistir. Arastirma kapsaminda yer alan degiskenler ve kosullarcaprazlandiginda 48 simulasyon kosulu belirlenmis ve her bir kosulda 100 tekraryapilarak 4800 veri dosyasi olusturulmustur. Her bir kosul icin olcme duyarliligi ve testguvenligi indeksleri hesaplanip birbirleriyle karsilastirilmistir. Yapilan analizlersonucunda, farkli madde kullanim sikligi kontrolu kosullarinda olcme duyarliliklariarasinda genel olarak buyuk farklilik bulunmamistir. Ilgili kosullarda madde kullanimsikligi kontrol yontemlerinden Asamali Dusurme yontemi secildiginde, test guvenligidaha iyi saglanarak madde kullanim sikligi dengeli dagilim gosterdiginden maddehavuzu kullaniminin daha verimli hale geldigi ve madde havuzu ifsâ olmadan bircokuygulama icin sureklilik saglayabildigi saptanmistir. Bu dogrultuda, BOBUTvuygulamalarinda farkli olcme kosullarinda, madde kullanim sikligi kontrolyontemlerinden BOBUT uygulamasindan once, tekrarli simulasyonlarla kullanimsikligi kontrol parametrelerinin belirlendigi, BOBUT uygulamalarinda en yayginkullanilan Sympson Hetter stratejisi yerine, BOBUT uygulanma surecinde kullanimsikligi kontrol parametrelerinin es zamanli olarak belirlendigi Asamali Dusurmeyontemi secildiginde, olcme duyarliligini dusurmeden madde havuzunu daha dengelikullanarak test guvenliginin daha iyi saglandigi sonucuna ulasilmistir.AbstractIn this research, it was aimed to investigate the effect of item exposure controlmethods on measurement precision and test security for dichotomous items withindifferent conditions (different ability distributions, test lengths, sample sizes) incomputerized adaptive testing applications. Current research which is designed as aMonte Carlo simulation research is a theoretical research. Within the scope of theresearch, short test was designed to have 25 items while long test was designed to have50 items. In addition to this, small sample was simulated to have 250 observationswhereas large sample was simulated to have 1000 observations. Regarding the testtakers’ability parameters, computerized adaptive testing samples which have normal,right-skewed, left-skewed and uniform distribution for each sample size condition inwhich ability parameters were within -3 and +3 range were generated. The condition inwhich item exposure wasn’t controlled for was taken as the reference condition andSympson Hetter and Fade Away methods were compared with the reference condition.When the variables and conditions within the research were crossed, 48 simulationconditions were established and 4800 data files were generated by doing 100replications for each condition. Measurement precision and test security indicescomputed for each condition were compared to one another. The analyses conductedindicate that measurement precision indices corresponding to different item exposurecontrol conditions didn’t create substantial differences. It was found out that when FadeAway method was used, item exposure was distributed in a balanced way and highertest security was obtained. As a consequence of this, it was concluded that item pool gotvimore effective and sustainable without revelation (disclosure) of the items. Accordingly,it was discovered that in computerized adaptive testing applications in differentconditions, instead of Sympson Hetter method which is frequently used beforecomputerized adaptive testing applications and in which item exposure controlparameters are estimated by iterative simulations, the use of Fade Away method whichensures the item exposure control parameters to be estimated simultaneously resulted inhigher test security without reducing the measurement precision.
- Correction
- Cite
- Save
- Machine Reading By IdeaReader
0
References
0
Citations
NaN
KQI