Turkish Journal of Physical Medicine and Rehabilitation 1999 , Vol 2 , Num 4

Temel Biyostatistik Kavramlar - II- Uygun Hipotez Test Seçimi -

Merih Eryavuz 1 ,Şafak Şahir Karamehmetoğlu 2 ,İlhan Karacan 3 ,Halil Koyuncu 4
1 İstanbul Üniversitesi Cerrahpaşa Tıp Fakültesi Fiziksel Tıp ve Rehabilitasyon Anabilim Dalı, İstanbul
2 İstanbul Üniversitesi Cerrahpaşa Tıp Fakültesi, Fiziksel Tıp ve Rehabilitasyon Anabilim Dalı, İstanbul, Türkiye
3 Serbest Fiziyatrist
4 İstanbul Üniversitesi Cerrahpaşa Tıp Fakültesi, Fiziksel Tıp ve Rehabilitasyon Anabilim Dalı, İstanbul

Besic principles in biostatistics II - The parametric tests are preferred in thestatistical analyses, because they are more powerful and reliable. If thedistribution of variables is fit the normal distribution in comparing groups, the parametric test could perform.If the veriable is continual and the number of case is enought, the distribution of this veriable can fit the normal distribution. Sometimes,smililarity of the normal distribution curves (equality of variances) is important. If the number of case is not enough or the variances are not equal, or the feature of data is artificial (score, i.e.), the nonparametric tests must be selected. These tests can be used for analysing the qualitative and quantitative values. Although indications of the nonparametric test are wider than parametric test, they are less reliable.

Keywords : hypothesis tests, selecting