Wednesday, April 29, 2015

Cara menulis laporan keputusan t-test satu sampel

I've read answers like the one below for one-sample t-test.
"The estimates (M=8.57, SD = 1.70) are significantly higher than the test value, t(28) = 15.2, p = .009."
You may be forgiven for wondering what exactly is the 'test value'. The answer did not provide enough 
information. A possible better answer is:

"A group of students were asked to freely draw a straight line without the aid of a ruler. They were instructed to draw a line which is 5cm long. The length of the lines (= 8.57, SD = 1.70) drawn by the students are significantly longer than 5cm (the line's intended length), t(28) = 15.2, p = .009, and the effect size is large, d = 2.1" 


Dalam SPSS, bagi one-sample t-test, 'test value' tu istilah umum - mungkin merujuk kepada nilai purata populasi, nilai ambang (threshold), nilai piawai, atau apa-apa yang ditetapkan oleh pengkaji. Silalah guna maklumat yang lebih khusus bagi setiap masalah yang dikaji. 'test value' istilah umumiah atau generaliziah; dalam jawapan kena beri yang khususiah atau specifikiah. Kalau tak, kuranglah markah kau Kiah. 


Kalau bakal suami kau lafazkan "aku terima nikah perempuan ini binti lelaki ini dengan mas kahwinnya X ringgit Malaysia tunai", agak-agak kau nak rasa hangin ke, atau rasa nak sedekah tazkirah 30 minit kat dia? Engkau bukan pemboleh ubah dalam sistem aljabar. Ada nama dan identiti sendiri. Dalam konteks nak menikah, kena lah guna nama sendiri, kan?

Kau nak nota dalam bahasa apa lagi Bedah? Hibrid Melayu-Telugu?

p/s What mistake can you detect in the t-test results?

No comments: