KEBOLEHPERCAYAAN

Kebolehpercayaan

merujuk kepada kestabilan dan ketekalan dalaman ujian atau alat ukur dalam mengukur sesuatu konsep atau aspek. Sekiranya penyelidik yang lain melakukan penyelidikan yang sama, keputusan yang diperolehi adalah sama. Ataupun soalan yang sama di beri pada masa yang berbeza, jawapan yang diperolehi bagi soalan tersebut tetap sama.

Di dalam kelas, banyak alat pengukur yang boleh digunakan untuk mengukur tahap kebolehpercayaan. Namun begitu, penyelidik tidak boleh mendapatkan data ketika mereka sedang sibuk bermain. Oleh itu, setiap tindakan penyelidik boleh memberi kesan kepada kebolehpercayaan data yang di perolehi dan penyelidik perlu mengambil langkah-langkah yang sesuai.

kesahan ialah suatu yang sangat sukar dan kompleks. Kesahan merujuk sejauhmanakah sesuatu ujian atau alat ukur itu mengukur atau menguji perkara yang hendak diuji dan tidak lebih daripada itu.

Tinjauan secara berterusan kepada soalan-solan berikut akan meningkatkan kesahan:

1. Sejauhmanakah soalan tersebut benar-benar menggambarkan tentang data yang
    diperolehi

2. Apakah terdapat bukti-bukti yang menyokong kesahan soalan, perancangan dan

    keterangan yang menunjukkan bagaimana data tersebut diperolehi.

 

Reliability..

 

is synonymous with the consistency of a test, survey, observation, or other measuring

 

device.  Imagine stepping on your bathroom scale and weighing 140 pounds only to find

 

that your weight on the same scale changes to 180 pounds an hour later and 100 pounds

 

an hour after that.  Base on the inconsistency of this scale, any research relying on it would

 

certainly be unreliable.  Consider an important study on a new diet program that relies on

 

your inconsistent or unreliable bathroom scale as the main way to collect information

 

regarding weight change.  Would you consider their results accurate?

 

Types of reliability.

 

1. Test-retest Reliability

 

 - Merujuk kepada ketekalan atau stabiliti markah ujian jika dilakukan pada masa yang

   berbeza

 

2. Internal Consistency Reliability

 

•Refers to the consistency of a group of individual’s scores on two equivalent forms of a test

 designed to measure the same characteristic.

•Menggunakan satu alat yang dibina dan satu lagi yang piawai.

•Ditadbir ke atas subjek yang sama dan pada masa yang sama atau masa yang lain.

•Equivalent form means that two tests are constructed so that they are identical in every

 way except for the specific items asked on the test.

•This means that they have the same number of items, the items are the same difficulty

 level, the item measure the same construct, and the test is administered, scored, and

 interpreted in the same way.

•The two set of scores are than correlated. If this reliability coefficient to be very high and 

 positive, that is the individuals who do well on the first form of the test should also do well

 on the second form, and individuals who performed poorly on the first form of the test

 should perform poorly on the second test.

3. Equivalent Forms Reliability

 

•Internal consistency refers to how consistently the items on a test measure a single

 construct or concept.

•The test-retest methods of assessing reliability are general methods that can be used with

 just about any test.

•Internal consistency measures are convenient and are very popular with researchers

 because they require one group of individuals to take the test one time.

•Two indexes of internal consistency:

oSplit half reliability

 

- Splitting a test into two equivalent halves and then assessing the consistency of the scores across the two halves of the test.

           - Divide the test into halves and correlate the scores from the two halves.

           - Compute the correlation between scores on the two halves of the test using

            Spearman-Brown formula.

            - The low correlation indicates that the test was unreliable, a high correlation

            indicates that the test was reliable.

 

oCoefficient alpha

 

- Lee Cronbach 1951) developed coefficient alpha.. Alpha Cronbach

Coefficient alpha tells you the degree to which the items are interrelated.

           Rule of thumb:

At a minimum, greater than or equal to .07 for research purposes and somewhat

greater than that value (e.g. ≥ .09) for clinical testing purposes.

 

 

Mempertingkatkan Kebolehpercayaan

Berikut adalah cara bagaimana pensyarah boleh meningkatkan kebolehpercayaan sesuatu ujian yang dibina.

a. Arahan hendaklah jelas dan tepat.

b. Pernyataan soalan mestilah jelas dan tepat.

c. Soalan hendaklah seragam (homogen).

d. Situasi dan masa ujian dijalankan hendaklah piawai, serupa dan terkawal.

e. Elak ganguan kepada pelajar disebabkan arahan dan soalan tidak jelas.

f. Elak kebimbangan pelajar semasa mereka menjawab soalan.