Fourth session

Section 1: Introduction

Definition of key terms
Define the terms that you use in the title of the skripsi and the research questions. Please view a certain term as a term, and avoid considering a term as consisting of several words. For instance, 'reading comprehension' is a term and you have to define it as 'reading comprehension'. Do not define it as 'reading' and 'comprehension'.

Please do not include terms that you do not use in the title and the research questions. For example, you don't need to define the terms 'TEYL', 'TEFL' or 'audiolingual method' in section 1 if you don't put them in the title and the research questions, although actually you conduct a study related to those concepts.

Section 2: Literature Review

You may not write anything about the experimental group in literature review. Instead, you should review theories and previous studies relevant to your research. You should describe the experimental group in the methodology section.

Section 3: Methodology

Instrument
Multiple choice & short answer tests are objective tests. You don't need to have two raters because the results are always the same.

Be careful in selecting the instrument. The instrument you use depends on the the research question you state in section 1. For example, if the research question is "Is there a significant difference between X and Y?" you must use a test to collect data in the form of scores. You cannot use observation as the research instrument.

If you use interrater reliability to estimate the reliability of your instrument, you use correlation
to find out the relationship between the results obtained from both scorers. Please choose one of the formulas that is most suitable. Do not write both Pearson r and Spearman r in your proposal.

One of you want to use split-half reliability to estimate the reliability of a reading test. This consists of one reading passage and ten multiple choice items. You cannot use split-half reliability for a very short test because you will possibly obtain a low coefficient. Split-half reliability is more suitable for longer objective tests, such as those consisting of 100 items.