Linguistics Resources

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Structure of a corpus-linguistic research report ======

A research report should describe at least five aspects of your corpus study: the **aims** of your study, the **materials** from which your data are derived, the **methods** by which you have extracted and annotated the data, the **results** of your analysis and **discussion** of how your results relate to the aims of your study. Each of these aspects must be described in sufficient detail for another researcher to understand your research design decisions and to replicate your study.

**Aims**. Describe your hypothesis or research question and explain how you plan to investigate it. In the case of a hypothesis, state specific predictions that can be tested using a corpus, in the case of a more general research question, state as precisely as possible what you will look for in your corpus.

**Materials**. Describe your corpus. If it is an existing corpus, state exactly which version of the corpus you used and explain why you chose this particular corpus. If it is a corpus you assembled specifically for your research project, explain as precisely as possible what texts you included in the corpus and how you selected them (if they are publicly available, state the source and where to find it). If your study is based on a sample rather than the entire corpus, explain how you selected this sample.

**Methods**. There are two broad steps that must be described: First, the procedure you used to extract the data from your corpus, second, the procedure(s) you applied to the data after extracting them. 

**//Data extraction//**. Explain how you extracted the data. If you used software, give the name and version of the software and state any information necessary to understand how you used this software; for example, describe the exact structure of your query or queries. If you did not use software, describe each step you performed in order to find the data in your corpus. If you only used a subsample of the extracted data, explain how this subsample was selected. If you checked your data manually, removing false hits, describe the kinds of data points that you removed and state how many of the original data points were removed in this way.

**//Data annotation//**. Describe how you categorized the data: Explain what categories you assigned, how each category was defined and how you identified it in the data.

**//Statistical evaluation//**. Describe the statistical test(s) you applied to the data and explain why you chose these particular tests.

**Results**. Show the results of your analysis. First, give descriptive statistics (frequencies, averages, etc.) in the form of tables (and, if you find this useful, additionally in the form of figures such as bar plots or line graphs; avoid pie charts). Then, give the results of the statistical procedure(s). Make sure to state all relevant information. 

**Discussion**. State whether your results support your hypothesis or falsify it. You should not not add any interpretation at this point, but you may state ways in which your design might have influenced your results and suggest additional ways in which your initial hypothesis could be tested.


A research paper consists of one or more such research reports. In addition, a research paper will include an introduction which provides background information, such as: Why are you investigating the hypothesis? How does it fit into previous research on the more general area in which it is located? Have other people investigated this hypothesis, and if so, what have they found? A research paper will also contain a General Discussion that summarizes the results and discussions of the individual research reports and interprets them – for example, by explaining how our knowledge about the research area is affected by your results or what additional research questions have been raised.
Linguistics Resources

User Tools

Site Tools

Page Tools
