| IN BROWSE MODE
|
To see a list of all available files
|
- Submit a search without specifying any of the speaker or speech event categories (i.e. leave all categories set to the default "all").
|
| IN SEARCH MODE
|
Types of searches allowed
|
- Single words or multi-word phrases
The wildcard character * may be used at the end (but not the beginning) of a search word to represent zero or more characters (e.g., typing in walk* will give you walk, walks, walked, walker, walkers, and walking). These characters appear separately to the right of the search word and can be sorted.
Current limitations: The search engine does not accept Boolean searches with OR (i.e., you cannot search for all instances of two or more terms, such as woman OR women).
|
To see concordance lines for any word or phrase used by any speaker in all transcripts
|
- Enter word or phrase in "Find" box on search page, and leave all speaker and transcript attributes set to "all" (the default setting).
|
To see concordance lines for any word or phrase used by any speaker in specific types of speech events only
|
- Enter word or phrase in "Find" box on search page, and select desired transcript attributes (e.g. speech event type, academic division, and academic discipline) in the "Transcript Attributes" to the right.
|
To see concordance lines for any word or phrase used by specific speakers in all transcripts
|
- Enter word or phrase in "Find" box on search page, and select desired speaker attributes (e.g. gender, academic role, native-speaker status) in the "Speaker Attributes" column in the middle.
|
HTML Transcript Conventions
|
- Search term is highlighted in red.
- Blue text indicates overlapping speech.
- Orange text indicates embedded turn (different speaker).
- <L> = laugh
- <P> = pause longer than 3 seconds
- <E> = contextual or non-verbal event (identified in utterance and full transcript views).
- For a full description of all transcription conventions, please see the Transcription Conventions page.
|
To re-sort results
|
- Use pull-down menus at top of search results page to specify primary (1), secondary (2), or tertiary (3) sort filters.
- Sort options include:
- 1st-5th word to the left or right (1L-5L, 1R-5R)
- The search term itself (e.g. in cases where the word may or may not be followed by a question mark)
|
To view entire utterance in which search term appears
|
- Click on "view" hypertext in the "View context" column to the right, and the full utterance in which the search term occurs is shown in the middle of the page.
|
To see speaker attributes for an utterance
|
- Click on "view" in the "View context" column to the right and a summary of the speaker characteristics is shown at the top of the page.
|
To get a quick overview of frequency data across speech events and speakers
|
- Click on "View results statistics" displayed at the top of the results page. This takes you to a new page that first lists simple descriptive statistics for the transcripts, including the raw and normalized (per 10,000 words) frequencies of the search term. The left-most column has a link to each transcript in which the search term occurs. The next tables provide the raw frequency of the hits according to academic division and interactivity rating. The final two tables display the raw frequencies distributed by gender and academic role.
|
To download results
|
- Click on "Download results in tab-delimited format" at the top of the results page, or on "Download results as XML" (only recommended for highly advanced users). Your browser will display a window called "Opening corpus", asking you to either (a) open the file directly or (b) save it to disk. We recommend that you save the result file onto your hard drive (option b). The file is automatically given the name "corpus" and is downloaded as a simple text file (.txt). To retrieve the file and save it as an Excel document, launch Excel and choose "open". Then go to the directory in which downloaded files are saved on your computer and select the right file. Remember to save your document. When you select "Save", the program will give you a few different options. Choose NO (to preserve the features) and save the file in the latest Excel format, with an .xls extension. Navigate to the directory in which you wish to save your file and give it a transparent name.
- For further information about the download feature, see Section 5.2 in the MICASE Manual version 3, which can be downloaded from the MICASE website.
|
| IN EITHER MODE
|
To select more than one item in a transcript attribute or speaker attribute category
|
- Hold down control key while clicking on additional selections.
|
To view entire transcript
|
- Click on name of the transcript in the far left "Transcript ID" column. At the top of the transcript view page, you also get some general information about the file (e.g. total number of participants). Click on the hypertext at the end of the file information header to "download entire transcript in XML".
|
Current limitations
|
- Note that limiting your search by too many transcript or speaker parameters may result in a very small number of hits, or none at all.
- A special threshold applies to searches for highly frequent words and expressions. Any time your results exceed 500, you are taken to a page that gives you four options as to how to proceed. Although you may continue with viewing the results in your browser, we recommend that you download the results as a tab-delimited file, and then open it in a program like Excel.
- If you discover something that doesn't seem to be working as it should, please let us know by sending a message to micase@umich.edu.
|