Once you have obtained access to the restricted area with the data, we suggest you start with a Simple Search. These searches are all conducted from within the password protected data section of Carolinas Miner. As you already know, all speaker names are aliases and must be used in any publication that arises from this data. Sometimes real names will be used in transcripts or sound files – these must never be used in publications.
When you reach the home page of Carolinas Miner, there are 3 options:
• install required components for sound to work
• filter speakers
If you want to listen to the examples, you need first to install the required components so that sound will work on your computer: you will be given prompts for this action.
In order to search for specific examples, choose the filter speakers option.
To do a simple search:
1) Go to the home page of Carolinas Miner
2) Click on filter speakers
3) A range of options is available: experiment with them. If fyou want to find all speakers with Diabetes, for example, click Condition or disease, and then click on Diabetes. A list of all speakers with Diabetes will be shown. You can further refine the search by clicking on other options. For example, if you next click on Ethnic affiliation, you can find all African American speakers with Diabetes.
4) Click the top left hand square to select all speakers
5) Click non-layered search
6) Type in the word or phrase you want to search for. If you type ‘doctor’ you will get a list of each time the word is used by each of the speakers.
7) Each item will have two icons next to it. The first will take you straight to the transcript, the second will take you to the transcript with audio or video. Because it can be slow to download the sound, it is usually better to check the transcript without sound before investigating the option with sound or video. If you click on the first icon beside one of the examples, this will open up the transcript. The line with the item you are searching for will normally be at the top of the screen.
8) If you want to listen to the sound, click on the second icon. This opens up the transcript with sound/video. If you click on it, several options will be available. If you click play, you can listen to what the speaker said. If a certificate appears, click "Trust."
9) Once you have accepted the certificate, there will be another wait while the sound or video loads. You will first be taken to the top of the transcripot where there are two progress bars. You can see the progress of the download on the top bar. Once the sound/video is fully loaded, the second bar will change to "Ready." The line with the item yuou are searching for will again normally be at the top of the screen. The line may appear at the top of the screen before playback is fully downloaded. Click on the line and several options will be available. If the option play is not available, the sound file is still downloading.
10) To play the whole of the transcript, click the play arrow at the top right of the screen; you may need to scroll back to the top of the page to find the play arrow. The stop button allows you to stop playing, and the back button goes back to the previous utterance on the screen. The scroll bar on the right of the browser window allows you to move back or forward within the transcript. If the interview has been broken into segments, Previous at the start of the section or Next at the end takes you to other sections of the interview. The back arrow of your browser allows you to return to the list of utterances and listen to other examples.
11) By default, the box "Only search transcripts for which these are the main speakers" is checked. The main speaker is always the interviewee. If you want to find out whether the interviewer used a word or expression as well as the interviewee, remove this check and re-do the search. You will find that some of the interviewers use the word doctor as well as some of the interviewees.
12) The search will find all examples of the expression you are searching for. If you search for like, you will also get all examples of dislike, fuss will find fussy or fussing as well as fuss. You may need to experiment with how you phrase the search item in order to find what is most useful. Searching for the, for example, will get all examples of then, them and there as well as the. If you search for ‘the ’ (the with a space after it) you will only find examples of the. The search is also case sensitive. Capital letters are only used for proper nouns. However, if a word may sometimes be a proper noun and sometimes not (such as marine) you can search for ‘[Mm]arine’. This searches for both ‘Marine’ and ‘marine’.
13) You can use regular expressions in your search. As already noted, [AB] will return words with either A or B in that position. Clicking Regular expressions will give you a list of the expressions you can use. Searching for ‘the [aeiou]’, for example, will extract all examples of the followed by words starting with vowels.
14) Limiting the speakers you are searching. The filter criteria on the filter speaker page allow you to choose a subgroup of the speakers from the total corpus. Click on each of the buttons to see the options. If you click on Role, for example, you can choose to search only the interviewees. (If you choose to search the community or student interviewers, you will need to uncheck the box that says only search transcripts for which these are the main speakers before you do your search.)
Saving the results of your search
If you want to save the results of your search, click the box that says Export results to Excel. This saves the search results as a CSV (comma-separated values) file which can be opened in Excel. (See figure 2 for an example spreadsheet. There are usually more tokens than shown in this example.) On some computers, the file will open automatically, on others you need to open it manually. If you add extra columns, analysis can be entered directly into the spreadsheet. The options link next to Export results to Excel allows you to specify the information you want extracted for each utterance containing the item you have searched for. The URL option should always be checked. This allows you to listen to each of your examples without redoing the search. In order to make the URL aterialstion on the formula bar and click Return (see figure 3). The URL will now show as a live link and will take you back to Carolinas Miner. Note that you will have to log into Carolinas Miner with your appropriate user code and password in order to listen to the sound. See the next section for another way to work with a saved list of search results. It is always a good idea to make sure you have refined your search before checking the box to export the results to Excel.
Figure 2 shows an example of a spreadsheet generated from a search for doctor. The Number column shows how many examples were found. The transcript column shows the interviews where the examples occurred, with the Speaker column showing the speaker’s name. Sync shows whereabouts the example is in the sound file (the first occurrence of doctors in figure 2 occurs 67.693 seconds from the start of the interview). The URL allows you to return to the example without redoing your search, and the Text column shows you the context of the search word.
Figure 2: part of an Excel spreadsheet containing results of a search for doctor.
Figure 3: The first sync in column E has been turned into a url. To make the second live, place the cursor in the window on the formula bar and hit Return.
You will find information on downloading under Downloading Materials, and additional information on searching techniques by pulling down the tab for Linguistic Search.