Style Guide for Transcribing and Editing
Overview
This style guide will teach you the basics of how to transcribe for CastingWords. You should also see the grading guidelines and example transcript to get a better feel for what we want. However you should never transcribe without reading through this style guide and following what it says.
Lastly, although we try to document and explain what to do in any situation we can't cover everything, just use your head and do what makes sense.
Goal
An accurate and readable transcript.
This means do not transcribe verbatim but be accurate. Leave out false starts and non-sense words but transcribe exactly what the speaker says not just the general idea of what they said.
Other Helpful Pages
What If Style Guide - This style guide goes into deeper detail on issues in transcribing. Following it will get you 9s.
http://turkers.castingwords.com/ - Many useful resources for transcribers and editors.
Required Guidelines
Your transcription HIT will be rejected if you fail to follow these guidelines.
Speaker Labeling
Speaker name and colon start each new speaker. If there is a timestamp it goes before the speech, after the colon.
Label speakers with their first and last name the first time they appear. Label them with just the first name after that.
If you do not know a speaker's name use a descriptive name such as "Man 1:", "Woman 1:", "Host:", "Interviewer:", etc. Do not use "Male:", "Female:" or non-descriptive labels such as "Speaker 1:".
Only relabel the speaker when the speaker changes or after a block non-verbal (A non-verbal that has a blank line above it and below it).
Example (with timestamps)
Jerry Coleman: [00:00] This is the beginning of the transcription so I am labeled with my full name.
John Gale: [00:31] Right, but after the first time you will be labeled with only your first name right?
Jerry: [00:57] Correct. As you can see I am labeled with just my first name now and will be labeled the same throughout the transcript.
Man 1: [01:14] Hi, I'm a new speaker and no one ever mentions my name so the transcriber should simply give me a descriptive name. In this case the only information that can be gathered on me is that I'm male so I'm "Man 1:"
John: [02:01] Ah, cool.
Formatting
The rules are simple but important. Thanks for your attention.
- Speaker name and colon start each new speaker. If doing timestamps put them in brackets after the colon.
- Put a blank line between speakers.
- Put a blank line between paragraphs of the same speaker.
- Do NOT use new lines to format the lines of text. Use word wrap. If your transcript looks all disfigured when you paste into into mTurk, you did not use word wrap and you need to fix the transcript before you submit.
Special Tags
[xx]
Mark unintelligible and inaudible words and phrases with "[xx]". Marking parts with "[xx]" will only hurt your grade if the grader can understand what was said where the "[xx]" is. But inaccurately transcribing the section instead of marking as "[xx]" will hurt your grade much worst then an semi-audible part marked with it.
Note: Editors must remove all "[xx]" tags from the transcript. If they cannot figure out what is said they must mark it as [indecipherable] or [unintelligible].
Example
John: I was going out to the [xx] and a pack of buffalo were on the opposite hillside.
[sp]
Use "[sp]" if you are unsure of a term or names spelling. You should not mark any words that can easily be looked up with [sp] and remember to always use spell check on your transcript before you submit it.
Note: Editors must remove all "[sp]" tags after double checking and researching (if necessary) the term/name.
Example
John: I went out to Wichita [sp] to see an old friend.
[?]
If you are pretty sure of what was said but not completely sure you should mark it with "[?]" this will tell an editor that they should closely review the section of text that it comes after.
Note: Editors must remove all "[?]" tags after they have reviewed and correct the text that the tag marked.
Example
White: I went out to a baseball game [?] at midnight.
Non-verbals
Anything that occurs on the recording but isn't represented by what is said should be marked in brackets.
Everything in brackets should be lower-case unless it is a proper noun.
Examples
- [bell ringing] (or other description)
- [laughter] (or other description)
- [musical interlude]
- [radio break]
- [song]
- [song: Happy Birthday by John Petersman]
- Host: [1:45] I went out to Wichita. (If requested)
Preferred Guidelines (Follow these to get a high grade)
Consistency
Spell things the same way each time, even if it's incorrect. It's easier for an edit to correct 1 mistake repeated nine times then nine unique mistakes.
Cleanup the Transcript (But don't Paraphrase!)
Don't Transcribe Filler Words
This means that you should "translate" speech into writing by leaving out filler words like "um", "uh", "like", and "kind of", while still keeping the personality and meaning of the original utterance.
If the filler words contribute to the meaning and tone of the speaker, leave them there. But in almost all cases, they are a distraction in written language and should be left out.
We suggest you look over our grading examples to understand the balance between cleaning up what is said while still being accurate.
Clean up Verbal Stumbling but Retain the Speaker's Tone
It's OK to occasionally clean up grammar but the transcript must be an accurate representation of what was said.
Start Each Line with a Capital Letter
Use proper punctuation. This means periods, commas, capitalization of proper nouns, everything just like your English teacher taught you.
Spell Out One through Nine but use Numerals over Nine
So use nine not 9. But if the number is over nine use numerals such as 10.
Use "..." to mark a break
Such as a speaker breaking off mid-sentence, starting mid-sentence. If the audio cuts off mark it as [cuts off] (see non-verbals).
Don't Transcribe Radio Commercial Breaks
There is no need to transcribe radio commercial breaks unless specifically requested; simply mark the break as [radio break].
Use written form for URLs and spelled-out names
Even if the speaker spells out the name for clarity use the written version.
Examples
"Hello, I'm John Smythe" not "Hello, I'm John Smythe, that's S-M-Y-T-H-E."
"Go to mturk.com" not "Go to M-T-U-R-K dot C-O-M."
Use Word Wrap!
Do not break lines at 80 columns; just keep typing and allow the software to do the line wraps. Separate paragraphs with a blank line, not with a tab. If you are not using word wrap the transcript will become broken and awkward when you shrink the window instead of nicely adjusting to the new size.
Have Short Paragraphs
Keeping paragraphs down to 500 characters or 4 or 5 lines is ideal. Sometimes this can't be done, but monster paragraphs will affect scores, as they affect readability.
Submit just the transcript, as the transcription.
We don't need comments, titles, or things other than the transcription in the assignment box. This is what the comments box is for, putting them in the transcript box just adds work for the editors and thus will negatively affect your grade.
Spelling
Editors: check the spelling of proper names against resources such as the podcast's website.
Preferred Spellings
| Incorrect | Correct |
| Okay | OK |
| alright | all right |
| Email/e-mail | |
| Web site | website |
| internet | Internet |
| & | and |
| colour | color |
More Guidelines
If you're reading this far, we love you!
Abbreviations
Dnt use abbrvs b/c thr diff 2 read. Spell out the words.
Hyphenation
Use standard rules (dictionary or style manual) for hyphenation.
The trend is toward fewer hyphens, so omit them when the meaning remains clear. Be consistent.
Some common hyphenation guidelines:
- Hyphenate two or more words that precede and modify a noun as a unit, especially if the words include a past participle, a present participle, a single letter, or a number: line-by-line scrolling, read-only memory. I-beam construction, 8-point font, eight-sided polygon, copy-protected disc, free-moving electrons.
- Hyphenate compound numerals and fractions: three sixty-fourths, twenty-first birthday.
- Do not put a hyphen between an adverb ending in -ly and the word
it modifies: Windows is a highly graphical interface.
"its" versus "it's"
"It’s" always means "it is" or "it has." If you can change the "its" in your
sentence to "his." If it works, you can use|leave out the apostrophe.
"your" versus "you're"
"You're" is always a contraction of "you are." If you can substitute "you are"
in the sentence, then keep the apostrophe. If it doesn't make sense, use "your" instead.