CodePen - Grader instructions.Round 3

Go to: Scorer Portal, Instructions, Transcription History, Transcribe, Change Password 1. Create an account at https://literably.com/scorer-signup if you haven't already.*2. Read the instructions below thoroughly.3. Transcribe a total of six training recordings. This should not take you more than 30 minutes.

CodePen - Grader instructions.Round 3 Literably Grading Instructions Literably Grading Instructions

Go to: Scorer Portal, Instructions, Transcription History, Transcribe, Change Password

Benefits of working as a Literably transcriptionist:
  • Can make over $10 per hour
  • Get paid every week
  • Work from anywhere
  • Make your own schedule
  • No experience necessary
  • Help teachers and students

Here's how to apply:

1. Create an account at https://literably.com/scorer-signup if you haven't already.*
2. Read the instructions below thoroughly.
3. Transcribe a total of six training recordings. This should not take you more than 30 minutes.

Quick notes:

  • *Please use the first name, last name, and email address attached to your Paypal account or we won't be able to pay you (if you're hired). If you don't have a Paypal account yet, use the first name, last name, and email address that you will use on your Paypal account.
  • Hiring is 100% performance based. We'll contact you if your performance on your training recordings meets our minimum threshold. Resumes and cover letters will not be reviewed.
  • You may contact us by emailing grading@literably.com. Note that due to the volume of correspondence, we are able to reply to very few of the inquiries that we receive. All the information you need to complete a successful application is found below.

Before starting the task described below, please be aware:

  • In order to complete the application, you will need access to a computer (desktop/laptop), an internet connection, and the ability to use Google Chrome (which is free to download).
  • It is not required, but we highly recommend using a high-quality pair of headphones for our transcription work.
  • You will not be able to complete your application using a smartphone or tablet.
  • Literably reviews applications on a rolling basis all year long and we encourage you to apply at ANY TIME if you think you might be a good fit for the role.
  • We hire transcriptionists all year long but tend to do large hiring rounds during our busy seasons, when transcriptionists’ pay rates tend to be the highest:
    • Early August through October
    • January through February
    • Late April through early June
  • It is very difficult to complete your application task successfully if you just skim the below information or the below instructional videos.
  • Currently, we are only able to pay transcriptionists via PayPal.

You'll be presented with a series of audio files of mostly children, but occasionally adults, attempting to read prepared texts. The audio files range from a few seconds to a few minutes in length, but most are about one minute long.

Your task is to produce a complete and exact transcript of the word and non-word utterances in the audio file.

The reader will usually try to follow the text. So, to save you time, we've built an online transcription interface that presents you with the original text and allows you to modify the text to reflect exactly what is said in the audio file.

For each reading, you'll see an audio player and a text.

To control the audio, use the following shortcuts:

SpaceStart/Stop
Left ArrowBack 2 seconds
Right ArrowForward 2 seconds

As you listen, modify the text by hovering over a word or a space and using one of the following shortcuts:

AAddHover over space.
SSubstituteHover over word.
DDeleteHover over word.
EEndHover over word.

In more detail…

AUse A to add a word or non-word utterance.
SUse S to substitute a word or non-word for the word in the original text. Note that S is just a combination of D and A. It applies a D to the word you're hovering over and also pops up the Add box.
DUse D to mark a word in the original text as unsaid/said (it's reversible) or to delete a word that you've added by accident.
EUse E to mark all remaining words as unsaid/said (it's reversible). E is also useful if the reader skips several lines. Press E at the first unsaid word and then press E again where the reader resumes.

When you're satisfied with your transcription, press = to submit. If it's a training transcription, we'll report your performance and highlight the errors you've made. Press = again to move on to the next recording after reviewing your errors.

Note: our transcription software does not currently integrate with "transcription foot pedals."

Main Transcription Rules*:

When and when not to mark a recording as unscorable:

1. For recordings longer than 15 seconds (usually, these recordings are 1 or 2 minutes long):

  • If you don't feel that you can produce an accurate transcript of a recording (due to multiple voices, whispering, background noise, or microphone issues) just click "Mark as Unscorable." In these cases, you will not be penalized for marking a recording as unscorable, nor will customers be negatively affected.
  • If the student does NOT begin reading within 15 seconds of the start of the recording (e.g., they start reading after 20 seconds have passed), mark the recording as unscorable.
    • If the student does begin reading within 15 seconds, transcribe the recording as you normally would.
  • If there are no words heard in the recording and the recording lasts longer than 15 seconds, mark the recording as unscorable.
  • If the student reads for the entire time in a different language, mark the recording as unscorable.
2. For "partial recordings" (usually 15 seconds or shorter):
  • Most "short" recordings you encounter should be scorable, unless there is a clear and major audio quality issue (e.g., multiple voices, whispering, background noise, or a microphone problem) with the recording.
  • If there are no words spoken in the recording and the recording lasts less than 15 seconds, transcribe the recording as you normally would:
    • 1. If you see text displayed, you will mark all words as unsaid using the E button.
      • As an example, Screenshot A below shows you how you would transcribe Recording A.
    • 2. If you do not see text displayed, press = to submit.
  • If you don’t see text displayed for a given recording, you can still transcribe a recording and NOT have to mark the recording as unscorable. For example, if you hear the student read “we like to climb” within 15 seconds but you see no text on your screen, please use the “A” button to add “we like to climb” and then press = to submit.
    • As an example of this, Screenshot B below shows you how you would transcribe Recording B.

    Screenshot A

    Audio File A

Screenshot B

Audio File B

Transcribing the text title and formatting:

  • If the reader attempts to read the title, transcribe their attempt before the first word in the text.
  • Do not use punctuation except for apostrophes that are part of a word, like "isn't."
  • You can choose to transcribe in exclusively lowercase or to capitalize in the usual way, with one exception: please capitalize proper nouns!
  • If the reader substitutes a single word/sound for a whole string of words, it does not matter where you locate the substituted word/sound within the string of unsaid words. For example, imagine the text says "The brown cat runs," but the reader says "The klab runs." In this case, you could press S over brown, S over cat or A over any of the three spaces. All are equally correct as long as brown and cat are marked as unsaid and klab is inserted.

What if there are multiple ways to fit what the student said over the words in the reading? Sometimes there are multiple ways of transcribing the audio that are both accurate because you can choose to either mark a word said or make that word appear as an inserted word while marking that word unsaid. In these cases, we ask that you choose the option that leads to the student appearing to read the text as accurately as possible.

Ambiguous recordings:

  • There are times when it's not totally clear whether a reader is pronouncing a word correctly. In these cases, we'd like you to lean towards marking the word correct. Teachers get upset if they feel that their readers are being penalized for accent/pronunciation differences. So, in borderline cases, please mark the word as correct.

*There is an FAQ found below which includes other transcription rules to be aware of.

When the reader says a non-word, you should transcribe the sounds phonetically. You may be unfamiliar with phonemes and phonetic transcription. A "phoneme" is the smallest unit of human speech - like "ch" or "b." All words are made up of phonemes. For example, the word "shin" has three phonemes: "sh" "i" and "n."

Note that minor pronunciation errors that emerge from a student's accent or overall speech pattern should not be counted against the student and do not constitute a non-word. For example, imagine a student who generally pronounces her "th" sound as a "d" sound. If you hear "du" in place of "the," you should simply treat the word "the" as having been read correctly. You will often have to decide whether a student is reading a word incorrectly or whether the way she is reading the word is just part of her normal speech pattern.

We've adopted this phoneme chart. You should study the phonemes and their spellings carefully. You may want to print the phoneme chart for reference. We've also prepared a video to teach you the phoneme sounds and spellings. You must use these phonemes to transcribe non-words uttered; transcribing what you hear in a way that "looks right," but does not match the phonemes in the chart, is not acceptable.

When transcribing a non-word, you will simply string together the appropriate phoneme spellings. For example, if a child says "b" "l" "i-" "k," you should write "blik."

Two clarifications:

1. Sometimes, your phonetically transcribed non-word will coincidentally have the same spelling as a real word with a different pronunciation. In these case, just add a hyphen "-" to indicate that the spelling should be interpreted phonetically. For example, if you hear "i-" followed by "s," you should write "is-", "-is" or "i-s," because "is" will be interpreted as the word "is," which is pronounced "iz". Similarly, if you hear an "i-" phoneme all by itself, you should write "i-" not "i," because "i" will be interpreted as the word "I," which is pronounced "ie".

2. Because there is some overlap between phonetic spellings (e.g. "e" and "ee"), there will be times when the same phonetic transcription could be interpreted in multiple ways. For example, the non-word "keer" could be interpreted as "k" "ee" "r" or as "k" "e" "e" "r." Our parser starts at the left and prioritizes two-letter phonemes ("ee") over one-letter phonemes ("e"), so "keer" will be interpreted as "k" "ee" "r." As a result, if you hear "k" "e" "e" "r," you should be sure to separate the e's using hyphens. You could write "k-e-e-r" "ke-er" "k-e-er" or "k-e-er." All that matters here is that the e's are separated.

Here's a short YouTube video showing how to transcribe a recording. We recommend you view this video before transcribing your training recordings.

Initially, you will be transcribing training recordings. Training recordings are recordings that we have already transcribed. When you submit your transcription for a training recording, we'll show you our transcription below your transcription, and highlight the word errors in your transcription. We will also report the percentages of words/phonemes that matched our transcription. Note that there may be a delay of a few seconds while your transcription is analyzed.

We do not expect your transcription to match ours perfectly. Sometimes our transcriptions even have minor errors. In general, we are looking for <2% word differences and <10% phoneme differences, however, when we review your transcriptions we also consider qualitative factors (e.g., were very simple mistakes made). When you get to the point where you are consistently transcribing at this level of accuracy (or sometimes slightly worse), then we will personally review your transcriptions. If we think you're ready for real recordings, then we'll offer you a position as a Literably contractor. We'll have you agree to a contractor agreement and then start delivering you real recordings.

Once you are hired, you will be transcribing a mix of real recordings and training recordings, and you'll never know whether the current recording is a real recording or a training recording until after it's scored. If, after you are hired, one or more of your recordings reveals that your transcription accuracy has declined, we will be forced to permanently de-authorize you from continuing to work for Literably.

You can view a list of the recordings you've completed by going to https://literably.com/transcribed, which you can also access from the top of the transcription page. If you accidentally submitted a real recording too early or otherwise made an error, you can correct it by clicking on the link to the recording and re-submitting it. Sometimes it takes a few minutes of processing before a recording can be modified. You cannot resubmit a training recording.

If you're curious about how many recordings you've done, you can see a count on this page as well. This may be useful if we've asked you to complete a certain number of recordings.

We will pay you via Paypal once per week, generally Fridays, based on the total number of audio minutes accurately transcribed since your most recent payment. We cover all Paypal fees; you will not pay any fees as the recipient. The price per audio minute will fluctuate depending on the volume of audio to be transcribed, and can always be found at the Scorer Portal. The current range of price fluctuation is $0.30-$0.80 per minute of audio. As of December 2020, the average pay per audio minute over the past four months is about $0.51.

We highly recommend using a high-quality pair of headphones for transcription work. This can really help for picking up hard-to-hear words and phonemes. A great option for under $20 are these: SONY MDRZX100.

How should I handle pauses as students sound out words? As long as the reader says all of the right sounds in the right order, mark it as correct.

I hear a parent or teacher inserting a word, what should I do? You should ignore what the other voice(s) say and transcribe what the student says. The one exception is a case where the parent/teacher appears to be "giving away" the word, leading the student to say the word correctly. In this case, if the student then says the word correctly, please simply mark the word as unsaid (by crossing out the word and not inserting it). If you are unsure what to do or find this difficult, please mark the recording as unscorable.

The recording is cut off on a partially pronounced word, what should I write? Transcribe the partial pronunciation phonetically.

I can transcribe an utterance either using the phonetic transcription or using an actual word. Which should I use? We prefer that you use actual words when possible, but you will not be penalized on a training recording for failing to do so since we'll decompose any inserted words into their phonemes to determine your phonetic error rate.

If a student has a side conversation in the middle of reading, should I transcribe this? If a student makes utterances that are clearly not related to their attempt to read the text (e.g., talking to someone else in the room), you can leave those utterances out of the transcript. However, all utterances that are related to the student’s effort to read the text should be included.

What if there are multiple correct pronunciations of a word? For example, "bow" can be pronounced as "b-oe" or "b-ow". If the student's pronunciation matches a sense of the word that is unambiguously wrong given the context, mark it incorrect. Otherwise mark it correct. If there is any chance that the student's interpretation is correct, mark the word correct. For example, in the story "We Read," you should let the student pronounce the word "read" as either "red" or "reed"--each of the sentences could be interpreted as either the past tense or the present tense.

Should I transcribe numbers as words or numerals? Use numerals to save space if the numerals accurately describe what the student said. If there is no way to accurately describe what the student said with numerals, do a phonetic transcription.

What if the student doesn't attempt the reading? If you are asked to grade a recording where you don’t hear the student reading, please follow the guidelines found in the Main Transcription Rules section above.

What should I do if the audio recording makes it difficult or impossible to hear what the student is saying? Please mark those recordings as unscorable. This should be done, for example, when you hear pure background noise or inaudible whispering. In these cases, do not simply submit a transcript with every word unsaid or every word said.

What if the student starts recording after she starts reading? Please score the recording as you normally would. If the text starts by saying “The boy goes to school” and you only hear the student read "to school," please mark the words "the boy goes" as unsaid.

Should I pay attention to vocal stress in grading? No. As long as the student has the correct phonemes in the correct order, mark the word as correct, even if she emphasizes the wrong syllable(s)

What is the status of my application? When will I hear back from you? We generally contact successful applicants within two to twelve weeks. At any time, you can check your application status at the top of your transcription history page. We receive a very large volume of applications, so questions about the status of any individual application will be ignored. We also cannot guarantee that we will contact you until we are ready to move forward.

I keep getting a "no recordings" message, or the price is a lot lower than I want it to be. Most of Literably's audio files are generated from K-12 students reading out loud in a school setting in North America. For this reason, volume fluctuates a lot. There will be extremely few recordings available from the last week of June to the first week of August and around the times of U.S. Thanksgiving and Christmas. We expect the highest volume, and thus the highest prices, to be during February, May, September, and October.

I didn't use my Paypal email for my account. Can I change it? To be paid, your first name, last name, and email address on Literably must exactly match the credentials attached to your Paypal account. If you did not set these properly when you initially signed up, you can change them to match your Paypal account in the Scorer Portal once you have been hired. If you have not yet been approved yet, there is no way to change your Literably account information.

Can children transcribe on Literably? To avoid potential liability issues, we require that all transcriptionists be at least 18 years of age.

What information can I use for tax filing purposes? Literably's EIN is 46-1838070. Literably's mailing address is: Literably, Inc. 2261 Market Street #4177, San Francisco, CA 94114

Can I do more than six training recordings as part of my application? You may do more than six, but out of fairness, we compare applicants primarily on the first six recordings. We'll look at your later unpaid training recordings only if you're on the border of being approved. Please don't do more than 25 recordings. We don't want to waste your time or our time.

I accidentally screwed up one or two recordings and they are bringing down my scores. First of all, review your work before submitting and please don't submit recordings with errors! The students we serve need impeccable transcriptions and we take our responsibilities to them seriously. This is not a typical transcription job where a modest level of error is acceptable. Second, we have designed our review process to account for the possibility of a flukey training transcription. Please don't ask us to give it special attention beyond this; out of fairness to other applicants, we can't.

What locations can I work from? As of November 15, 2019, we will no longer engage individuals located in the following U.S. states as independent contractors: California, Connecticut, Massachusetts, Montana, Nebraska, or New Jersey. By signing up, you are confirming that you neither live nor will be performing services for Literably in any of the listed states. Your representations to this effect may be subject to ongoing confirmation by Literably.

Can I have more than one grader account with Literably? Graders should not have more than one active account with Literably. Submitting duplicate applications may lead to having your grader accounts suspended.

I was not approved. May I re-apply? Unfortunately we only have one application right now, so you will need to wait until we put together a new set of application materials. Feel free to check back in 1 year.

IMPORTANT: if you have not been contacted by Literably about being a super grader, you can skip this section.

You can access Literably's super grader directions HERE.

We receive a huge volume of email! As a result, we have worked hard to provide answers to every frequently asked question in our training materials. Before asking for help, please make sure you're using the latest version of Google Chrome and that (1) you've read all of these instructions thoroughly, (2) you've reviewed the phoneme chart, and (3) you've watched the YouTube instructional videos. If we don't reply to your question, it's likely that it's because the answer is found somewhere in the instructional materials.

You may contact us by emailing grading@literably.com. Note that due to the high volume of correspondence, we are able to reply to few of the inquiries that we receive.

ncG1vNJzZmirY2Ourq3ZqKWar6NjsLC5jqWgrZ2ilq%2BtxYyfoKWdo2S0s63DnqmYnJmnsqTAyKilrGaYqbqt

 Share!