Tagged: omr

Thoughts from latest iteration on OMR

Well, we run the latest OMR sessions (numeracy & literacy) – there does seem to beย  little problem we need to iron out if I’m going to recommend this to staff and others for a more significant pilot. The problem is one of scaleability. After you upload the image (TIFF file) via our photocopier / scanner. Some of the scans need to be reviewed, after which when you open the csv file with the results others have not had the results included … so you need to manually add these to the csv (results) file. When you then check these records in the software (Form Return) they are listed as not yet processed … this is an inconvenience on small numbers (about 20% occurrence with the two problems on a submission), but it would simply not work for larger cohorts.

My plan is, with one more submission left, can I tweak things, use the software more effectively to reduce the rate of this occurring. Fingers crossed ๐Ÿ™‚

OMR – 5th iteration for Health

We are on the 5th iteration in Health. Some lessons learnt to date (these are all so obvious ….);

  • reduce the printing costs through remove the unique id for each test. So you can re-use the question booklets
  • get the discipline team to finalise the question booklet, and write answers before sending to us to create as create the answer sheets
  • include a box for them to add their name, as well as there candidate answer
  • include a clear title on the answer sheets which is generic enough not to cause questions. This is important when two papers and being tested in the same session
  • check the spreadsheet to investigate any blank cells for the answer – it seems some get through, not sure why this happens. But a responsibility should be for the person scanning the answer sheets to make sure all cells are entered, unless left blank on the answer scripts. this will let you double check, so don’t assume the system will do it for you ๐Ÿ™‚
  • ensure the sensitivity setting is appropriate on the answer sheet – to reduce the likelihood of scan errors
  • include a visual display of what we mean by shade the whole bubble – we still get lots with lines across etc.,

OMR (return form) for elevate team session evaluations

OK – I’m keen to start rolling out the use of OMR for objective feedback and assessment. So, given we are using the UCS Corporate Development session evaluation template, I’ve OMR’d it ๐Ÿ™‚ This will include some free text responses (which we’ll need to key).

Also, given it’s a survey and we aren’t 100% on who will attended, we’ll need to undertake a little csv file work.

The process is;

Before the session

  • print off the evaluation forms from DropBox >> EDU >> Projects >> OMR – Elevate Evaluation

At end of session

  • get to complete. Tell them, if going to identify themselves, please use the UCS Username, only other key information is the Data
  • scan and send to Andy as a TIFF (very important) via a MDF device

After session …. needs to be Andy until get lots of cash to buy some more packages

  • save and upload scanned file
  • uploaded scanned file to Dropbox using date as file name
  • process and clean up if required
  • export data as csv with the following settings
  1. Publication_ID >> rename as Event Title
  2. Form_ID >> keep as default
  3. Form Score >> uncheck
  4. Processed Time >> uncheck
  5. Source Date >> uncheck
  6. Captured data >> keep as default
  7. Individual scores >> uncheck
  • open csv, and add the following
    1. Event Title >> set and copy to something obvious
    2. Add column after Event Title >> date >> fill in with today’s date
    3. Add column after Date >> UCS ID >> fill in the UCS Username
    4. Add column after UCS ID >> facilitator >> fill in with our initials
    5. Add column at end >> free text >> type in free text response
    6. cut and paste into Google Spreadsheet >> The Elevate Event Evaluation Data
  • clean (remove) the data from the report in FormReturn so ready for next scan
  • delete csv … the data is in Google

OMR – FormReturn – Iteration 3

Authoring was really smooth this time ๐Ÿ™‚

The Elevate Team responsibilities are as follows;

  • create the answer sheet as a PDF and email to the discipline expert

This time, no changes to the question booklet, so we can run again. Discipline expert supplies the users (applicants in this case). Therefore, the process for us is …

Step 1 >> create the applicants_list_month_year_csv and store in Dropbox

Step 2 >> in Form Return, open the generic segments for both the Maths and Literacy Test (stored in the dropbox folder)

Step 3 >> make a new form, and include the segment. Add a barcode and the text

Step 4 >> add source data, a new table (follow emerging naming convention), and import the applicants_list_month_year_csvย  from Step 1

Step 5 >> publish and test (print some, complete, scan, upload data and download) – post test, delete the data

Step 6 >> email the answer sheet PDFs to the discipline expert

All done in 30 mins ๐Ÿ™‚

OMR Iteration 2 …initial thoughts

Well, we’ve run another two sets of interview tests for candidates using the OMR (FormReturn) system. From my perspective this was much better than the first iteration … good news ๐Ÿ™‚

  • I thought the answer sheets were better laid out
  • it included negative making (which worked)
  • it included candidates entering their applicant number as opposed to writing their name. This meant I could auto join the results with the person, as opposed to cut and paste
  • the processes and responsibilities are much clearer (nothing to do with the software but important for scaleability). The discipline team have taken responsibility for managing and printing the question booklets

Interestingly, this time we had a significant increase in the number of answer sheets which hadn’t been processed correctly (either it couldn’t read an answer or their application number). So there was an increase in the time I spent having to work through the unprocessed sheets in the software adding in the answer based on their paper copies. This isn’t a pain for me given the numbers (30 candidates). However, it wouldn’t scale particularly well. I assume, we could reduce this time by getting them to complete use a pen to fill in the bubbles at the end of the exam. Hence the contrast would be better than a pencil. I also, think there is a setting around the sensitivity. I’ll look into this for the next time.

Another need for next time is to auto generate the results (final score) as a percentage, as well as a absolute figure

The final enhancement for next time is I’d like to send the answer sheet as a PDF to the discipline team and they can print it out. I’ll then be able to design a work flow process.






Responsibilities & Work Flow Document: OMR

The following a worked up responsibilities and work flow document for deploying an MCQ Exam via the OMR service

Step 1: Initial start up meeting

Owner: Discipline Team, and Elevate Team

  • Discuss the process, identify work packages and agree on people involved. A timetable is agreed based on the following work flow. Agree on results format (what they need), and who should receive the results.

Step 2: Writing the questions, mocking up the answer sheet

Owner: Discipline Team

  • write the question booklet (include the questions, the instructions, give the test a unique number)
  • mock up the answer sheet (identify the number of available answer options, the correct responses, any marks / negative marking, and the unique test number)
  • pass the completed question booklet and mocked up answer sheet to The Elevate Team

Note >> the discipline team is responsible for writing and printing the question book. The Elevate Team are given the final copy of the mock up’s answer sheet

Step 3: User Accounts

Owner: Discipline Team

  • based on earlier conversations in Step 1, the discipline team will need to send (as an csv file) the user accounts for the exam (first name, last name and email address) to the Elevate Team

Step 4: Writing the answer sheet, and testing the system

Owner: The Elevate Team

  • create and test the answer sheet
  • sign off … by printing off the answer sheets

Note >> the Elevate Team will need 5 working days to create and test the answer sheet

Step 5: The Exam

Owner: The Discipline Team

  • print all the question booklets, invigilate the exam, ensure students follow the instructions
  • hand the completed answer sheets to The Elevate Team (we will try to collect from the room at the end of the session)

Step 6: Process the Results

Owner: The Elevate Team

  • Check the submissions and scan them
  • process the results
  • create the results doc (as agreed in step 1), and upload to Google Docs to be shared with the appropriate people
  • hand back the completed answer sheets to the discipline team



OMR FormReturn – Session 1 – observations

Well, we used the FormReturn OMR system in anger with two candidate interview tests (numeracy / literacy – MCQs)

It all went very well from a technical perspective,

  • The forms scanned and emailed to me as a PDF (24 submissions in one PDF file)
  • uploaded to FormReturn, which created the results.
  • Only one answersheet which did not scan / decode
  • I exported to csv, and open in excel to undertake an data manipulation
  • I flicked through the answer sheets to see if they’ve changed an answer – and check these / re-score in excel
  • In this version I got the student to write their name on the answer sheet – so I also needed to assign the person to the answer sheet in the excel sheet – cut and paste job
  • then saved the results and emailed to staff coordinator
  • the scanned pdfs / and csv answer sheets are stored in the myd dropbox area

The next iteration will include:

  • the user adding their unique ID, so saving me the cut and paste job
  • ensuring the alignment is better on the answer sheets (improved in latest version)
  • devolving the writing / printing of the question book to the discipline team … who will need to provide it to me by a set date
  • rolling this out to other members of the team
  • testing how long it takes to process the results … I’d imagine if installed on a laptop, which is taken to the room, we can have this sorted within minutes ๐Ÿ™‚
  • improve the reporting and analysis … discipline team lead on this
  • upload to Google Docs (spreadsheet) for online version … so not sending attached docs around the group

So overall … this is great, it works … it is simple ๐Ÿ™‚