Once a set of endorsements or ratings is entered - manually or using the harvester - staff and interns are responsible for checking that the information on the scorecard matches what we have in admin and on our website. No matter how careful we are, mistakes are bound to happen. That's why checks are so important. If possible, webchecks should be performed by people other than the importer.
Ratings
Common Manual Entry Errors
The first webcheck should be comparing the information on the original SIG scorecard with the information on Vote Smart's internal website. Enter the date and the initials of the person checking onto the CEC Google drive tracking sheet. If there are no errors, click "Release to live web" on admin. Spot check on the live web to double-check the data properly displays. If there are errors such as the score being entered incorrectly, a rated official/candidate is missing, or a rating is attributed the wrong rating/candidate - return to admin and enter the correct information. Examples include deleting candidates that are not on the original scorecard, adjusting scores if incorrect, and adding candidates that are on the original scorecard, but not on VoteSmart.org.- Accidentally entering a score for somebody that was not in office during the scorecard's session
- Candidates/officials with the same first and last name
- Missing officials (especially for older scorecards with people out of office)
- Empty ratings
- Missing the percent sign (%)
- Sort the rating list on admin in a way that matches the scorecard as closely as possible - i.e., if a scorecard sorts scores by last name and chamber, you will want to select "by Office, Last Name"
Checking Harvested Scorecards
Common Harvested Entry Errors
Errors as a result of automated processes like scraping tend to follow a pattern. Most likely these issues are not one-off errors. Look for patterns.
Checking harvested scorecards differs from checking the manual entry in that it requires extra steps. These extra steps mimic steps taken during entry. You will find what scorecards have been harvested by checking the "type of entry" column in the current CEC tracking sheet in Google Drive. Once you go to the SIGs page, you will notice ratings that have been harvested will have a date, but not a title. Click on the date to open the scorecard. You will need to enter the title of the scorecard, select the appropriate "Rating Text Name", and lastly fill in the appropriate categories and tags for the rating. Categories should match the categories given to the SIG. The ratings should only receive different categories if the original scorecard is completely different than the mission (ie. a veterans group releases a scorecard about taxes). After you have completed entering blank fields, save the rating and release it to the internal web. Once released to the internal web, check the scorecard as you would a manual entryErrors as a result of automated processes like scraping tend to follow a pattern. Most likely these issues are not one-off errors. Look for patterns.
- Poor rounding
- Common last names that multiple people share
- hyphenated names, accents, tildes
- Nicknames
- Missing officials (especially for older scorecards with people out of office)
- Empty ratings
- If you find errors - especially errors that fall into a pattern - and the scorecard has been harvested, you will need to edit the .csv file uploaded to the harvester and reharvest the scorecard.
It is preferred to do the checks on the computer. However, if needed, checks can be performed on paper - it's easy to get tired eyes staring at a computer for a long time. In this case, you will compare the original-printed scorecard with a proofing report.
- In the "Rating for Candidates/Officials: Number of errors:" field, enter "0" (zero) and click on "Generate Proofing Report"
- Print the Report - if possible, use recycled paper.
- The Proofing Report is a representation of what appears on our website. All current officials will appear first, and anyone no longer in office will be last on the list. Go through this list and be sure that all of your numbers/grades/strings match.
- Correct errors in admin and release to live web.
Endorsements
Similar to checking ratings, checking endorsements requires you to check the original endorsement (press release, list, etc.) against what is on our website.
Note
The "Release to Live" button must be selected to release the endorsement to the website. Only clicking "Save" will unrelease the endorsement back to admin.
First, you will check the endorsements on Vote Smart's internal website. Second, check that everybody on the original endorsement list is on our site under the correct year and that nobody is on our website who is not on the endorsement list. Once you have checked for errors, go to admin to fix all errors. Finally, release the endorsements to live. Make sure all of the endorsements have been released to live web. Occasionally one or two elections will be missed releasing to live web and will not show up. Fill in the tracking sheet with the date checked and your initials.The "Release to Live" button must be selected to release the endorsement to the website. Only clicking "Save" will unrelease the endorsement back to admin.
Checks Using SQL and Spreadsheets
Sometimes checks can be more quickly and accurately done through queries and a spreadsheet. There are queries in the network drive that are for pulling rating and endorsement data from our database with extra data not seen in admin. After you run the query, copy and paste the results into a spreadsheet to sort the database rating/endorsement to match the original. Correct any errors found in admin and then release the rating or endorsement to the live web. Make sure to check the website to confirm everything is displaying correctly. This should only be done by experienced staff and interns with staff permissionFinal Notes
Errors are bound to happen. Take your time checking to verify that our information accurately reflects the group's scores. Make sure that the right person is being rated and that they are receiving the correct score. For example, there are two 'Dan Sullivan's' in Alaska - one is a mayor and the other is a senator. Often times, the rating for the Senator is attributed to the mayor. Although an easy error to make, it is critical that we catch it. Thoroughness and accuracy are of utmost importance; we do not want to misrepresent an official, candidate, or group.If you notice that an intern/staff member is frequently repeating the same mistake, please let that person know, he/she may not be aware they are making that error.
Back to SIGs Guide
