Civil Service Update
By: Francis D. Kelly, Secretary
On February 21, 2003, the Civil Service Commission voted to adopt “Statistically Valid Grouping” as the new Rule replacing the Rule of Three Scores for Certification of Eligibles for promotion applicable to the uniformed ranks of the San Francisco Fire Department.

The vote was 4-to-1 to adopt the new Rule. However, since the initial vote, one of the Commissioners has recanted. The Rule will now be revisited sometime after this writing. We will attempt to keep you posted at the April Union meeting. Here is verbatim language sent to us by Civil Service explaining the new rules:

“...Important points in Rule 313 - Certification of Eligibles are listed below:”

“...The Rule on the Certification of Eligibles for promotions in the San Francisco Fire Department is the Statistically Valid Grouping also known as the ‘Sliding Band.’ Statistically Valid Grouping is based on an established and court-tested statistical formula to determine that eligibles have comparable knowledge, skills and abilities and the difference in their scores is not significant.”

“...Those eligibles in the grouping are referred to the Fire Chief for consideration for selection. The Fire Chief may consider the scores in the grouping as equally qualified for selection. Additional eligibles may only be considered when the top score is exhausted.”
“...In the selection process, the proposed Civil Service Commission Rules require the Fire Chief to develop secondary criteria that are:

• Job related,

• Based on experience, training and employment history, and

• Not based on relationship, race, religion, gender, national origin, ethnicity, age, disability, gender identity, political affiliation, sexual orientation, ancestry, marital status, color, medical condition, or non-merit factors or otherwise prohibited nepotism or favoritism.”

Implementing the new Civil Service Rule of Statistically Valid Grouping to the 1994 Captain’s Exam is quite revealing. With this new formula, the Band for that particular exam would be 181 points. The top score for the 1994 Captain’s Exam was 900 points out of 1000 point test. The Band therefore would mean that the Candidate who scored 719 (or finished 43rd) is equal to the Candidate who finished first with a score of 900. According to Civil Service these candidates “would have comparable knowledge, skills and abilities and a difference in their scores is not significant.”

A difference of 181 points is not significant? Is there no difference between a B+ and a D-? GOD! I wish my father was alive today so I could tell him “...hey Dad, I really was a B+ student!”

Theoretically, Civil Service argues that the candidate who scored 900 may have had a good day and therefore 900 is not the candidate’s true score; he/she is comparable to the candidate that scored 719. What Civil Services doesn’t theorize is that the candidate who scored 719 also may have had a good day, and therefore that candidate’s true score of 538 is not comparable. One could refer to the new rule as statistical voodoo rather than statistically valid.

Careful attention must be paid to Civil Service’ “bureaucratic ESE”. We need to scrutinize every sentence of their explanation for Statistically Valid Grouping. Consider the last sentence of Paragraph No. 2 in the aforementioned explanation.

“...additional eligibles may only be considered when the top score is exhausted.”

“May only” is the key phrase of the previous sentence. The appearance is that Civil Service is protecting the candidate who applied themselves and placed higher on the list than other candidates. Let’s revisit the 1994 Captain’s Exam and consider the impact to the List when the top score is exhausted or the first candidate is promoted. The difference in the scores between the first two candidates was 44 points; therefore the Band would slide to 675 points on the list, picking up an additional 28 candidates to add to the original 43. Now there would be a pool of 70 candidates available for the second promotional position. Our letter grade has now moved from a “B” to an “F”. The second promotional position may be filled by a lieutenant who actually failed the exam.

It’s extremely difficult for me to fathom the concept of statistically valid grouping. Is there a clandestine plot to return to the “Spoils System”? Nah, it’s outright blatant! Political appointments will rule the day. It would be very difficult to identify favoritism or to prevent nepotism. Parameters within the Civil Service Rules already permit the appointing officer to use broad discretion in the selection process. Now Civil Service has made it virtually impossible for truth and fairness to prevail.

Local 798 has no alternative but to seek arbitration. I hope we are successful, because if not, Statistically Valid Grouping will have a devastatingly negative impact on the San Francisco Fire Department for many years to come.


Main Menu