Leaving Footprints in the Data - Large Database Validation

[Mnts2C]

Hello,

Currently working as an audit manager in a shipyard. We have a large database where deficiencies are entered from various sources, over 60,000 entries per year. Seems to be a normal, large database that allows a variety of analyses.

One issue under discussion is the value of entering data where something occurred and had no deficiencies. Roughly, the database records deficiencies as level 4 = no deficiency, level 3 = minor, level 2 = major, and level 1 = critical.

It takes effort to ensure the accuracy and validity of the database. One opinion is that the effort to validate level 4 entries is non-value added. Another opinion is that the level 4s provide a relation to deficiencies in similar areas. (ex: two level 2s for the year = better take action now. But two level 2s in relation to 48 no deficiencies = the process seems manageable.) Recording the level 4s can be referred to leaving footprints.

Was hoping some Covers would share some experience on the issue. Thoughts ??

Thanks in advance !!

 

[Miner]

A lot depends on what you intend on doing with the data. What you currently have is sufficient for analyses (e.g., Pareto) of the audit deficiencies themselves. However, you limit other analyses that depend on the number of level 4 audits, such as the percentage of discrepancies.

In other words, you have evidence of things done wrong, but no evidence of things done right (i.e., effectiveness).

 

[Ninja]

Another way to look at this is "negative record".

Being a company making stuff or doing stuff, I presume there is a record already of how many of each thing was done...either in the work logs, or Invoice line items, or completed job tickets, or something...

If 500,000 "jobs" were done, and 60,000 non-4 records were created, it may be good enough to assume that the lack of record for the other 440,000 are 4's.
Not a positive record of no defect...a "negative record" or no defect...and that approach while a bit ethereal may be the most value for what you're trying to analyze.

Positive records are more compelling...but positive records cost money to create.

 

[Mnts2C]

Thank you both.

Ninja is highlighting one facet of the discussion where data, no matter if the data is negative or positive, there is a cost associated with the data.

Is the cost of negative data or footprints worth the effort or ROI?

Good input, thanks.

 

[Ninja]

FWIW, the type of negative record (meaning "lack of data" being used to guide a decision) is free....
It may, however, be worth as much as it cost.

If you do nothing, you have your negative record which you can use to presume something. Doing nothing is free in this context.
{maybe not reliable...but free none-the-less}

 

[Sidney Vianna]

hoping some Covers would share some experience on the issue. Thoughts!!

If the organization truly embraces the preventive action approach and has the resources to timely analyze the events level 4 (near misses/incidents), it should benefit from this data collection and analysis. The loss causation pyramid model shows that.

But, if level 4 events are not being analyzed, then, I agree, it is a wasted effort.

 

[Mark Meer]

I agree with previous replies: no analysis or purpose, no point.

It takes effort to ensure the accuracy and validity of the database. One opinion is that the effort to validate level 4 entries is non-value added.

That being said, IF you do have some marginal use for it (like the example you give), perhaps it is possible to decrease the requirements for entries of this level?

In other words, why does data validation take so much effort? For the actual deficiencies, presumably it is because there is a bunch of required data to be entered. ...but for level 4, perhaps most of this data is unnecessary (because you don't have to detail a deficiency, evaluate it, take actions, etc...).

If your system allows for data requirements for this level to be scaled back, you could potentially keep logging them, without the validation overhead.

MM

 

[Mnts2C]

Appreciate the feedback. It is helping our discussion at work.

Thanks again to all !!