Need Help Checking Weights

[TacoSauce]

I found that the Cuban Cigar Website has a nice list of average weights per "Factory Name". It is really a nice resource. Checking the data, however, suggests that some of these values might not be correct. For instance, if I plot approximate volume against reported weight there are 4 obvious outliers.  There is of course a large variation in weights across individual cigars, but these outliers are >60% off the mark. I am guessing that the reported weights for these are typos, however since I don't have access to some of the actual sticks, I cannot weigh for myself. 

If someone has some of these sticks and a scale I would be interested to know your observed weights. The expectation here is that the reported weights are all too low.

• Duke 54 × 140mm (5½″): Reported Weight: 7.14 grams
  • Partagas Serie E No.2
• Hermosos No.4 48 × 127mm (5″): Reported Weight: 5.30 grams
  • Choix Suprême, Connie 1, Picadores, Exhibicion No.4
• Petit Coronas 42 × 129mm (5⅛″): Reported Weight: 4.14 grams
  • Regalias, Mille Fleur

• Palmitas 32 × 152mm (6″) Reported Weight: 3.30 grams
  • La Gloria Cubana MdO No 4

 

[NSXCIGAR]

I believe the weights (or more accurately masses) on the list come from official HSA statements. Actual average masses for each vitola have probably never been calculated. 

[TacoSauce]

1 hour ago, NSXCIGAR said:

I believe the weights (or more accurately masses) on the list come from official HSA statements. Actual average masses for each vitola have probably never been calculated. 

Thanks, that is good to know. I think the dataset is generally a good one except for those 4 outlier points. I intend to use these values for further analyses and so am trying to fix those last 4 points.

I was able to dig up a Choix Supreme and it came in at exactly 10 grams, which is almost twice the stated mass. I can use that as my updated estimate for a Hermosos No.4, but I still need to get better estimates for the others. Worst case, I can just use the prediction from the regression, but it is always better to have at least one measurement.

[NSXCIGAR]

Did you check the ReFoHrence weight section? I'm not sure what CCW is using but here it has the CS (and all Hermosos No. 4) at 10.86g. 

[TacoSauce]

2 hours ago, NSXCIGAR said:

Did you check the ReFoHrence weight section? I'm not sure what CCW is using but here it has the CS (and all Hermosos No. 4) at 10.86g. 

Ho Ho! Yes! Great find and thank you! The fact that the actual weight data is encoded in images explains why those posts didn't show up in my earlier searches.

 

Thanks to @NSXCIGAR for pointing out the data in the ReFoHrence section. Updating the points with the highest residuals using the FoH values gives a much improved dataset.