Sandwich Weight Control

Considerations to be reviewed in line with the proposal to include sandwiches in the average weight marking.

It is clear from our members that weight compliance with specifications agreed by retailers is achieved when considering total net weight. The variability of the component weights is complex and as variable as the components used. The simpler products that may have a portioned (scooped or machine deposited) mix e.g. egg mayonnaise will show relatively little variability in weight and there being only one process involved can be tightly controlled.

Variability will increase with the number of components each having their own portioning variability and inevitably will affect the total product weight. For example a chicken salad sandwich the components may be chicken pieces, salad leaf, tomato slices all of which are not infinitely divisible and proportional, but in discrete pieces and this adds process variability.

Taking real data collected by a BSA member and using the Packers rules:

the actual contents of the packages should not be less, on average, than the nominal quantity;
the proportion of packages which are short of the stated quantity by a defined amount (the “tolerable negative error” or TNE) should be less than a specified level; and
no package should be short by more than twice the TNE.

The following data were collected:-

3920 individual product weights were taken over 34 different products over a period of 7 months (Sep 2009 to Mar 2010).
Average weight was 203g against a weighted specified of 204g. This would show good compliance to nominal weight (average actual weight = declared [specified weight]), and would therefore satisfy a contents control challenge.
If this population was subject to this Regulation’s control, the 2xTNE lowest limit at which no package is acceptable would be 195g.
In actual fact, 307 samples (7.8%) of this set weighed greater than 5% less than specified weight.
Taking an example of one product where 32 weights were recorded, the specified weight was 175g, the actual average weight was 182g, and the standard deviation was 8.1g.
Taking 0.5TNE, this equates to approximately 4g, the variability clearly exceeds this value.
The specified quantity would have to be increased to the greater of:
[175 – 8 (+ 2 x 8.1) = 183.2g] or [175 – (2 x 8) + (3.72 x 8.1) = 189.1g]
In this case the target weight would have to be raised to 189g.

Applying this rule to the whole set, the value would be the same; an additional 14g would have to be added to the target weight to comply with the packers rules at a declared weight of 175g. This would equate to an 8% increase in ingredient weight to meet the Regulatory control required for weight marking, if the current net weight was declared. Conversely, the declared weight could be reduced, such that the current net weight becomes the target weight (i.e. there is no change to current specifications, the declared weight is adjusted to account for process variability). This would then mean that the declared weight would be of the order of 161g.

Sandwiches are relatively light in weight, and therefore the absolute tolerances are small (in g). It is also worth noting that our members data show that If they use the TNE table, the average weight of a sandwich commonly falls in the 100-200g bracket, so 4.5% should be applied.

If a 180g sandwich has a T1 171.9g, T2 164.16g - the bread going into that sandwich already has a ± 4g per slice, so making 8g without any other ingredients even being included, so is already in the T1 weight bracket and this is not under any control of the sandwich manufacturer.

In multi-component assembly at reasonable line speeds with discrete pieces of these components, variability is too great against the legal TNE. This means that target weight will have to be increased, or declared weight will have to be reduced. This renders the whole exercise pointless, given that the first packers rule states that on average, the pack contents should be equal to the declared weight (albeit, not less than the declared weight).

Discrepancy between the weight declared and the actual weight would only serve to cause confusion when examined by enforcement authorities. The declaration of nutrition is done per 100g and per pack, enforcement officials would find interpretation of per pack nutrition difficult from the per 100g figure, since the declared weight would not be the actual weight.
Many of our smaller producers rely on small scales to check weigh a sample of products to ensure continued compliance to specifications either created for in-house control or in contractual agreement with their customers. The capital equipment costs for automatic inline check-weighers would be considerable especially where there may be a number of packing lines to be covered.

It is our belief that since a sandwich is taken as a light meal the consumer is not so concerned with the actual weight of the product (since most will pick it up and consume the whole pack) but they are more likely to be interested in the nutritional values provided and this may affect what they purchase. Therefore given the extra investment and monitoring costs associated with the proposal, it is difficult to see what justification or value would be gained from putting the weight on the pack.

Sally Higgins PhD, British Sandwich Association