Tom Finch - Co-founder/Archivist, Manchester Coffee Archive
Aim
To investigate whether or not there is any perceptible difference in flavour between vacuum sealed coffee and coffee stored in a knotted plastic bag, where both are stored in a freezer for a period of 6 months.
Background Information
We wanted to investigate whether the low cost option of a tied plastic bag could be used as an alternative to vacuum sealing, which is currently one of the most popular methods of packaging coffee for freezer storage. Additionally, since a tied plastic bag is similar storage method to a closed reusable container (such as a plastic specimen container) it was hoped that the results may help determine whether these are also a viable option for long term freezer storage.
Method
216g of roasted coffee (Square Mile - Filter Blend) was packaged individually before placing into a -20°C freezer:
Coffee A: vacuum sealed plastic (12g in 12 bags)
Coffee B: plastic bag which was knotted, inverted and then knotted again. (12g in 6 bags)
After six months of freezer storage, six double-blinded triangle tests were prepared: a cupping was prepared with six sets of three bowls, in which two contained Coffee A and one contained Coffee B. The aim for each taster was to pick the odd one out.
Results
Taster 1 (Tom) - 4/6 Coffee B samples correctly identified
Taster 2 (Amy) - 2/6 Coffee B sampled correctly identified
Discussion
In order to interpret the results, I used the Coffee Ad Astra’s Wolfram Alpha Cupping Widget, featured on his “Statistics and Blind Tasting” blog post, to generate the N-sigma significance for each result. This was then converted to a confidence value.
| Number of Intruder Cups Identified | Number of “Failures” | N-sigma significance (via Coffee Ad Astra) | Equivalent Confidence Value | What this means… |
| 0 | 6 | 0 | 0% | No significant support for the hypothesis that Coffee A tastes different to Coffee B. Would expect up to 2/6 from chance alone (1/3 chance for each triangle test) |
| 1 | 5 | 0.1 | 8.0% | |
| 2 | 4 | 0.5 | 38.3% | |
| 3 | 3 | 1 | 68.3% | Still no statistically significant support for the hypothesis that Coffee A tastes different to Coffee B |
| 4 | 2 | 1.6 | 89.0% | |
| 5 | 1 | 2.4 | 98.4% | Supports the hypothesis that Coffee A tastes different to Coffee B, with an acceptable confidence level (>98.4%) |
| 6 | 0 | 3.2 | 99.9% |
Therefore, for the findings to be statistically significant to an acceptable confidence value (>95.4%) a taster would need to pick out 5-6 cups of Coffee B. As neither taster achieved 5/6, the results do not support the hypothesis that Coffee A and B had a perceptible difference.
While the results do not prove that there was no difference in flavour, they are nevertheless encouraging as it may indicate that a tied up plastic bag can be yield similar a preservative effect to a vacuum sealed storage when frozen. However, it is important to note that this possibility is not supported statistically by the data obtained from this experiment.
Nevertheless, given my own experience of the tasting I would personally not hesitate in using a tied plastic bag for freezing coffee for a period of up to six months as the results may indicate that any difference in flavour are minimal. Similarly I would also not expect coffee stored in a frozen reusable container, with minimal headspace, to taste significantly different to frozen, vacuum sealed coffee. After completing this experiment I was so encouraged by the results that I started using reusable containers to store single doses of coffee in my freezer and generally use this method for any coffee I am planning on using in the next few months.
After reviewing the data, the main concern I have is that we did not generate enough data to form a reliable conclusion. While it was not possible to involve more tasters due to COVID precautions, the number of cups was also concluded to be too few. After inputting further data into the Cupping Widget it was concluded that a total of eighteen triangle tests may increase the reliability of the results: a method for a further experiment has been written but is yet to be started. However, as reusable containers are most relevant to coffee storage this has been chosen as the alternative to vacuum sealing that will be tested: reusable containers are generally low cost, and have the added benefit of reducing plastic waste.
Conclusions
The results did not support the hypothesis that there is a perceptible difference in flavour between vacuum sealed coffee and coffee stored in a knotted plastic bag, where both are stored in a freezer for a period of 6 months. However, the findings may change with a larger number of tasters or with different coffees.
Further work
A method for a second experiment has been written, in which eighteen triangle tests will be completed (see here). In this experiment, reusable containers will instead be compared to vacuum sealed bags as it is hoped the results will have more real world applications.
If you would like to participate in this experiment, please take a look at the method and email us so we can share our data and collaborate on the interpretation of the findings.