Apologies for being so slow with this. For those who can't be arsed to wade through it all, you can just read the abstract
Abstract:
A batch of American Pale Ale was split into two secondary tanks after fermentation, and each was treated with a different late-hopping technique. One was dry-hopped, and the other had a hop infusion added according to the “Hot French Randall” technique. Blind tasting with three beer-literate but experimentally naïve participants suggested that the two techniques give noticeably different final characteristics; mean scores from this sample indicate that late-hop additions by the dry-hopping method are preferred to those made by the Hot French Randall method.
Experimental question:
With other variables held constant, how does the Hot French Randall (HFR) method of hop addition compare to dry-hopping (DH) when an equal quantity of hops are used?
Recipe (5 gallon batch, to be split into two):
7.5lb Maris Otter
0.5lb Carapils
0.5lb Crystal 60
1oz Perle @ 60
1oz Cascade @ 10
1oz Cascade @ 5
Plus 2 x 1.2oz of Cascades for hop addition
US-05 yeast
Mash @ 153F
After 2 weeks in primary at 68F, the beer was split into two identical secondaries. For the HFR batch, 1.2oz of Cascades were steeped in a French Press (1 pint capacity). This was placed in a water bath and held at 160F for 30 minutes. The resultant liquid was then added to the HFR secondary. For the DH batch, 1.2oz of Cascades were added to the DH secondary, together with a French Press full of water that had been held at 160F for 30 minutes (to control for any oxygen loss in the heated water in the HFR batch). After a week, the two batches were separately bottled with priming sugar to give 2.5 volumes of CO2.
Tasting:
Tasting sessions were conducted in a university laboratory. Three adult males with an interest in craft beer were asked to take part in the tasting session for this experiment. Participants were tested separately. All were aware that they would be tasting samples of beer, but none were aware of the hypothesis being tested, or of what beers they would be sampling. They were asked to answer any questions they were asked as fully and honestly as possible, and otherwise to note their impressions of each of the samples tasted. Although there were no obviously apparent visual differences between the two beers, opaque red plastic cups were used for the tasting to minimise the influence of any visual comparisons the tasters made. For all testing, participants did not see the samples being poured.
Phase one: triangulated testing:
Each participant was given three cups of beer (each containing approximately 150ml of liquid) and was told that two of the cups contained the same beer, whilst the third contained a different type. They were asked to indicate the two cups that they believed contained the same beer. The purpose of this testing was to see whether there was any perceivable difference between the two batches; if there was no noticeable difference between beers, then any further subjective judgements of the two beers would run the risk of simply reflecting demand characteristics or other non-interesting noise, rather than actual subjective taste differences between the two batches. For this phase of the experiment, participants A and C had 2 HFR samples and one DH sample; participant B had 2 DH samples and one HFR sample.
Two of the three participants (B and C) correctly identified the two paired samples, whilst the third (A) was incorrect in his pairing. Given the small sample size, no meaningful statistical comparison against chance is possible. However, these findings are consistent with, if not direct evidence for, the idea that the two batches are noticeably different.
At the conclusion of this part of the experiment, participants were given iced water and dry crackers to eat before the next session.
Phase two: subjective ratings:
Participants were told that they would be tasting different types of beer, and that they should write down scores for aroma, flavor, and overall impression for each beer they tried. These three categories were taken from the standardised
BJCP scoresheet for beer; the categories of “appearance” and “mouthfeel” were not used, both to keep the procedure simple, and because they were not expected to differ between beers. Again, to keep the judging procedure simple, participants were asked to give scores out of five for each category.
In order to help keep the participants naïve to the experimental hypothesis, they were told that they were going to be given four types of beer to judge. The DH and HFR batches were sampled and rated first to avoid any influence on the palate from the other beers; after this, participants rated two additional beers (one an English Bitter, one a Belgian Pale Ale). Data for these beers is not reported.
Participants A and B were given the HFR sample first, followed by the DH sample. Participant C was given the DH sample first, followed by the HFR sample.
Dry-Hopped Batch
Aroma....A:5...B:4...C:4...Total:
13
Flavor....A:3...B:4...C:3...Total:
10
Overall...A:4...B:4...C:4...Total:
12
Hot French Randall Batch
Aroma....A:5...B:3...C:3...Total:
11
Flavor....A:4...B:2...C:3...Total:
9
Overall...A:4...B:3...C:3...Total:
10
Participants' preferred beer:
A: HFR by 1pt
B: DH by 4pts
C: DH by 2pts
Mean scores for each beer:
...Aroma...Flavor...Overall
DH....4.3....3.3....4.0
HFR...3.7....3.0....3.3
Discussion:
Taking the mean averages of the two samples as an index of subjective preference, the DH batch scored more highly than the HFR batch on the three measures used. On account of the small sample size, it is not possible to judge whether this difference would be replicated in the broader population. However, it would appear from the subjective scores that the different late-hopping techniques do indeed impart noticeably different characteristics to the final beer.
An obvious limitation to this study is the small sample size. No meaningful statistical comparisons can be made without testing further participants, so the findings should be interpreted with caution. However, notes made by the brewer are consistent with the idea that the DH method yields a more pleasing final product (at least, when used with this recipe, and with these techniques). These brewer’s notes are reproduced in the Appendix. Although these are obviously subjective, a common theme which may help to explain the findings from the blind tasting is that the HFR beer appeared to have a slight vegetal note that did not sit well with the style of beer being made. In comparison, the DH beer did not have this taste. Whether this vegetal taste would be more welcome in other styles of beer is a question for further study.
A further limitation to this study is the lack of a third “control” beer - one in which no late-hop additions were made. Including this third condition would allow not just a comparison between the two methods of late-hop additions, but between each of them and the same beer with no hop additions at all. This would allow for a better index of how the extra hops in the DH and HFR beers change over time.
In conclusion: when other factors in the recipe are held consistent, these results suggest that the dry-hopping technique yields a more favourable final product. This may not, however, be because of additional perceived hoppiness; instead, it may be due to the absence of a “vegetal” taste component that the HFR technique introduces into the beer.
Appendix: Brewer’s notes
Recorded at bottling: “Both batches were pretty much fine. The HFR batch I think has a slight vegetal taste to it. When it’s chilled and carbed it’ll probably be fine, but next to the DH version I don’t think it tasted as fresh and fragrant. The DH version seemed really good – even flat and warm it’s very nice.”
Recorded 3 months after bottling: "There’s not a huge difference between these two in terms of appearance or aroma. There may be a slight aroma like hot, wet hops (rather than the more usual floral or citrus aroma) on the HFR batch, but it really could just be me imagining it as I look desperately for some difference, however slight. That said, the flavor difference is more noticeable. In the HFR batch there’s a slightly odd taste. It isn’t quite cabbagey, but it’s in that direction. In comparison, the DH batch is cleaner tasting. Both are nice and perfectly drinkable beers, but I think I’d describe the flavour in the HFR batch as a flaw rather than a complementary flavor. This component might come across better in a different style of beer, but for an American Pale Ale, I think the DH version is closer to what I’m used to tasting. I also think the DH version is more pleasant tasting than the HFR beer.
Recorded 6 months after bottling: “Both batches have aged well. They’re well carbonated and have a nice head retention. When the beers are very cold, there’s not a lot to choose between them, and both are surprisingly very clear. As the beers warm up slightly, the HFR batch reveals a slight vegetal note to the aroma, and this is also present in the flavor. In comparison, the DH version has a more familiar hop aroma, and a very clean taste with a good-to-moderate hop flavor. The vegetal flavour in the HFR batch is not unpleasant or overpowering, but neither is it something that I would choose to add to the beer. The HFR batch can stand on its own as a decent beer, and I would happily drink it. But next to the DH version (of this recipe, at least) my preference is for the DH version. It doesn’t wow me with hops, but it doesn’t have the slightly stewed off-note of the HFR version.”
).... added to fermenter the next AM.... and I think I gave it an infection by doing so. Had the makings of a filmy Brett pellicle on top. 
