"Data: More or Less?" by OA Science Manager Andy Rogan
Perception of our jobs, as marine biologists, varies enormously and is almost always fairly wide of the mark. The, sadly fairly inaccurate, vision tends to be some variation of endless weeks spent cruising over flat-calm and crystal clear deep blue oceans, interacting daily with a myriad of charismatic ocean life, hair going lighter and skin going darker under an infinitely azure and cloudless sky.
These days, the reality is that less and less time is being spent in the field collecting data. Field work can be incredibly expensive, and more time on the ocean means less time in the office fund-raising. Nowadays there are hundreds of groups collecting data: from big oceanographic institutions such as Woods Hole & SCRIPSS, universities and other academic groups, government organizations such as NOAA and non-profits such as Ocean Alliance. This is of course a wonderful thing. Competition inevitably leads to a more efficient, productive industry. Yet, for the most part, we’re competing for the same resources: more hands being put in the same pots for smaller sums of money.
What this means is that when the time for data collection comes along we need to take full advantage. This puts more pressure on us to collect all the data required in that short period of time, and brings to the fore a familiar question in the scientific world. Data: more or less? Collect too little and you could miss out on crucial information, a seemingly innocuous data point which forms the vital piece of the puzzle. Collect too much, and every extra piece of information becomes more time-consuming, more complex and leaves more room for error (and if you are too busy with your head down recording information, you risk not even seeing the whales in the first place!).
One of the major advantages of SnotBot is that it allows us to collect a high number of samples: a good sample size. Small sample sizes are major bottleneck to most data collection techniques which involve collecting physical, biological samples from large whales. SnotBot changes this, by allowing the researcher to race over to a whale, collect a sample/multiple samples from the same whale, race back to the research vessel, wait for the sample to be removed and appropriately stored before flying off to the next whale and repeating the process. Of course, if I had my head down recording every single variable of each flight, this would considerably slow the process. So where does the balance lie?
This was a conundrum we had to figure out, to decide what was important and what wasn’t. With well-established measures of data collection, many scientists live by the expression, ‘less is more’. Collect only the vital pieces of information, making for a more efficient and more easily analysed data set with far less room for error. SnotBot is not a well-established measure of data collection. Indeed much of the value of these early expeditions is about testing different drones and collection devices in an effort to determine the most effective and practical ways of collecting as much exhaled breath condensate, or ‘blow’, as possible. As we take these first steps into the world of SnotBot, we don’t know what the most important factors will be in shaping whether or not we get a sample and how large that sample will be. As we look to establish SnotBot as a mainstay of marine mammal research, it is imperative that we collect as much data as possible. In 5 years’ time we don’t want to look back and say, ‘if only we had collected this piece of information or that on every SnotBot flight’. We will be able to look back at mountains of data and determine what the most important factors really are. Then, and only then, will we be able to look up a little more often and get to enjoy the experience of being in the company of whales.
Having said that, it might be a stretch to say that we didn’t have many good encounters with whales…