I would also suggest ticketing for Coombe Junction is artificially ...
That's a massive problem with these data sets. The consultants put a lot of work in to apportionments, but with the best will in the world the figures remain (as they are described) as estimates. Lelant is another huge growth station, but look at the St Ives lines as a whole, and read the text, and the changes are more due to the redistribution of journeys than a fourfold change on the ground. And these shiftings of passengers around disproportionally register with the lesser station in the exchange.
As well as RickWarwicker's comments about unticketed journeys, I note that long distance journeys from Coombe often cost the same as long distance journeys from Liskeard to the same destination. Imagine I live halfway between Coombe and Liskeard. Walk down to Coombe to start my journey if it coincides with a rain calling - but can I buy my ticket on the train, and if not do I buy a Liskeard one (same price anyway) or a Coombe return even if on my return I'm not going to be connecting to Coombe. Whatever I do, I'm distorting figures ...