Tales of animals going nuts in the lead-up to an earthquake are pretty common, so it's no wonder scientists have been trying to figure out what's going on. Now an analysis of studies on the phenomenon suggests the evidence is shaky.

For the apparent wealth of research on the matter, there are surprisingly few studies that dig into the statistics behind unusual animal behaviours preceding earthquake activity.

In fact, a significant number tend to dwell on possible mechanisms rather than ask how solid that link truly is.

So a team of researchers from the GFZ German Research Centre for Geosciences gathered what analyses they could find and ran the numbers correlating observations in our pets and livestock with the scale and location of tremors.

"Many review papers on the potential of animals as earthquake precursors exist, but to the best of our knowledge, this is the first time that a statistical approach was used to evaluate the data," says hydrogeologist and lead author of the study, Heiko Woith.

The team collected 180 studies covering 729 reports of animal weirdness related to 160 earthquakes, and analysed them with respect to details on the quake's magnitude and distance, foreshock activity, and the quality of the publication itself.

Most of the reports centred on just three events; the Darfield earthquake in New Zealand in 2010, the 1984 Nagano-ken Seibu earthquake in Japan, and Italy's 2009 L'Aquila earthquake.

Overall, the studies recorded the observed actions of more than 130 species, from dogs to cows to even silkworms.

In spite of this apparent wealth of data, the researchers concluded the bank of evidence suffered from a critical limitation – all but 14 were based on a single observation in the lead up to an earthquake.

A typical report would detail an observer's reflections of what they figure to be unusual behaviours some time prior to an earthquake.

This lack of a timeline makes it hard to objectively evaluate how this 'special' action differs from a baseline of normal behaviours, leaving researchers with no way to rule out confirmation biases - such as noticing patterns that aren't really there.

With such a variety of animals covering diverse behaviours that were reported up to several months before a quake, it's impossible to sort which reports are meaningful and which are random.

Among those which might have something to add to the discussion, the researchers suggested it was likely the animals were simply more sensitive to the initial shaking.

"The animals may sense seismic waves – it could P, S or surface waves – generated by foreshocks," says Woith.

"Another option could be secondary effects triggered by the foreshocks, like changes in groundwater or release of gases from the ground which might be sensed by the animals."

None of this means we categorically rule out the chance that some sort of animal super-sense exists, and the scientists have a suggestion on how we can continue to get to the bottom of the question.

By noting the activity of a population of animals over a long period, one that extends long before and long after an earthquake, we might have better information to judge how changes relate to the tremor's properties.

"Up to now, only very few time series with animal behavior exist at all, the longest being just one year," says Woith.

So for those keen on proving Fido has a gift for predicting a quake, the key is to get their everyday behaviour on record long before that shaking begins.

Having a way to predict earthquakes well in advance simply isn't possible at the moment, so if animals know something we don't we'd love to know the secret.

This research was published in Bulletin of the Seismological Society of America.