John Eppley asks what I make of this:

Eppley is guessing the negative spikes are searches getting swamped by holiday season shoppers.

John Eppley asks what I make of this:

Eppley is guessing the negative spikes are searches getting swamped by holiday season shoppers.

I think it’s students trying to find information for classes. It’s got a classic academic-year shape, with a big break in the summer and ramping-up shapes during the Fall and Spring semesters. The negative spike is presumably just after finals — I’m not sure why you’d need to make an argument about shoppers to explain it…

This is correct. Nearly every academic subject has the same correlation over time.

If you use the search terms “storks” and “parenting” you get almost as strong a correlation

Seems to me that the trends are highly tied to school semester dynamics, with increases near the end of both Spring and Fall terms and the large decline right at Christmas/New Year’s when there are very few students at their computers. I’d be more interested in trying to explain what I perceive to be a general decline over the time period in search frequencies… Are students getting smarter or just less interested in statistics and psychology?

I didn’t know about Google correlate. I checked the correlation between mathematics (my field) and several of the correlates. All the comparisons show the same down spikes at the same times – sharp at year’s end, there but broader at midyear

(http://www.google.com/trends/correlate/search?e=mathematics&e=understanding&t=weekly&p=us#default,40). Comparing kittens

(http://www.google.com/trends/correlate/search?e=kittens&e=ryobi&t=weekly&p=us#default,40)

with its (surprising to me) correlates shows the end of year downspikes but not the midyear ones. Do academic disciplines slow down in the summers?

I tried these terms for interesting patters possibly related to the interpretations aired here: amazon, garmin(think holiday travel), ebay, research paper, university, refworks, apa style. Seems to me the academic semester argument is consistent with many of these terms.

Does this pattern differ from the internet norm? Perhaps most word pairs show this correlation so this isn’t anything special.

People applying for college?

On the overall pattern, I oncur with other speakers. It’s students looking for information. You can even discern the Thanksgiving dip in November and the Spring Break drip in March. The downward trend may reflect the growing popularity of Google among the general population vs. early adopters: contrast with http://www.google.com/trends/correlate/search?e=sushi&t=weekly&p=us#default,20 Likewise, the trend for “mathematics” is down, with spikes around mid-April and early December, and declining activity beyond mid-April (university students); but the trend for “math” is up, with persistently high activity through the last week of May (middle-school students).

A peculiar and distinct pattern could be seen in “trigonometry”, “linear algebra”, and particularly “precalculus”, where significant peaks occur not at the end of the term, as in statistics, but in the beginning. “Precalculus” has huge peaks in late August and then smaller peaks in earlier January. Presumably, lots of high school students are googling these terms to find out what the heck those classes are and why they should take them.

With respect to the correlation between statistics and minorities, the link may be stronger when directed from minorities to statistics: “minorities” is a kind of word that would rarely be used outside the subjects of demographics and, more generally, statistics.

I think Raymond is correct, and also Andrew. This is a normal behaviour for any term, and it’s probably influenced by the holidays. See, for example, this comparison: http://www.google.com/trends/correlate/search?e=politics&t=weekly&p=us. None-election years show the same pattern we see in the initial comparison, pretty much.

I also found the downward decline in overall searches interesting. I wonder if this also an artefact of the system – maybe as overall searching grows, each individual search-word becomes less significant.

Also, look at THIS troubling trend: http://www.google.com/trends/correlate/search?e=itch&e=white+spots&t=weekly&p=us

What’s the deal with the itch epidemic of August 2007?

It looks like nobody is interested in statistics or minorities unless it’s for a term paper.