The Impact of Weather on Activism and Social Movements

(Note: this article is currently being written. It is a draft! Added Campus Activism events analysis section on March 20, 2012.)

How much do social movements depend on good weather for their success?

Will the Occupy Movement be able to survive the winter? If so, when will it come back a full-strength?

To answer these questions, I examine several indicators of social movement strength over time. I focus on the United States as it generally (outside of Hawaii, some territories, and parts of the West Coast) has a strong variation in temperature between summer and winter.

I use Google Search Trends as a measure (or proxy) of movement strength. It is probably biased towards events that get more media attention - and the occurrence of those events may not always be correlated to the actual strength of social movements.

I do not know how much of this variation is due to students doing research - and is thus reflective of the academic cycle, not the strength of social movements. The best way to separate this factor is to look at non-search trends - for instance newspaper articles.

"Activism" in the United States

There is a very clear pattern of peaks that occur in the spring (April or less often March) and the fall (November). The reason that there was a 2003 peak in March instead of April was the first anniversary of the Iraq War (March 20, 2003). There is also a hard to explain general decline in searches for "activism".

"Activist" in the United States

This keyword has weaker spring and fall peaks, possibly because it is more focused on specific activist individuals and less on movements. In 2005, 2007, 2008, 2010, and 2011 there is a small dip in March. My guess is that this could be due to Spring Break - and that the number of students doing searches for research papers (or other studies) declines in March.

"Civil Disobedience" (red) versus "Activism" (blue) in the US

Similar results. Civil disobedience has a strange peak in January 2008. Could be the International Day of Action to shut down the Guantanamo prisons. Civil disobedience has a strong Fall 2011 due to the Occupy movement. I'm not sure why civil disobedience searching hasn't set a new record as I expect the number of arrests in the US in 2011 was a record (5000+), possibly the larger number of arrests since 1970 or before. It could be that most people don't think of protester arrests as civil disobedience. I haven't found a good search term for "protester arrests" that has enough traffic to show results.

"Social Movement" in the United States

Also shows a spring and fall peak. In two years the spring peak is in May.

"Protest" in the United States

Of the search terms I've examined, "protest" has the most activity - so you might expect that it would show the clearest pattern. You'd be wrong. It looks like the "protest" term is heavily influenced by specific protests that get tons of media coverage. There are massive peaks in February 2011 (due to the Wisconsin protests), and August 2004 (due to the Republican National Convention protest). The results generally agree with the spring/fall peak.

"Protesters" in the US

This is very heavily influenced by specific protests. The three largest peaks are connected to the Republican National Convention (2004), Wisconsin February 2011 protests, and the Occupation Movement (Oct-Nov 2011).

"Peace Movement" and "Environmental Movement" in the United States

These two terms show a very similar trend.

"Labor Movement" in the United States

Shows a slightly different trend from the peace and environmental movements. It tends to peak a month or two earlier in the spring (February or March), possibly due to when US labor contracts expire?

"Student Movement" (blue) and "Student Protest" (red) in the United States

Both of these terms show the familiar March/April and Oct/Nov peaks. By contrast "student activism" is a lot flatter.

Using Google Search Insights at higher resolution
What happens to the April/March peaks if we look at them in higher resolution? Eg change the chart to display several months at a time - instead of 7 years.

If you look at one year at a time. Peaks for "Activism"

Spring Peaks
2004: Mar 14-20 ties Apr 18-24 (Apr 4 avg)
2005: Apr 10-16, ties April 17-23 (Apr 16.5)
2006: Apr 23-29 (Apr 26)
2007: Apr 29-May 5 (Apr 32)
2008: Apr 27-May 3 (Apr 30)
2009: Apr 19-26 (Apr 22)
2010: Apr 11-17 (Apr 14)
2011: Apr 24-30 (Apr 27)
Average Peak Date: April 21.4

Fall Peaks
2004: Nov 14-20 (Nov 17 - aka Dec "-13")
2005: Oct 30-Nov 5 (Nov 2 - aka Dec "-28")
2006: Dec 10-16 (Dec 13)
2007: Dec 2-8 (Dec 5)
2008: Nov 30-Dec 6 (Dec 3)
2009: Dec 6-12 (Dec 9)
2010: Dec 5-10 (Dec 8)
2011 (too early to tell)
Average Peak Date: Nov 29.6

Anyone remember N30?

Cross Country Comparisons
Google Search Insights lets you pick a country or even a state in the US to look at trends.

Comparing Occupations
You could do a comparison of local Occupations to identify the impact of weather (temperature, precipitation, and possibly wind) on the movement. You could include a dummy variable to mark the start of the local protest (because the start often had a very high level of activity) and a second one for a major eviction, crackdown, or arrests. By comparing across 10 or 20 local occupations, you could identify how bad weather impacts movements - independent of external influences that affect the entire movement.

A possible external influence is how a movement would naturally grow or decline over time. For instance my research on several student anti-sweatshop sit-ins showed exponential growth over the first 10 days (none of the sit-ins lasted longer than that - at a certain point they would have started to decline).

Another possible external influence is a political event - for instance if a progressive Democrat were to challenge Obama in the Democrat 2012 primary.

Google Predicts Activism in 2012
I think it is hilarious that Google makes search trend predictions for the next year. You don't seem them in the embedded charts. But if you go to the Google Insights page then you will.

Student Power Cycle - Using Newspaper Articles
In 2003, I did a study of online newspapers at several universities over several years. I found that the spring activism peak was in April and the fall one was in November. This exactly matches my Google Search results for the search term "activism".

For student activism, there was a dip in activism due to Spring Break and Fall Break. Generally student activism grows stronger until the semester is almost over and student workload increases dramatically with term papers and other projects being due.

Student Power Cycle Essay

Summer and Winter Holidays
Holidays play a critical role in decreasing activism. In the United States (as in many countries), people tend to take holidays over the summer and a shorter one over the winter. This effect is hard to differentiate from the impact of hot and cold weather as it occurs at the same time. However, you can test this hypothesis if you look at a place with a temperate climate - if you see the same size of decline in activism in the summer or winter as one with a more extreme climate then weather is not a factor.

Relative Weather
If weather is a factor it is likely that there are two significant factors: absolute weather conditions and the difference between those and expectations. For instance, if the weather is abnormally warm say 65-70F in winter (which happens occasionally on the East Coast in DC) it provides a big boost to any planned events.

Event Data from
I look at the Campus Activism event data from January 1, 2004 to December 31, 2011. During that period there were 2146 events listed on our website for the United States. This is good sample of events with a strong bias towards student events.

Non-conferences include protests, days of action, convergences, and some speakers.

For general trends, you can see that spring is stronger than fall. There is evidence of spring break, and to a lesser extent fall break. Activity decreases at the end of school year (week 19) and at the beginning of the school year (weeks 34-36).

The Spring has two different trends. Conferences have roughly equal peaks before and after spring break. But non-conferences show a general increase through spring until the peak on week 16 (which is Apr 15-Apr 22 -- this is very close to my April 21.4 peak that I found by looking at Google Search Trends!), as warm weather encourages protesting. By contrast in the Fall the trend for non-conferences is flat.

For the Spring if you look at all events you get a different peak of Week 13 (March 25-April 1) that is three weeks earlier than the peak for non-conferences (April 15-Apr 22).

The Fall non-conference peak is October 14-27 (a two week tie) which slightly over 5 weeks sooner than the peak I found by looking at Google Search Trends (Nov 29.6). This is probably due to the Campus Activism data being more influenced by the university cycle, and less events are planned for the end of the semester.