Here is the reason why you need to build libraries

Conclusion: The further people have to travel to a library, the less likely they are to use it.

Statistic: Negative correlation between library catchment area and library visits.

Pop culture reference: “If you build it, he will come.” (Field of Dreams, 1989)

Now you know the punchline, let’s go back to the beginning and understand how we got here.

The ever-inquisitive ALIA President Jane Cowell asked me a question the other day about library use and how people in some areas seem to use libraries more than others. It got me thinking about analysis I did a few years back which explored issues like average travel time to the nearest library, average population catchments per branch and correlations between library visits and other library metrics (see ANALibrarYSE references at the end of this piece).

It also struck a nerve in relation to a just-finished project where we identified a couple of suburbs in an otherwise high library use area where library membership and use was distinctly low. Why? All of the libraries were at the other end of the LGA (Local Government Area), so these residents simply travelled to a nearby library in a neighbouring Council area and signed up there. Proximity matters.

Anyway, I wondered if there was a link between the density of libraries in an LGA and the level of library use.

For starters, we know that there are more libraries per square km (sq km) in inner city LGAs than there are in outer metropolitan areas (mainly due to historical reasons). For example, the City of Port Phillip in Melbourne (think St Kilda) has 5 libraries in an LGA of 20.6 sq km. Yes, every Port Phillip resident lives within 2km of a library. Out in the City of Hume on Melbourne’s fast-growing northern fringe there are 5 libraries in an area of 504 sq km. If Hume had the same library density as Port Phillip (1 library every 4.1 sq km, not 1 every 101 sq km) it would have 122 libraries! Bonkers!

But surely – I hear you say – Port Phillip also has higher population density than Hume? Of course it does – 5,300 pop per sq km vs only 522 per sq km – but here we’re looking at a 10:1 ratio whereas the library density has a ratio of 25:1. Which means that Port Phillip has one library for roughly every 20,000 residents while Hume has one for every 50,000 residents. This sort of explains why three of Hume’s libraries are very large (two are very small) and why Hume is looking at building libraries and alternative service models in the new growth areas.

So I pulled together the following data for every LGA in metropolitan Sydney (2022-23) and Melbourne (2023-24):

Total population
Area sq km
Number of library branches
Number of library members
Number of library visits
Number of loans
Number of attendances at library programs.

I calculated 5 ratios (which take the population variation out of the equation):

Library density (sq km per library branch)
Library membership per capita (on average around 0.30-0.35 or 30-35% of the population)
Library visits per capita (approx. 3.0 to 3.5)
Library loans per capita (approx. 4.5 to 5.5)
Program attendance per capita (approx. 0.20 to 0.25).

And putting my big-boy statistical hat on I got Excel to do some simple correlations. Which for those who are long past their Year 10 maths is a statistical technique that asks if there is a numerical relationship between two sets of numbers. That is, if one number goes up:

Does the other number also go up? (highly correlated – like wealth and ownership of fancy cars)
Does it go down? (negatively correlated – like summer temperatures and eating belly-warming casseroles)
Are the two sets of numbers simply not related to one another? (zero correlation – like shoe size and blood pressure).

Here are the results, remembering that perfect positive correlation = 1.0 and perfect negative correlation = -1.0.

Without going into all of the detail of hypothesis testing, measures of significance and t-tables here is the bottom line.

A negative correlation of around -0.35 indicates a significant negative relationship between the two variables. That is, the larger the catchment area for a single branch library, the lower the number of visits or loans per capita. Or in simple terms, the further people have to travel to a library, the less likely they are to use it.

Now in the interests of full disclosure I should note the following.

Correlation does not imply Causality. That is, just because two numbers are correlated does not mean that one thing caused the other. It says that the two numbers move in predictable ways (e.g. both go up together, or in this case one goes up when the other goes down).
It might just be that the type of people who like to use a library live in inner city areas where the libraries are closet to their home. I probably won’t buy that for the thousands of young families that are more likely to be found in outer urban areas, or the older readers who live in established LGAs that are neither inner or outer metropolitan. However, I will acknowledge that people in inner city areas live in smaller dwellings so the desire to find a ‘third place’ to exist outside the home is strong.
Not all of the number in the table above point to the same conclusion. Notably, two of them (membership of Sydney libraries and program participation in Melbourne libraries) are not statistically significant, being so close to 0 (zero correlation) that we can’t say there is any positive or negative relationship between library density and use. However, with six correlations at or below -0.32 I’m prepared to say that there is something real happening here.

So what does all of this mean?

Well, it doesn’t mean you maximise library use by building a library on every street corner. But it does suggest that if there is a large area in a metropolitan LGA where there is NO library you should not be surprised if people in that area have much lower levels of library use. Going to the library is a discretionary act. If you make it difficult for people to get there because the nearest library is too far away, they will just not come. In the same way that anyone who manages a new-build or newly-refurbished library will tell you that this results in a boost to library use, the opposite is true. Too far to go, no public transport or parking, too small or unappealing leads to lower use.

Most importantly, this data suggests that if you’re in an outer metro LGA where houses are popping up in once far flung rural areas you better start planning to build a library for the new community. Because if someone thinks the new residents will travel to the library in an established area 20 minutes away – tell him he’s dreamin’.

Here is the reason why you need to build libraries

Emerging library service models put the customer in control