Methodology

How the index is built

From counting places to grouping neighborhoods, in plain language. Six steps and a few honest caveats.

1

Finding the places

We started by mapping every place in Connecticut where children can find books: 5,191 access points in all. They come from four families of sources.

  • Out-of-school — places open to everyone: public libraries, bookstores, museums.
  • In-school — school libraries, from unstaffed shelves to full library programs with certified librarians.
  • Nonprofit — book distribution networks: Read to Grow partners, Reach Out and Read sites, book banks, Little Free Libraries.
  • Childcare — licensed childcare centers and programs, where many children spend the most formative years for early literacy.
2

Measuring accessibility for each type

A bookstore and a public library offer a very different quality of access, so each type of location gets a weight. Three criteria drive it:

  • Free or paid — can a family leave with a book at no cost?
  • Large or small selection — thousands of children's titles, or a single shelf?
  • Guided access or not — is someone there to help a child find the right book?

A public library is free, holds thousands of children's books, and staffs librarians to guide the choice: it earns the top weight, 5. A Little Free Library is free too, but it offers a couple dozen titles and no guiding hand: it counts 0.5.

Out-of-School
Public library 5
Museum 0.5 to 1
Bookstore 0.5
In-School
School library 1 to 6
Nonprofit
Read to Grow partner 2
Reach Out and Read 2
Book bank 2
Little Free Library 0.5
Childcare
Childcare center 1

Museums count 0.5, except children's museums, which count 1. Botanical gardens and zoos are not included.

Schools are weighted by staffing. A school library counts 1 if no librarian is on staff, climbing to 6 as the student-to-librarian ratio improves.

no librarian on staff
1
2
3
4
5
6
high ratio of librarians to students
3

Measure how far

For every neighborhood we measure how long it takes to reach each place, by car and by public transit (bus and rail), and a place's value fades the longer it takes to get there.

1.0 0.5 0 0 10 20 30 40 Travel time (minutes) How much it counts ~9 min: half value
The same library, two modes:
5 min 0.67
22 min 0.17

Technical: an exponential travel-time decay with β ≈ 0.08 per minute. A place about 9 minutes away counts about half as much as one next door.

4

Add it up

A neighborhood's score is the sum, over every reachable place, of its weight times how close it is. More places, and closer or higher-weighted places, both raise the score.

YOU Library w 5 RTG partner w 2 Book bank w 2 Bookstore w 0.5
PlacewCarCountsPublic transitCounts
Public library55 min3.422 min0.9
RTG partner28 min1.135 min0.1
Book bank212 min0.840 min0.1
Bookstore0.54 min0.428 min0.1
Score≈ 5.7≈ 1.2

The same neighborhood, scored two ways: strong by car, thin by public transit. That gap is what the index is built to measure.

Numbers above are illustrative. Real scoring uses the project's measured travel-time matrices.

5

Sort into easy categories (bins)

Raw scores are hard to read, so we rank every neighborhood into five equal-size groups (Very low, Low, Moderate, High, Very high), and we do it separately for car and for public transit, so each map reads clearly on its own.

By car
Very low
Low
Moderate
High
Very high
By public transit
Very low
Low
Moderate
High
Very high
No public transit access

Public transit has an extra "No public transit access" group for neighborhoods nothing reaches by bus or rail. Unpopulated areas (no resident children) are set aside entirely.

An honest caveat: because the two modes are binned separately, a public transit "High" is high for public transit. In absolute terms a car still reaches far more, almost everywhere.

One more caveat: because groups are ranked against the current statewide data, a neighborhood's group can shift when other neighborhoods change, not just when it does. These groups compare places to each other today. They are not built to track improvement over time. Tracking change is a separate exercise, done on the underlying scores.

6

Does reaching books depend on a car?

The index values the whole set of book sources a neighborhood can reach, weighted by type and faded by travel time. It is not the trip to the nearest book. So we read each neighborhood in two steps.

Every neighborhood
Q1 · Is the reachable set meaningful at all?measured by the car score, the best case
No, the set is thin
Car very low → [6] Very few book sources by any mode
Car low → [5] Few book sources by any mode
Yes, the set is meaningful
Q2 · How strong is public transit on its own terms?public transit bin, against public transit across Connecticut
Top fifth → [1] Among CT's best public transit access
Fourth fifth → [2] Good public transit access, for CT
Middle fifth → [3] Some public transit access
Bottom two fifths or no public transit → [4] Limited, mostly needs a car

Where the reachable set is thin to begin with, public transit cannot widen it. Where the set is meaningful, public transit's own statewide quintile decides how much a car-free family can reach.

An honest anchor. Public transit tiers are relative to public transit across Connecticut, which is modest statewide. Even strong public transit access reaches a narrower set of book sources than a car does. The point is not that public transit matches a car. It is that some neighborhoods give a car-free family a real option, and others leave them with almost none.

A suburb on a rail line with local bus reads as strong public transit access. A dense town center reads good to strong, depending on routes. A large-lot exurb with roads but few buses reads limited, mostly needs a car. A remote rural town reads as few or very few sources to begin with.

For district views we average each neighborhood's score by where children live, then route the district through the same two-question tree. Averaging can hide contrast inside large districts. A district holding both a public transit-served center and car-dependent outskirts will read as middling, even though few of its neighborhoods are actually middling. The neighborhood view is the honest grain; the district view is the summary.