Research
How we research
Attendly does research so district partners, charter schools and county offices of education can act on evidence, not hunches. This page is the commitment, then the steps. The counts are from our first review, on letters and texts to families.
What we will not skip
-
Every number traces to a source
If we print a figure, you can follow it to a paper, a statute or a state file. We use AI research tools to search and draft. Every number is checked against the original source before we publish.
-
Null and negative results stay in
A trial that found no effect, or that made absences worse, is part of the evidence. Leaving those out would make the review easier to read and less true.
-
We say what we do not know
Gaps, missing papers and numbers that are our own arithmetic are labeled. A clean page that hides the holes is not useful to a district.
-
This is not legal advice
Statute text is read as of the review date. Confirm current wording with counsel before you change a required notice.
The steps
The same path for every review. The notes under each step are how it worked on the family outreach review, November 2026.
- 01
Start from a real district question
We pick a question attendance staff actually have. For this review it was whether letters and texts to families reduce chronic absence, and what that means for California truancy notices, SART, SARB and Attendance Recovery.
- 02
Plan the sections
We turn that question into an outline and a checking plan. This review used seven sections, in the order a district would read them. The rule for every page: no number without a source.
- 03
Search a large research index
We search a large index of published research. For this review that was two searches, about 100 papers back. All five studies we started with were found, and about ten more were added, including 2025 and 2026 trials.
- 04
Read the original papers
We open the published paper, federal report or working paper and check every number against a quote and a link. For this review that was 236 source lookups across mailers, texting, research reviews and California law.
- 05
Get a second, independent review
We run a second, independent review on the same question, so we are not relying on one pass. For this review it screened 284 papers and included 100. The conclusion matched. Some details still had to be read in the original papers.
- 06
Cross-check the public record
We re-read federal evidence reviews in the What Works Clearinghouse, California statutes and state data. For this review that meant four federal evidence reviews, the first truancy notice law in full, and the 18.3% statewide chronic absence rate confirmed with the California Department of Education.
- 07
Correct what summaries overstated
Abstracts and memory are a starting list, not the finding. Five starter-list claims were corrected before this review was published.
- 08
Rate each study by ESSA evidence tier
A district has to know what it can cite. We rated 15 studies. The federal clearinghouse gave three of them Tier 1 and one of them Tier 3. Our own readings are labeled as ours. Eight studies were unrated. Three found no effect.
- 09
Chart the checked numbers
We draw the figures from the checked numbers, not from a summary of the summary. This review has eight charts: dot plots, bars and one trend line, with a table view under every chart.
- 10
Write for a district administrator
The review is in plain language, in seven sections, with 46 linked sources and an outreach plan in seven ordered steps.
- 11
Do a final check before publishing
We check that the charts run, that every source cited in the text is in the list, and that we look at the page the way a reader on a phone would. Then we publish.
This review, in numbers
Family Outreach Evidence Review. Only counts from the process record.
24studies in the review
46linked sources
8charts
15studies rated by ESSA tier
ESSA tiers
3Tier 1, from the federal clearinghouse
1Tier 3, from the federal clearinghouse
8unrated
3found no effect
Corrections
Five starter-list claims were corrected. These four are the ones written down in the record.
- 15% drop p = .056
- 40% better an assumption
- 12% attendance class periods, not days
- two-way trial a 45-student pilot