A Systematic Review of Maintenance Measurement in Early Childhood Autism Spectrum Disorder Research.
Follow-up tests in early autism studies are a free-for-all—lock in clear labels, timing, and repeats before you start.
01Research in Context
What this study did
The team read 103 single-case studies on early autism interventions.
They looked for any follow-up tests after teaching stopped.
They tracked what authors called these tests, when they did them, and how many times.
What they found
Every study did follow-up its own way.
Some said "maintenance," others "follow-up," "generalization," or new words.
Gaps ranged from days to months, and most probes happened only once.
How this fits with other research
Bolte et al. (2013) saw the same mess in broader autism measures.
They counted 289 different tools used only once.
The new review shows the chaos never left; it just moved to follow-up tests.
Hong et al. (2018) found tablet studies claimed strong maintenance, but like Gandhi et al. (2022) note, the plans behind those claims were thin.
Both papers agree: without clear rules, "maintenance" is just a word.
Why it matters
You can’t say an effect lasts if you never check the same way twice.
Pick one label, one schedule, and repeat the probe.
Write it in your protocol so the next team can copy it.
Then your data will stack up with others and parents will know the skill really sticks.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Add a fixed maintenance label and two repeat probes to your next study plan.
02At a glance
03Original abstract
Single case research designs (SCRDs) are integral to identifying evidence-based practices (EBPs) for young children with autism spectrum disorder (ASD); however, the field lacks guidance on measuring response maintenance within SCRDs. We identified 103 studies in which researchers used SCRD to investigate the maintenance of behavioral intervention outcomes for children with ASD ages 0-5. Findings include: (a) maintenance conditions across most EBP categories, (b) limited within-case replication of maintenance assessment, (c) inconsistent use of maintenance terminology, (d) varying frequencies of maintenance assessment, and (e) wide range in latency to first and last maintenance probe. Results indicate a pressing need for the regular inclusion of maintenance conditions in behavioral research to increase our understanding of programming for and assessing maintenance.
American journal on intellectual and developmental disabilities, 2022 · doi:10.1352/1944-7558-127.4.313