What we asked
This summary brings together first-time-family Parenting Ladder
outcomes and returning-family Program Improvement feedback. The
instruments answer complementary questions and are summarized separately
because their identifiers have not been linked.
Parenting Ladder
Are reported parenting outcomes and gains consistent across
annual cohorts after accounting for participant
characteristics?
The analysis uses each participant’s earliest valid, unambiguous
Parenting Ladder survey in the available export. The primary sample
contains 1,256 participants, with 4 fiscal-year cohorts represented
(2022-23, 2023-24, 2024-25, 2025-26). Participants retrospectively rated
their skills BEFORE PACT and NOW at the same assessment. These are
annual cohorts, rather than repeated measurements of the same families
over time.
Main findings
- Participants reported positive BEFORE-to-NOW gains in the three
Parenting Ladder factors in every represented cohort.
- Adjusted annual NOW outcomes were broadly similar. Modest year
differences remained for knowledge, confidence, and abilities and for
reading frequency.
- Age, ethnicity, and primary language jointly contributed to
explaining outcomes. Models adding education, employment, and income
provided sensitivity checks.
- Exploratory tests did not detect changes in income- or
education-related associations across years.
- Internal consistency supported retaining the established factors,
although it does not establish measurement equivalence across years or
languages.
These findings remain preliminary pending residual-diagnostic review.
Similar estimates or nonsignificant year tests do not establish
equivalence. The data do not estimate causal program impacts or
participation dosage.
Annual reported
gains
Gains compare matched retrospective ratings on the 1–7 Ladder scale.
They represent perceived change, not change measured from a prospective
pretest.
Outcomes after
accounting for cohort composition
Core models adjust for retrospective BEFORE ratings, age, ethnicity,
and primary language. Adjusted annual estimates standardize years to the
same participant-characteristic distribution within each outcome’s model
sample.
| Access to services |
6.22 |
6.26 |
1173 |
No |
| Knowledge, confidence, and abilities |
6.36 |
6.49 |
1178 |
Yes |
| Managing parenting stress |
5.92 |
6.12 |
1165 |
No |
| Reading |
5.88 |
6.14 |
1132 |
Yes |
| Singing |
6.39 |
6.57 |
1124 |
No |
| Social support |
5.43 |
5.59 |
1171 |
No |
| Talking |
6.31 |
6.41 |
1130 |
No |
Ladder outcomes use 1–7 ratings. Reading, talking, and singing use
0–7 days per week. Year-difference classifications use joint year tests
adjusted across seven outcomes; they do not identify specific
significant pairwise contrasts.
Reliability and data
coverage
| Access to services |
NOW |
0.774 |
0.840 |
| Access to services |
BEFORE |
0.868 |
0.918 |
| Knowledge, confidence, and abilities |
NOW |
0.897 |
0.913 |
| Knowledge, confidence, and abilities |
BEFORE |
0.933 |
0.938 |
| Social support |
NOW |
0.821 |
0.898 |
| Social support |
BEFORE |
0.849 |
0.909 |
Alpha describes internal consistency using complete factor responses.
It does not measure test-retest reliability. Social-support NOW alpha
declines descriptively across cohorts; this has not been formally
tested.
Income is missing, declined, or unclassified for approximately 38–50%
of each cohort. Income-adjusted findings therefore describe a smaller
complete-case sample and are treated as sensitivity analyses. Excluded
participants differ from included participants on some observed
characteristics.
Program Improvement:
returning-family feedback
How do returning families describe their program experience
and family outcomes across annual cohorts?
The Program Improvement export contains 880 raw records and 609
distinct Record IDs. Preparation retained 872 unambiguous assessment
episodes. Primary annual summaries use the latest unambiguous assessment
per ID/year, yielding 721 annual records. These are not all unique
families across years.
| 2021-22 |
1 |
1 |
2 |
| 2022-23 |
177 |
177 |
214 |
| 2023-24 |
229 |
229 |
303 |
| 2024-25 |
170 |
170 |
199 |
| 2025-26 |
144 |
144 |
154 |
Record ID is a provisional person/family identifier. Repeated survey
records describe observed assessments; they do not establish attendance,
service intensity, or the number of times a family participated. The
earliest fiscal year has only two raw records and should be read
descriptively.
Services and
Staff
The charts retain the item groupings used in prior annual reports.
Percentages show the highest agreement category on the four-category
response scale, using each item’s valid-response denominator. Scroll
within the chart to view all items. Hover over bars or points for
estimates and sample sizes; click legend entries to show or hide fiscal
years.
Parent and Family
Outcomes
These items describe perceived family outcomes. They are not paired
BEFORE/NOW ratings. The two reporting domains are not presented as
validated latent factors.
Activities with
children
Activity ratings share the Parenting Ladder’s 0–7 days-per-week
scale. The groups have not been matched, and this summary does not
interpret their differences as effects of repeat participation.
Repeated records and
next questions
There are 178 IDs with multiple retained assessments, including 102
observed in more than one fiscal year. Person/family identifier
continuity is awaiting confirmation.
Within-person change estimates are withheld until identifier
continuity is confirmed. Confirmed IDs would support earliest-to-latest
and repeated-assessment analyses; participation exports would support
more direct dosage questions.
Preparation and
comparability
Of six same-date ID groups requiring review, four agree on numeric
outcomes but differ in open text. Their numeric responses count once;
all original text is retained for review. Two groups have conflicting
numeric outcomes and are excluded pending adjudication. The separate
technical module exports these records, item distributions, and an
all-episode sensitivity summary.
Primary annual summaries select one assessment per ID/year, so counts
can differ from prior reports that used every survey row. No new
thematic findings are asserted from the open-ended responses. No
demographic variables beyond survey language are included in this
export.
What this means for the
program
The available data support examining the recurrence of reported
parenting gains and broadly similar outcome levels across annual
cohorts. The modest remaining year differences warrant examining their
magnitude and program context. The Program Improvement results add
returning-family feedback and a larger bank of repeated assessments.
Verified identifier continuity and participation history are needed to
extend the work to within-family change and dosage.
Methods in brief
Fiscal years use July–June survey dates. Duplicate ID/date/outcome
records are collapsed; conflicting response sets on a date are excluded
pending review. Established factors use items 1–11, 12–13, and 14–15;
stress management is separate. Factor scores use available-item means,
with complete-factor and earliest/latest record sensitivity checks in
the technical analysis. Models use HC3 standard errors and
multiple-comparison adjustment. No untreated comparison group is
available. The technical RMD retains the detailed modeling and QA
documentation.
