Overall attainment in the first three years of primary in Aberdeen is continuing to rise but pupils with the biggest advantages when they start school continue to stretch ahead.
Gaps between children on free meals and their more advantaged classmates remain significant, according to the latest findings from the Performance Indicators in Primary Schools (PIPS) tests.
The city was the first in Scotland to use the standardised Durham University tests and now has data on two full cohorts of pupils who went through P1-P3 between 1999-2002 and 2000-2003.
Anne Horgan, principal officer for education statistics, says the evidence from the two cohorts shows that differences in reading and mathematics are significant at the start of P1, grow by the end of the year and widen by P3.
Disadvantaged pupils still do better than they used to but early intervention strategies tend to favour those who are better placed to take advantage, a finding confirmed by national research studies on intervention.
Mrs Horgan, however, said there was optimism that the current P3, the third group, are closing the gap. Disadvantaged children in this latest cohort are not falling as far behind, while P1 children in the past two years have already closed the gap.
"This is a very positive indication that social inclusion policies are beginning to have an effect on attainment in the early years of primary education in Aberdeen," she says.
Findings on the two previous groups show little variation in performance over the three years, although one cohort did better in the first year of schooling while the other made more progress over the three years.
Differences between boys and girls remain "very significant", Mrs Horgan reports. "The girls' advantage seems to be getting wider each year in reading. They are actually starting with an advantage in P1 and by P3, the gap is wider, but it is not every school."
Mrs Horgan says that other evidence from 5-14 tests and National Qualification results confirms that early differences in performance are never narrowed.
In maths at the start of P1 in 2000, boys and girls had the same average scores. Boys made faster progress during P1 and were ahead by the end of the year. But girls made faster progress in P2 and P3 and finished P3 at the same average level as boys.
Meanwhile, a straight comparison between pupils at Walker Road primary shows that those taking partial immersion French in P1 are much better readers.
In 2000, one P1 class took up French and the other did not; both started with similar average reading scores. The French learners made faster early progress and had opened a gap by the end of their first year. By the end of P3, they maintained that difference, mostly due to the advances in P1.
There is no significant difference in maths scores by the end of P3.