Impact of 20 Hours ECE on Playcentres 2008

3 - Change in participation

Six sets of figures are available to determine the extent that attendance at Playcentres may have changed due to them not being included in the 20 Hours ECE scheme.

First, Playcentres were asked for the number of families attending their service just before implementation of 20 Hours ECE (29 June 2007) and four months later (31 October 2007).  If 20 Hours ECE has resulted in significant volume change, it should be shown over this period, although there may have been some volume change before (in anticipation of 20 Hours ECE) and after the period.  (See sub-section 3.1)

Second, Playcentres were asked whether their number of children had increased, decreased or stayed the same since 1 July.  This was in regard to under-three year olds and children aged three years or over.  (See sub-section 3.2)

Third, in regard to the number of under-three year olds and children aged three years or over who had left, services were asked the reason for their departure, with one of the reasons being that parents had moved them to a service providing 20 Hours ECE.  (See sub-section 3.3)

Fourth, Playcentres were asked how they thought 20 Hours ECE had affected them overall, with one of the responses being that children had left.  (See sub-section 3.4)

Fifth, Playcentres were asked for the number of under-three year olds and children aged three years or over on their regular roll on 31 October 2007.  This can be compared with roll data collected each year through the RS61 process.  The advantage of these data is that they put recent changes into the context of the long-term trend, as does the next set of data.  (See sub-section 3.5)

Sixth, data are available through the funding system, both on funded-child-hours and what these are as a percentage of the maximum number of funded-child-hours Playcentres are entitled to.  This not only gives long-term trend information, it also reflects both enrolment numbers and the average time children attend for.  (See sub-section 3.6)

Each set of figures tends to support the likelihood of participation having fallen after the implementation of 20 Hours ECE, most probably because of it.  However, the extent that participation has changed varies across the data sources.  Each set of figures is now presented and discussed separately, with the overall results being summed up in sub-section 3.7.

3.1 Change in the number of families

Of the 157 respondents, 153 gave information on the number of families for both 29 June and 31 October 2007.

These figures show the number of families to have fallen 2.8 percent over the period.  However, Playcentres were distributed widely around this average.  While a large proportion showed fewer families (45.8 percent), almost as large a proportion showed an increase (41.8 percent).  Figure 1 shows the distribution around the average, both for the percent change and the change in the number of families.

Figure 1: June-October change in number of families attending surveyed Playcentres – percentage change and change in number of families

Three factors contributed to the overall change of -2.8 percent, i.e. the fact that it is negative. 

• More services experienced a fall than a gain in family numbers
• Services losing families lost a higher percentage of families on average (17.6 percent) than the percentage gained by services gaining families (16.8 percent)
  
• Services losing families tended to be larger (they were made up of an average of 27.3 families in June 2007) than those gaining families (21.6 families).
  

Despite the net loss of families for Playcentres as a whole, the wide distribution of services shows a wide range of experiences.  The average change cannot be said to describe the typical Playcentre experience, as there does not appear to be a typical experience.  The wide distribution also lessens the statistical significance of the change.

3.2 Playcentres’ views on whether their rolls have changed

Of the 157 respondents, 153 stated whether their under-three roll had changed and 151 stated whether their three-and-over roll had changed since 20 Hours ECE was introduced.  Considerably more respondents said that their under-three roll had risen, with 41 percent more responding that it had increased than responding that it had decreased.  The opposite was the case for the three-and-over roll, with a 15 percent more services saying it had fallen than said it had risen.  See Figure 2.

Figure 2: Percent of Playcentres by whether they thought their rolls had fallen, risen or remained the same since 1 July 2007, by age group

These figures suggest that 20 Hours ECE is likely to have had a significant impact on rolls, given that only children aged three years and over are eligible.  The change for the three-and-over roll stands in marked contrast to the majority of services saying that their under-three roll has risen.

As an aside, the response about the under-threes roll is curious.  It stands in marked contrast to some other figures and is further discussed in sub-section 3.5.

3.3 Playcentres’ views on why children left

Section 4 examines the reasons given to explain why children had left.  One of the response options was “Parent(s) moved them to a service providing 20 Hours ECE”.  In the case of under-three year old children, 24 percent of services made this response, with the children involved being 2.3 percent of the under-three year old roll at the end of June 2007.  In the case of children aged three years of over, 45 percent of services made this response, with the children involved being 9.3 percent of that age group’s roll at the end of June 2007.  As is discussed in that section, the actual figures may be somewhat lower for two reasons.  First, it appears likely that some services may have been referring to the destination of the child rather than the reason for leaving.  Second, some of these children may have left anyway for other reasons.

3.4 Playcentres’ views on whether 20 Hours ECE has caused children to leave

One question put to Playcentres was, “Overall, how do you think 20 Hours ECE has affected your Playcentre?” with one of the response options being “Children have left”.  Twenty nine percent of Playcentres ticked this response, although no indication is given of the number or percent of children that have left.  (See Section 6)

3.5 Change in the number of enrolments

All of the 157 respondents provided data on the number of children on their regular roll on 31 October 2007.  Annual enrolment data were obtained from the RS61 for the years 2002 to 2007 for all but three of these Playcentres, with the three exceptions not being open the full period and hence excluded from the following analysis.

Figure 3 shows the number of enrolments each period for the 154 Playcentres analysed.  Figure 4 presents the same data as an annual percentage change (or four-month change in the case of the figures for 31 October 2007).  The graphs show a post-implementation fall in the number of enrolments for both age groups.  The fall was 4.8 percent for under-three year olds, 4.6 percent for three year olds and over, and 4.7 percent in total.

Figure 3: number of enrolments in surveyed Playcentres, by age

Figure 4: annual percent change in number of enrolments in surveyed Playcentres, by age

(Note, change to 31 October 2007 is over four months, not the year)

It is important to bear three factors in mind when considering this recent change in roll numbers.

• First, the length of the period being considered.  The interval between the last two periods is four months, which is a third of the interval between each of the previous consecutive time-points.  It is quite possible that the roll-drop may be even greater once a full year’s of data are available.
• Second, seasonality.  The last time-point shown in the graphs is for the end of October, which is different from all the other time-points shown, which are for the end of June.  It is quite possible that the fall in roll numbers between June 2007 and October 2007 merely reflects the fact that it is a different month of the year, i.e. that it is seasonal.  Indeed, analysis of funded-child-hours indicates fewer hours in October than in June (see Figure 9 in sub-section 3.6 below), although this may be the result of non-attendance due to winter illnesses and other reasons rather than a fall in the regular roll.
• Third, the long-term trend.  Recent changes need to be put into the context of the long-term trend, which has been downwards, especially for the roll of children aged three years or over.  It is also important to consider the extent that the long-term figures fluctuate over time, as the more they fluctuate, the less certainty there is that recent changes are not part of this normal volatility in the figures.
  

All three factors lessen the certainty of any conclusions based on these figures alone.  This is why several data sources and several ways of analysing the data are considered in this paper.  Availability of the 2008 RS61 data later in the year will address the problems raised in the first two of these points.

Another way of assessing the change is to examine the trend in the percentage of services with an increase in enrolments and those with a decrease, as is shown in Figure 5.  This shows a break in trend, with the proportion of services with a fall in the last four months (55.2 percent) being higher than for any other period and occurring after four years where it was relatively stable.  The break in trend is more easily seen in Figure 6, which subtracts the proportion of services with an enrolment drop from the proportion with an enrolment gain.

Figure 5: percent of Playcentres surveyed, by whether enrolments rose, fell or stayed the same over the year

Figure 6: percent of Playcentres increasing enrolments less the percent with falling enrolments

In terms of the overall change in enrolments, these four graphs support the findings of the previous sub-sections, which showed (1) a fall in the number of families and (2) more services saying that their three-and-over roll had fallen.  They do not however support the view of services that the under-three roll has grown.

As was mentioned before, the fact that the majority of services thought their under-three roll had risen was curious.  Not only does it contradict the enrolment numbers they gave, it also contradicts the fact that slightly more of them think that family numbers have declined rather than risen.  Having said this, however, it is the case that the enrolment figures show a higher net proportion of Playcentres with a fall in their three-and-over roll than in their under-three roll, but the difference is not nearly as large as that suggested in sub-section 3.2.  See Figure 7.

Figure 7: comparison of Playcentres’ impression of their enrolment change and what their enrolment numbers show, by age group (percent of services with gain, fall or no change)

Another way of examining the roll-change figures is to look at how changes vary across Playcentres.  The extent of the variation around the average is shown by the two graphs in Figure 8, which distribute the number of Playcentres by their percentage change in roll numbers.  The graph on the top shows the change in the 12 months to the end of June 2007 and the one on the bottom shows the change over the four months to the end of October 2007.

Figure 8: roll change at surveyed Playcentres – 12 month change to June 2007 and 4 month change to 31 October 2007  (percentage change in roll number)

The two graphs show there to be a wide spread across Playcentres, distributed around a mode of little change.  The distribution in the bottom graph is tighter (i.e. services tend to be closer to the centre) than the top graph, probably reflecting the fact that only four months’ change was considered and fewer services would have experienced change over such a shorter period.  While it is difficult to see, the bottom graphs do show a slight leftwards movement compared with the top graphs, suggesting a greater proportion of services experiencing a drop in roll numbers.

3.6 Funded child hours

The Ministry collects daily information on the number of funded-child-hours, which is the number of hours the Ministry funds children attending the services.  These figures reflect both the number of children and the average number of hours they attended.

Unfortunately, the funded-child-hours figures are very volatile over time, reflecting both seasonal and other factors such as some holidays falling on different months in different years.  This makes analysis difficult.  Two graphs are given showing the long-term trend.  Figure 9 shows the monthly number and the 12-month moving average.  Figure 10 shows the annual (month-on-month) percentage change.¹   In both cases, a vertical line is drawn to show when 20 Hours ECE was implemented.  Both graphs reveal that funded-child-hours have fallen since the introduction of 20 Hours ECE, but that this continues a downwards trend that may have been present since early 2007 if not before.

Figure 9: Number of Playcentres’ funded-child-hours per month and 12-month moving average

Figure 10: Annual (month-on-month) percentage change in number of Playcentres’ funded-child-hours

Another way of examining the funded-child-hours, which is less affected by volatility, is to divide them by the maximum number of funded-child-hours that these services are licensed for and can be paid.  The annual (month-on-month) percentage change of these figures is shown in Figure 11.  This gives a smoother line than that in Figure 10.  It suggests that growth rates began falling in late 2006, becoming negative in early 2007 and significantly more negative after the implementation of 20 Hours ECE.  Assuming there was no change in the maximum number of hours services were licensed for, Figure 11 shows that funded-child-hours fell at an annual rate of between one and two percent after the implementation of 20 Hours ECE.²

Figure 11: Annual (month-on-month) percentage change in number of Playcentres’ funded-child-hours divided by their maximum number of funded-child-hours

3.7 Change in participation – discussion and conclusion

The different data sources all suggest a reduction in participation since the introduction of 20 Hours ECE.

The extent that participation may have fallen can be examined from two perspectives: the proportion of services that are likely to have lost children due to 20 Hours ECE; and the proportion of children that are likely to have left due to 20 Hours ECE.

In regard to the proportion of services with a reduction in the number of children since the introduction of 20 Hours ECE, Figure 12 summarises results from the different data sources.

Figure 12: percent of Playcentres with a volume fall following implementation of 20 Hours ECE – various measures compared

Figure 12 shows that, while most of the data sources indicate that more services had a fall rather than rise in participation, the estimates vary considerably.  The results can be grouped under three headings: the perceptions of services about the impact of 20 Hours ECE on their service; the perceptions of services on the extent of their roll change; and the actual change in roll and other measures of volume.

• The first three measures shown in Figure 12 indicate the perceptions of Playcentres on whether 20 Hours ECE has resulted in children leaving their service.  The best of these measures is the first, which shows that 29 percent of services thought that 20 Hours ECE had resulted in the departure of children.  As was discussed previously, the second two measures may exaggerate the number.  They show the proportion that gave the response, “Parent(s) moved them to a service providing 20 Hours ECE” as the reason that children had left since 1 July 2007, but some services are likely to have mistakenly viewed this response as referring to the destination of the child rather than the reason they left.
• The next two measures show Playcentres’ perceptions of how their roll has changed since the introduction of 20 Hours ECE, with the net figure shown being the proportion of services who thought rolls had dropped minus the proportion who thought rolls had increased.  Fifteen percent more services thought their three-and-over roll had decreased than those who thought it had increased, which is fairly similar to the proportion shown by the actual roll numbers.  It is not surprising that the figure is lower than the previous figures, as services with children leaving because of 20 Hours ECE may still have experienced roll growth (or no change) for other reasons.  However, as has been previously mentioned, the very high proportion of services that thought their under-three roll had risen does not fit well with the other figures.
• The last five measures use the actual volume changes to show the net proportion of services with a volume fall.  The first three of these relate to roll numbers, the fourth to family numbers, and the fifth to funded-child-hours (as expressed by annual change in occupancy rates, i.e. change in funded-child-hours / maximum-funded-child-hours over the year to October).  A net 18 percent of services experienced a fall in their roll in the four months since the introduction of 20 Hours ECE, with the figure for three-and-over children (13 percent) not being very different from services’ own perception.  The figure for funded-child-hours (8.5 percent) is lower, although it is for a somewhat different time period and not strictly comparable.  The figure for the number of families (3.9 percent) is lower still, which indicates that some families may choose to keep their younger children in Playcentres but send their older children to services offering 20 Hours ECE.
  

In regard to the actual change in participation since 1 July 2007, Figure 13 summarises results from the different data sources.

Figure 13: percent fall in volume following implementation of 20 Hours ECE – various measures compared

Figure 13 shows that the fall in roll since the introduction of 20 Hours ECE was around five percent, and the fall in family numbers just over half that.  The fall in occupancy rate, which reflects the number of funded-child-hours, is slightly less again, but is for a somewhat different period of time.

In interpreting these volume change figures, it is important to be mindful of several factors.

• Seasonality.  Four of the five measures in Figure 13 compare different months of the year, which may vary purely for reasons of seasonality.  Figures are not available to test the impact of seasonality in regard to regular roll and family numbers, but occupancy rates have a definite seasonal cycle.  Occupancy rate figures using funded-child-hours show that participation tends to be lower in the July-October period than in the period before; while the median July-October 2007 occupancy rates was down 1.9 percent on the same period the previous year, they were down 2.6 percent on the four months ended June 2007.  However, it is possible that the seasonality in funded-child-hours is not reflected in the regular roll figures and family numbers; e.g. they may be purely due to winter illnesses or just non-attendance.
• Long-term trend.  Recent change needs to be put into the context of the long-term volume trend, which appears to be slightly downwards.  The regular Playcentre roll in June 2007 was one percent down on the previous year and five percent down on its high four years before.  Growth in occupancy rates began falling late 2006 and became negative early in 2007.  On the other hand, given that the post implementation period is only four months (a third of a year), the impact of this factor is not very great.
• However, there is another context provided by the long-term trend, and that is its volatility.  The more volatile the figures are over time, the less certainty there is that recent changes mark a break from trend.  While many of the above graphs indicate considerable volatility, not all do, and all give a consistent result of a fall in volume since 1 July 2007.
• Wide distribution of services.  Services differ widely in both the direction and extent of their volume change, which lessens the statistical significance of any result.  However, despite this factor, the recent change does appear to be a distinct break from the long-term trend line and is consistent across most measures.
  

It is also important to note the comment of several services that the downwards trend appears to have continued into 2008.  Furthermore, some report that there is less interest from potential new families and their waiting lists are down.  If this is the case then, when the June 2008 RS61 figures are analysed, they may show the impact to be even greater than what has been shown here.

In summary, 29 percent of services reported that 20 Hours ECE had resulted in children leaving, but the net proportion experiencing a volume fall in the four months since 1 July was less, at 18 percent.  Rolls fell around five percent in those four months.  The fall in family numbers has been somewhat less, indicating that some families may be choosing to keep their younger children at Playcentres while sending their older children to other services.  While these figures indicate a negative impact on volume, the changes vary considerably across individual Playcentres, suggesting that there is no typical Playcentre experience.

 

Footnotes

1. The January holiday months are excluded as they distort the totals.
2. A cursory examination of the maximum-funded-child-hours trend shows that, if anything, they have fallen slightly, which would suggest that the change fall in funded child hours may be even greater than the change in occupancy rates.