You are on page 1of 22

Steamboat Springs School District – Enrollment Forecast

Developed by Jim Looney, Square Cubed Consulting, Ltd. 10/29/16

Scope

Square Cubed Consulting has been hired to develop a five-year K12 forecast for the
Steamboat Springs School District (SSSD). To do this, Square Cubed Consulting analyzed
historical enrollment, births, and new school impacts. Then, three different forecasting
methods were used to develop a five-year forecast.

Square Cubed Consulting also agreed to provide student yields for different residential
unit types. Even though the forecasting methods described in this document account
for natural residential development, these yields are important for any future
developments that fall outside what is considered normal. Outside the norm is a
development that significantly increases the number of closed units in one given year,
or a large apartment community that adds 100+ units in one year.

Square Cubed Consulting also agreed to provide a scatter plot map for each of the
schools that report to the Colorado Department of Education. The scatter plot maps are
based on the geocoded student addresses that Square Cubed Consulting received in
September 2016. Square Cubed will include the scatter plots in the presentation.

Key Data Points

1.0 Historical Enrollment: Between 2000 and 2005, the K12 enrollment for SSSD added a
total of 32 students. However, in 2001 and 2003 overall enrollment declined for those
specific years. Then between 2006 and 2010, the K12 enrollment saw a dramatic uptick
adding 254 students. The increase in enrollment continued between 2011 and 2015,
when another 283 students were added. For 2016, the K12 enrollment grew by only 10
students from 2015 (Chart 1.1).

Chart 1.1: SSSD K12 Enrollment


3,000

2,500
2,516
2,000 2,233
1,979
1,947
1,500

1,000

500

0
2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

Square Cube Consulting 1


The K12 enrollment growth rate has only seen declines in two years since 2000. In
2013, the K12 enrollment nearly set a new growth rate record at 3.5%. Since 2013, the
rate has steadily declined over the last three years (Chart 1.2).

Chart 1.2: K12 Enrollment Growth Rates


5.0%
4.0% 3.8% 3.5%
3.0% 3.1%
2.5% 2.8% 2.8%
2.0% 2.1% 2.2% 1.9%
1.7%
1.0% 1.2% 0.9%
0.5% 0.4%
0.0%
-1.0% -1.1%
-2.0% -1.8%

-3.0%
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Besides the declining growth rates there is also a plateau moving through SSSD. In Chart
1.3, the bubble is centered on the 2010 2nd grade with enrollment at 201. By 2016, the
same bubble is now in 8th grade and has 230 students. In fact, this cohort in 2016 is the
single largest grade for any grade any year between 2010 and 2016. Over the next five
years this bubble will graduate and leave SSSD.

Chart 1.3: 2010 Enrollment by Grade


250

200

150

100

50

0
Kin 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th

Chart 1.4: 2016 Enrollment by Grade

Square Cube Consulting 2


250

200

150

100

50

0
Kin 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th

2.0 Births: Birth data is sourced from the Colorado Department of Health and includes
all births within the SSSD boundary. All birth counts are moved forward five years to
correspond with that year’s kinder count (i.e., children born in 2005 are counted in the
2010 forecast). This is why you will see birth counts for upcoming years. For the rest of
this analysis, any mention of birth counts will be for this 5-year cohort, not birth year,
unless specifically called out as birth year. This allows us to see the relationship
between birth counts and kinder population.

Between 2000 and 2020, SSSD has averaged 132 births per year, peaking in 2010. Births
stayed at this elevated level over the next five years, leading to a mini baby boom.
Then, due to the recession families started having fewer children. This was seen, not
just within SSSD, but also the state of Colorado and the U.S. as a whole.

Chart 2.1: Birth and Kinder


200
180
160
140
120
100
80
60
40
20
0

Births Kinder

As you can see in Chart 2.1, the kinder tends to move in conjunction with the birth
count. In fact, as expressed as a correlation coefficient, births to kinder has a coefficient

Square Cube Consulting 3


of 0.90. In mathematics, a coefficient of 1 means there is a positive linear correlation
between the two variables. A coefficient of -1 means that the variables have a negative
linear correlation, or simply stated, the variables move in opposite directions. A
coefficient of 0 means that the variables have no linear correlation.

When forecasting kinder enrollment, having a coefficient of 0.90 means that the kinder
enrollment should move in the same general direction as the birth counts. It does not
mean they have to be an exact 1 to 1 gain or loss.

3.0 New School Impact: A new Montessori program opened in the SSSD boundary in
2016. At build out it is expected to be a K-8 school. For the opening year, there were 28
students in the kinder class. Of the 28, it is estimated that 20 are SSSD students. For
the remaining 1st through 5th grades, there are 50 SSSD residing students. For
forecasting purposes the 50 students were distributed more heavily in the lower grades.
This was done for two reasons. First, in general families like to have a continuity of
education. 4th and 5th grade students are less likely to leave for a new school than lower
grades.

Second, when looking at the Cohort Comparison Rate for 2016 the kinder and 1 st grade
rates were below 1.0 for the first time in six years. This is most likely due to the opening
of the new school. Below is the estimated grade-by-grade breakout for the new
program. These numbers will need to be taken into account when looking at each of
the forecasting methods.

Chart 3.1: Expected SSSD Reside Children Attending Montessori Program


Year Kind. 1st 2nd 3rd 4th 5th 6th 7th 8th Total
2016 20 20 15 10 2 3 70
2017 20 20 20 15 10 2 3 90
2018 20 20 20 20 15 10 2 3 110
2019 20 20 20 20 20 15 10 2 3 130
2020 20 20 20 20 20 20 15 10 2 147
2021 20 20 20 20 20 20 20 15 10 165
2022 20 20 20 20 20 20 20 20 15 175
2023 20 20 20 20 20 20 20 20 20 180

Forecast Methods

The goal of any method is not to hit a specific number exactly. When forecasting, time
is best spent on identifying the key data points and gaining certainty on how those key
data points will impact your forecasting number. Chasing down every minute data point
will not make a forecast more accurate. In fact, it is likely to cause issues due to time
taken away from the key data points.

Square Cube Consulting 4


The goal of producing a forecast is to ensure that all potential risks have been identified
and accounted for. When forecasting, K12 enrollments the major risks are cohort
bubbles, declining birth rates, and new school impacts. By properly identifying the
impact of such risk, the stakeholders can then make plans to mitigate any risk.

There are three main forecasting methods that can be used. Each of these methods has
inherent risks and benefits.

Regression: The regression method is standard practice in the field of statistics. The
method calculates relationship between multiple variables. A classic example is the
price of gas and oil production. In this example, the price of gas is the dependent
variable. Oil production is the independent variable. How good of a predictor oil
production is for gas prices, is defined as the r-squared value. The closer the r-squared
value is to 1.0 the better a predictor it is.

For enrollment forecasting there are two independent variables and one dependent
variable. The two independent variables are Total Population and Student Age
Population, ages 5-17 (SAP). The dependent variable is the K12 enrollment.

The State Demographers office provided the two independent variables. In my


experience, the forecast provided by the state are significantly overstated. At the state
level the 10-year Total Population forecast has a Median Absolute Percent Error (MAPE)
of 7-8%. When forecasting at a smaller geographic size, such as a county, the 10-year
can grow up to 10%. It is recommended that an error rate of 4% be applied to the
state’s Total Population five-year forecast.

The SAP is has an even larger MAPE. It is estimated that the 5-year error rate is around
10% and should be applied to the state’s SAP forecast.

Chart 5.1: Original DOLA Routt County Forecast. Major Axis is SAP. Minor Axis is Total Pop.
4,000 30,000
3,900
25,000
3,800
3,700 20,000
3,600
15,000
3,500
3,400 10,000
3,300
5,000
3,200
3,100 0
2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

SAP TOTAL Pop

Square Cube Consulting 5


The last census was in 2010. Everything since 2010 is an estimate. As Chart 5.1 clearly
shows, the state demographer has the SAP estimates since 2010 growing at a much
faster rate relative to the Total Population. This is highly unlikely. So not only is the
five-year forecast (2017-2021) inaccurate, it is very likely that the 2011 to 2015
estimates, that the forecast are based on, are also inaccurate.

Chart 5.2: Revised DOLA Routt County Forecast. Major Axis is SAP. Minor Axis is Total Pop.
3,900 26,750 27,000
3,850 26,500
3,843 26,000
3,800 25,500
3,750 25,000
24,970
3,700 24,500
3,684 24,000
3,650 23,500
3,600 23,000
2016 2017 2018 2019 2020 2021

SAP TOTAL Pop

The Square Cubed revised Routt County forecast flattens out the growth in SAP, as the
full impact of the baby bust starts to be felt in 2019. The revised 2021 numbers are
what are used for the dependent variables in the regression analysis.

The full regression analysis is included in Appendix A. Below are the key data points
used to forecast the 2021 K12 enrollment.

Forecast = (SAP * V1) + (Total Pop * V2) + Intercept

Variable Coefficients
Intercept -3468.832209
X Variable 1 (V1) 1.164941536
X Variable 2 (V2) 0.06745717
Multiple R 0.97626495
R-Squared 0.953093252

By applying the formula above to each of the revised Total Population and SAP, you will
get the following forecast.

Year Total Pop. SAP K12 Forecast


2017 24,871 3,727 2,550
2018 25,376 3,780 2,647
2019 26,018 3,804 2,717

Square Cube Consulting 6


2020 26,518 3,834 2,777
2021 27,026 3,875 2,837

Please, keep in mind that for any given year’s K12 forecast, the corresponding number
of students from Chart 3.1 should be removed. The reason for this is that the regression
method does not take into account the SSSD residing students attending the Montessori
Program.

There are two main issues with using a regression method for student enrollment. First,
the enrollment forecast is based off of a forecast for Total Population and SAP. In turn,
the population estimates are based of estimates for years 2011 to 2015. The last good
number was the 2010 census. If the population forecasts or estimates are off, there will
be a larger error rate for the enrollment forecast.

The second issue has to do with the fact that the regression model provides no insight
into the any specific K12 enrollment trends. The biggest that comes to mind being the
size of particular cohorts moving through the years. As shown in Charts 1.3 and 1.4,
there is a bubble moving through the district. The regression model does not do a good
job of capturing the bubble exiting the school system.

Growth Rate: A second method for calculating a forecast is to apply a growth rate based
on historical averages. As shown in Chart 1.2, the growth rate for SSSD has been
slowing since 2013. This slow down should be accounted for in any forecasted growth
rate. Below are the growth rate averages for different time frames.

2000 to 2016 1.7%


2010 to 2016 2.3%
2012 to 2016 2.1%
2014 to 2016 1.7%

Reviewing the growth rates above, a growth rate of 1.7% is the safest rate to use. If you
apply the rate of 1.7% for each of the years between 2017 and 2021 you get the
following products:

Year Forecast
2017 2,569
2018 2,613
2019 2,657
2020 2,702
2021 2,748

Obviously, this is a very simple method and does not take into account any nuances that
might be taking place within certain grades or impacts from new schools opening.

Square Cube Consulting 7


Cohort Comparison Rate (CCR): Also known as cohort survival, is the method used by
most K12 enrollment demographers. The CCR identifies the rate for any given grade by
comparing that grade with the previous year’s lower grade. An example is the 2015 5th
grade has 192 students. By dividing the 192 from the 2014 4th grade enrollment number
of 183, a rate of 1.05 is produced.

One of the drawbacks to using CCR is that kinder has no cohort. A demographer has a
few options to account for this issue. One, they could simply apply a growth rate to
kinder based on the historical kinder trends. This works if there are not any wild
fluctuations in the number of births for any given year. However, as shown in Chart 2.1,
the births for SSSD do vary greatly. In addition, we also know that there is a strong
correlation between births and kinder enrollment five years later.

For that reason, Square Cubed uses the birth counts gathered by the Colorado
Department of Health as a proxy for the kinder cohort. Below are the grade-by-grade
counts (Chart 6.1) for every year starting in 2010 and the corresponding Cohort
Comparison Rates (Chart 6.2).

Chart 6.1: K12 Enrollment


Year Births Kin 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th Total
2010 165 161 187 201 163 163 182 170 178 174 164 159 158 173 2,233
2011 146 165 174 200 206 178 169 189 172 182 167 168 151 161 2,282
2012 161 161 185 182 196 206 185 181 193 165 179 165 171 151 2,320
2013 162 174 181 182 187 195 211 194 181 195 179 178 165 179 2,401
2014 157 190 195 188 191 183 194 213 190 190 193 190 176 175 2,468
2015 151 176 207 201 192 197 192 195 221 201 193 187 179 175 2,516
2016 154 166 162 203 193 201 192 197 204 230 203 192 190 193 2,526

Chart 6.2 Cohort Comparison Rates


Year Kin 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th
2011 1.13 1.08 1.07 1.02 1.09 1.04 1.04 1.01 1.02 0.96 1.02 0.95 1.02
2012 1.00 1.12 1.05 0.98 1.00 1.04 1.07 1.02 0.96 0.98 0.99 1.02 1.00
2013 1.07 1.12 0.98 1.03 0.99 1.02 1.05 1.00 1.01 1.08 0.99 1.00 1.05
2014 1.21 1.12 1.04 1.05 0.98 0.99 1.01 0.98 1.05 0.99 1.06 0.99 1.06
2015 1.17 1.09 1.03 1.02 1.03 1.05 1.01 1.04 1.06 1.02 0.97 0.94 0.99
2016 1.08 0.92 0.98 0.96 1.05 0.97 1.03 1.05 1.04 1.01 0.99 1.02 1.08

Once the CCR’s have been identified, the demographer must decide on what CCR to use
for the forecast. This is where the art comes into play. Picking the wrong CCR will have
implications for all subsequent years and grades.

To identify what CCR to use, Square Cubed Consulting identified the six-year average (All
Year), the three-year average, and the two-year average (Chart 6.3).

Square Cube Consulting 8


Chart 6.3: Cohort Comparison Averages
Average K 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th
ALL YR 1.11 1.08 1.02 1.01 1.02 1.02 1.03 1.02 1.02 1.01 1.01 0.99 1.03
3 YR 1.15 1.04 1.02 1.01 1.02 1.01 1.01 1.02 1.05 1.01 1.01 0.98 1.04
2 YR 1.12 1.00 1.01 0.99 1.04 1.01 1.02 1.04 1.05 1.01 0.98 0.98 1.04

The kinder 2016 CCR of 1.08 is much lower than the previous year’s CCR. This is due to
the opening of the Montessori program. Square Cubed Consulting believes that moving
forward the CCR for kinder will be more in line with the 2016 rate of 1.08.

The 1st grade 2016 rate of 0.92 is also due to the opening of the Montessori program.
Excluding 2016, the 1st grade rate averaged 1.11 between 2011 and 2015. Square
Cubed Consulting believes that the 1st grade rate moving forward will rebound nearer to
the average of 1.11. This is expected because 2016 is the first year that the Montessori
program opened. Moving forward it is expected that students who wish to attend the
Montessori program will enter the program as kinder students.

The reasoning for the 1st grade CCR also holds true for the 2nd and 3rd grade CCR’s.
Reviewing the 2011 to 2015 CCR’s for 2nd grade, the average was 1.03. Three of the
years were above the average. For this reason, Square Cubed Consulting expects the 2nd
grade CCR to be 1.04.

The 3rd grade is expected to have an average CCR of 1.03 moving forward. Again, the
2016 rate should be discounted due to the opening of the new Montessori program.

Below are the CCR’s by grade that Square Cubed Consulting used to calculate the 5-year
forecast (Chart 6.4).

Chart 6.4: CCR’s used for Forecasting


Kind 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th
1.08 1.12 1.04 1.03 1.04 1.04 1.03 1.02 1.05 1.01 1.01 1.00 1.04

The CCR’s take into account any normal development activity related to the
construction of new residential units. For any given year, there are a certain number of
new units constructed and closed, which will then possibly add students to the SSSD
enrollment.

Chart 6.5: Five-Year Forecast based on Assumed CCR’s


Birth K 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th Total
2017 135 146 186 168 209 201 208 198 201 214 232 205 192 198 2,558
2018 135 146 163 193 174 217 208 214 202 211 216 235 205 200 2,584
2019 115 124 163 170 199 180 225 214 219 212 216 219 235 213 2,589
2020 147 159 139 170 175 207 187 232 218 229 216 218 219 244 2,613
2021 150 162 178 145 175 182 214 192 236 229 242 218 218 227 2,618

Square Cube Consulting 9


The only grade where an adjustment was made was in the 9th grade. Once the
Montessori program is built out, the 8th students will reenter SSSD the following 9th
grade year. The expected impacts can be seen in Chart 3.1.

Recommended Forecast

Square Cubed Consulting suggests that SSSD uses the forecast produced from the CCR in
Chart 6.5. This method accounts for all of the major factors related to enrollment
forecasting. Those factors are: birth rates, historic enrollment trends, cohort bubbles,
new school impact, and natural residential development.

The regression forecast is a quick mathematical way to forecast enrollment. However,


even though the regression model used has a very high r-squared value, it relies on
future Total Population and SAP forecast to be accurate. As discussed above, Square
Cubed Consulting is concerned about the original forecast provided by the state
demographer. We have tried to revise them as best as possible, but the inherit risk is
still great.

The Growth Rate method should also be discounted. It happens to be the quickest way
to produce a forecast, but also ignores major impacts such as lower birth rates and new
school impacts.

In the experience of Square Cubed Consulting, the Regression and Growth Rate methods
should be calculated because of the ease of the calculations and to assist in framing
enrollment trends. Typically, these two models over project the total enrollment.

Student Yields: Student yields are calculated by geocoding current student addresses,
and then dividing the number of students within a particular residential unit type by the
total number of residential units by type. This allows a demographer to apply a yield to
a new development to understand the impact the development will have on enrollment.
It is particularly important for large developments that fall outside the historical typical
development cycle. The yields are provided here so that SSSD and the committee can
calculate the impact of any new development without having to contract with a
developer in the future.

One other note, the yields are based off of the type of unit built, not if the unit is owner
occupied or being rented out. For this reason, Square Cubed Consulting thought it
important to create yields based on regions. In particular, it was important to breakout
the resort area from downtown.

Square Cubed Consulting received a current student address file from SSSD in
September 2016. The file was geocoded and successfully found each student’s address.
However, students living in the county have a higher placement error (off by a certain

Square Cube Consulting 10


number of feet), then those living in the city of Steamboat Springs or near the resort.
For this reason, the yields listed below for the county regions should be taken with a
grain of salt.

Chart 7.1 Single-Family Detached Yields


Region Students Units Yield

Resort 228 985 0.23


Downtown 592 1,473 0.40
North 327 1,088 0.30
South 108 621 0.17

The single-family detached yields are what Square Cubed Consulting expects (Chart 7.1).
The yield is highest in the downtown area. In fact, the downtown yield is very similar to
the single-family detached yield seen in Denver. The large number of vacation homes
can explain the lower yield, for the resort region.

Chart 7.2 Single-Family Attached Yields


Region Students Units Yield
Resort 36 1,431 0.03
Downtown 43 472 0.09
North 0 38 0
South 0 17 0

The single-family attached yields are also what Square Cubed Consulting expects for the
downtown and resort areas (Chart 7.2).

Chart 7.3 Condominium Yields


Region Students Units Yield

Resort 52 4,494 0.01


Downtown 12 631 0.02
North 0 0 0.00
South 0 43 0.00

Square Cube Consulting 11


Chart 7.4: Apartment Yields (All are in the Downtown region)
Name Students Units Yield
Hillside Village 21 90 0.23
215 Old Fish Creek 4 8 0.50
Hillcrest Apts 0 8 0.00
603 Pahwintah 3 6 0.50
Total 28 112 0.25

The KOA campground was reviewed and no students are currently residing within in the
property. The mobile home park just west of the campground was also reviewed.
There are 30 students living in this mobile home park.

In addition, it was recommended that the western portion of Downtown be analyzed


separately from the rest of downtown. The Single-family yield for this area is 0.36. The
only other residential type was single-family attached. The yield for single-family
attached is 0.58.

Square Cube Consulting 12


Appendix A: School by School Forecast

North Year K 1ST 2ND 3RD 4TH 5TH 6th 7th 8th 9th 10th 11th 12th Total
Routt 2017 10 12 9 11 11 11 11 11 11 0 0 0 0 97
Charter 2018 10 10 12 9 11 11 11 11 11 0 0 0 0 96
2019 10 10 10 12 9 11 11 11 11 0 0 0 0 95
2020 10 10 10 10 12 9 11 11 11 0 0 0 0 94
2021 10 10 10 10 10 12 9 11 11 0 0 0 0 93

Soda Creek Year K 1ST 2ND 3RD 4TH 5TH 6th 7th 8th 9th 10th 11th 12th Total
Elementary 2017 70 93 78 112 105 110 0 0 0 0 0 0 0 568
School 2018 70 81 97 81 115 109 0 0 0 0 0 0 0 553
2019 58 81 85 100 83 118 0 0 0 0 0 0 0 525
2020 76 68 84 87 102 86 0 0 0 0 0 0 0 503
2021 78 87 71 87 89 107 0 0 0 0 0 0 0 519

Strawberry Year K 1ST 2ND 3RD 4TH 5TH 6th 7th 8th 9th 10th 11th 12th Total
Park 2017 66 81 81 86 85 87 0 0 0 0 0 0 0 486
Elementary 2018 66 72 84 84 91 88 0 0 0 0 0 0 0 485
School 2019 56 72 75 87 88 96 0 0 0 0 0 0 0 474
2020 73 61 76 78 93 92 0 0 0 0 0 0 0 473
2021 74 81 64 78 83 95 0 0 0 0 0 0 0 475

Steamboat Year K 1ST 2ND 3RD 4TH 5TH 6th 7th 8th 9th 10th 11th 12th Total
Springs 2017 0 0 0 0 0 0 187 190 203 0 0 0 0 580
Middle 2018 0 0 0 0 0 0 203 191 200 0 0 0 0 594
School 2019 0 0 0 0 0 0 203 208 201 0 0 0 0 612
2020 0 0 0 0 0 0 221 207 218 0 0 0 0 646
2021 0 0 0 0 0 0 183 225 218 0 0 0 0 626

Steamboat Year K 1ST 2ND 3RD 4TH 5TH 6th 7th 8th 9th 10th 11th 12th Total
Springs 2017 0 0 0 0 0 0 0 0 0 231 200 184 187 802
High 2018 0 0 0 0 0 0 0 0 0 215 230 197 189 831
School 2019 0 0 0 0 0 0 0 0 0 215 214 227 202 858
2020 0 0 0 0 0 0 0 0 0 215 213 211 233 872
2021 0 0 0 0 0 0 0 0 0 241 213 210 216 880

Yampa Year K 1ST 2ND 3RD 4TH 5TH 6th 7th 8th 9th 10th 11th 12th Total
Valley 2017 0 0 0 0 0 0 0 0 0 1 5 8 11 25
School 2018 0 0 0 0 0 0 0 0 0 1 5 8 11 25
2019 0 0 0 0 0 0 0 0 0 1 5 8 11 25
2020 0 0 0 0 0 0 0 0 0 1 5 8 11 25
2021 0 0 0 0 0 0 0 0 0 1 5 8 11 25

Square Cube Consulting 13


Appendix B: Birth Data

Cohort
Year Births
2000-01 94
2001-02 105
2002-03 94
2003-04 99
2004-05 110
2005-06 150
2006-07 119
2007-08 119
2008-09 132
2009-10 120
2010-11 165
2011-12 146
2012-13 161
2013-14 162
2014-15 157
2015-16 151
2016-17 154
2017-18 135
2018-19 135
2019-20 115
2020-21 147
2021-22 150
Note: The 2021 cohort year corresponds with 2016. The 150births is an estimate.

Square Cube Consulting 14


Appendix C: Out of District Counts

K12
Year Count
2007 48
2008 45
2009 56
2010 115
2011 104
2012 109
2013 111
2014 122
2015 129
2016 142

2016 by Grade
Kinder 9
1 10
2 12
3 13
4 15
5 9
6 8
7 7
8 14
9 11
10 11
11 10
12 11
Total 140

Square Cube Consulting 15


Appendix A: Regression Analysis

Regression Statistics
Multiple R 0.971675552
R Square 0.944153378
Adjusted R
Square 0.93556159
Standard Error 52.88720633
Observations 16

ANOVA
Significance
df SS MS F F
7.16935E-
Regression 2 614738.2643 307369.1321 109.8902085 09
Residual 13 36361.73571 2797.056593
Total 15 651100

Standard Lower Upper Lower Upper


Coefficients Error t Stat P-value 95% 95% 95.0% 95.0%
- - - -
Intercept -3468.832 583.461 -5.945 0.000 4729.323 2208.341 4729.323 2208.341
X Variable 1 1.165 0.265 4.394 0.001 0.592 1.738 0.592 1.738
X Variable 2 0.067 0.021 3.287 0.006 0.023 0.112 0.023 0.112

RESIDUAL PROBABILITY
OUTPUT OUTPUT

Observation Predicted Y Residuals Percentile Y


2000 1902.929499 44.07050103 3.125 1911
2001 1905.820353 5.179647007 9.375 1912
2002 1972.134164 39.13416362 15.625 1930
2003 1949.249179 37.24917859 21.875 1933
2004 1938.738371 8.738370752 28.125 1947
2005 1964.02474 14.9752602 34.375 1979
2006 2021.876908 0.876908491 40.625 2021
2007 2112.150237 35.15023722 46.875 2077
2008 2175.877564 33.87756413 53.125 2142
2009 2217.097655 65.09765507 59.375 2152
2010 2278.936818 45.93681753 65.625 2233
2011 2218.226366 63.77363361 71.875 2282
2012 2230.221221 89.778779 78.125 2320
2013 2317.321862 83.67813758 84.375 2401
2014 2449.054241 18.94575879 90.625 2468
2015 2570.340822 -54.3408218 96.875 2516

Square Cube Consulting 16


Strawberry Park Elementary School Scatter Plot

Square Cube Consulting 17


Steamboat Springs Middle School Scatter Plot

Square Cube Consulting 18


Steamboat Springs High School Scatter Plot

Square Cube Consulting 19


North Routt Charter School Scatter Plot

Square Cube Consulting 20


Soda Creek Elementary School Scatter Plot

Square Cube Consulting 21


Yampa Valley High School Scatter Plot

Square Cube Consulting 22

You might also like