Professional Documents
Culture Documents
Blazy Data Analysis
Blazy Data Analysis
Blazy Data Analysis
Mina J. Blazy
Background Information
The following analysis is based on the end of the year second grade benchmark analysis that
exams all core standards in mathematics. This assessment for the K-5 Go Math was created by
the Houghton Mifflin Harcourt publisher for elementary. The teachers in the Adelanto
Elementary School District (AESD) to a consensus and chose the curriculum in Spring of 2014.
Since the Fall of 2014 AESD has used the Go Math curriculum for K-5 grade levels.
Go Math (Education and Learning Resources, n.d.) states that the three benchmark
assessments are aligned to the Core Mathematics standards (CCSS mathematics resources -
common core state standards (CA dept of education), n.d.). There is a beginning, mid and end-
of-year benchmark assessment available for K-6 grades. The AESD school district only uses the
benchmarks from Go Math for K-2. If students are proficient on the assessment the company
states that students will also do well on the Smarter Balanced Assessment Consortium (SBAC).
For this data set 77 students were given the second-grade mathematics end-of-year benchmark
assessment with 27 items on the test.
The mean on the assessment of student data is 38.16, medium is 47 and the mode is 40; the
variance is 66.08, the variance of error is 12.70 and the standard error of measure (SEM) is 3.56
leaving the reliability coefficient at .81.
The Wright Map is identified in Table 1.0; also known as the item-person map or variable map.
The logit scale below is an interval scale with equal distance. Any point on the vertical scale the
items or person data are also equal in size (Bond & Fox, 2015a, p. 69). The mean of the item is
set at zero (0). The vertical line from is set at -2, -1, 0, 1, 2, 3 for the data given. The students or
persons are on the left side and the items or questions are on the right side. The higher the
students the more likely they are to answer the items and the lower the students on the scale the
less likely they will be to correctly answer the items. The higher the item students will rarely get
the item correctly answered and the lower the item the more frequent the items will be answered
correctly.
The students who could correctly answer items I21 and I6 at 0, for example, have a 50%
probability of succeeded or failing on these items. Items I24, I16, I15, I4, and I7 are too easy for
this group of students and there a no items that are too difficult for all students.
3
TABLE 1.0 Second Grade EOY Mathematics Benchmark ZOU286WS.TXT Aug 5 9:22 2017
INPUT: 77 studentss 27 Qss MEASURED: 77 studentss 27 Qss 80 CATS 3.68.2
--------------------------------------------------------------------------------
Table 3.1 student measures for the end-of year mathematics benchmark for second grade has a
mean score of 1.06; when looking at table 1.0 you can visually see that the mean of the students
is higher than the mean of the items at 0. The students and items are independent of both the test
item and item difficulty.
The reliability coefficient is .81 because there are more moderately discriminating items and less
high and low discriminating items.
The standard deviation (SD) for students is .79 and the SD for the items is 1.00 from the mean.
The reliability RASCH model for students taking this assessment is .82 and the item reliability
for the RASCH model is .95; the Cronbach Alpha (KR-20) of raw score reliability is .84; This
information shows that the items and the students data are reliable based on the RASCH Model.
TABLE 3.1 Second Grade EOY Mathematics Benchmark ZOU286WS.TXT Aug 5 9:22 2017
INPUT: 77 studentss 27 Qss MEASURED: 77 studentss 27 Qss 80 CATS 3.68.2
--------------------------------------------------------------------------------
Table 10.1 for the second grade mathematics end-of-year benchmark assessment for items 22,
12, 24 and 14 are underfitting because the mean square (MNSQ) is either near or above 1.50.
The expected value of the mean square is equal to one (1). Items with a MNSQ near .70 or lower
is considered better than expected. The underfitting data for the data set in 10.1 is consistently
showing that there is noise in these data. While the items show easy a few students that should
have successfully answered these items failed at their attempts.
In table 10.3 for items 12, 24, 14, 19 and 20 have average measures that are not in ascending
order for the polytomous data. All the items are listed below the mean on the Wright logit scale
accept for I19 which is slightly above one (1) on the logit scale (table 1). Teachers may have
been undeciding of whether to give a 1 or a 2 when grading the assessment.
TABLE 10.1 Second Grade EOY Mathematics Benchmark ZOU286WS.TXT Aug 5 9:22 2017
INPUT: 77 studentss 27 Qss MEASURED: 77 studentss 27 Qss 80 CATS 3.68.2 Over 1.5 underfitting
--------------------------------------------------------------------------------
students: REAL SEP.: 1.94 REL.: .79 ... Qs: REAL SEP.: 4.28 REL.: .95
----------------------------------------------------------------------------------------------
|ENTRY TOTAL MODEL| INFIT | OUTFIT |PT-MEASURE |EXACT MATCH| |
|NUMBER SCORE COUNT MEASURE S.E. |MNSQ ZSTD|MNSQ ZSTD|CORR. EXP.| OBS% EXP%| Qs G |
|------------------------------------+----------+----------+-----------+-----------+---------|
| 5 133 77 -.59 .22|1.14 .7|2.47 3.1|A .17 .38| 71.4 74.1| I22 0 |
| 18 107 77 .43 .15|1.14 .9|1.94 2.7|B .42 .55| 59.7 59.5| I12 0 |
| 1 149 77 -1.88 .47|1.13 .5|1.92 1.5|C-.10 .19| 93.5 93.5| I24 0 |
| 13 118 77 -.57 .22|1.47 2.8|1.49 2.8|D-.06 .40| 44.2 64.4| I14 0 |
| 17 110 77 .32 .16|1.18 1.2|1.36 1.3|E .40 .53| 51.9 58.0| I5 0 |
| 4 136 77 -.38 .20|1.26 1.1|1.32 .6|F .27 .40| 75.3 78.1| I17 0 |
| 19 103 77 -.77 .24|1.22 1.8|1.24 1.6|G .07 .35| 58.4 66.7| I8 0 |
| 24 63 77 1.45 .17|1.19 1.4|1.23 1.5|H .33 .49| 37.7 48.7| I2 0 |
| 2 147 77 -1.25 .32|1.11 .4| .88 .1|I .18 .24| 93.5 93.5| I7 0 |
| 6 131 77 -.21 .18|1.04 .3| .81 -.2|J .43 .44| 70.1 71.7| I26 0 |
| 12 121 77 -.39 .21|1.03 .2| .97 -.1|K .41 .42| 64.9 64.9| I25 0 |
| 9 126 77 -.27 .19| .99 .0| .74 -.8|L .47 .43| 63.6 68.0| I18 0 |
| 25 61 76 1.41 .15| .94 -.4| .99 .0|M .57 .53| 46.1 43.1| I19 0 |
| 22 88 77 .68 .19| .97 -.1| .97 -.2|N .47 .44| 57.1 53.5| I3 0 |
| 27 36 77 2.12 .17| .96 -.2| .81 -.7|m .48 .43| 61.0 59.1| I13 0 |
| 3 145 77 -1.22 .30| .95 .0| .58 -.6|l .35 .27| 90.9 90.9| I15 0 |
| 10 123 77 -.05 .18| .95 -.2| .77 -.6|k .52 .47| 68.8 64.6| I6 0 |
| 11 123 77 -.22 .19| .95 -.2| .92 -.2|j .49 .44| 76.6 65.5| I20 0 |
| 7 130 77 -1.32 .25| .91 -.5| .95 -.2|i .44 .35| 75.3 71.6| I16 0 |
| 15 113 77 -.48 .22| .90 -.7| .90 -.7|h .50 .40| 70.1 63.0| I1 0 |
| 23 75 77 1.12 .16| .89 -.9| .89 -.7|g .60 .52| 45.5 44.2| I23 0 |
| 26 42 77 2.02 .17| .89 -.8| .83 -.8|f .54 .44| 62.3 56.9| I11 0 |
| 16 113 77 .30 .16| .85 -.9| .60 -1.2|e .64 .54| 64.9 59.8| I9 0 |
| 8 126 77 -1.25 .24| .81 -1.4| .78 -1.4|d .57 .36| 75.3 68.8| I4 0 |
| 21 99 77 .48 .17| .77 -1.8| .72 -1.8|c .66 .50| 44.2 50.8| I27 0 |
| 14 117 77 .05 .18| .75 -1.6| .62 -1.6|b .65 .48| 67.5 60.4| I21 0 |
| 20 103 77 .47 .16| .73 -2.0| .62 -1.9|a .70 .54| 57.1 51.3| I10 0 |
|------------------------------------+----------+----------+-----------+-----------+---------|
| MEAN 108.8 77.0 .00 .21|1.00 .0|1.05 .1| | 64.7 64.6| |
| S.D. 29.7 .2 1.00 .07| .17 1.1| .45 1.4| | 14.4 13.1| |
----------------------------------------------------------------------------------------------
6
TABLE 10.3 Second Grade EOY Mathematics Benchmark ZOU286WS.TXT Aug 5 9:22 2017
---------------------------------------------------------------------
|ENTRY DATA SCORE | DATA | AVERAGE S.E. OUTF PTMEA| |
|NUMBER CODE VALUE | COUNT % | MEASURE MEAN MNSQ CORR.| Qs |
|--------------------+------------+--------------------------+------|
| 5 A 0 0 | 5 6 | .76 .60 4.2 -.10 |I22 |
| 1 1 | 11 14 | .81 .24 1.2 -.13 | |
| 2 2 | 61 79 | 1.13 .09 1.1 .17 | |
| | | | | Average measure should be
| 18 B 0 0 | 22 29 | .52 .19 2.4 -.44 |I12 |
| 1 1 | 3 4 | 1.28 .23 1.0 .06 | |
in ascending order for each
| 2 2 | 52 68 | 1.28* .09 1.0 .40 | | item.
| | | | |
| 1 C 1 1 | 5 6 | 1.36 .27 2.0 .10 |I24 | Items, 12, 24, 14, 5, 11 and
| 2 2 | 72 94 | 1.04* .09 1.1 -.10 | | 19 are not in ascending
| | | | |
| 13 D 0 0 | 3 4 | 1.05 .12 1.5 .00 |I14 | order. These are a misfit for
| 1 1 | 30 39 | 1.13 .13 1.5 .07 | | the assumption of the way
| 2 2 | 44 57 | 1.01* .13 1.4 -.07 | |
| | | | | students should have
| 17 E 0 0 | 18 23 | .43 .17 1.4 -.44 |I5 |
| 1 1 | 8 10 | 1.33 .20 1.6 .12 | | answered.
| 2 2 | 51 66 | 1.24* .10 1.3 .32 | |
| | | | |
| 4 F 0 0 | 9 12 | .48 .20 1.3 -.27 |I17 |
| 2 2 | 68 88 | 1.14 .10 1.2 .27 | |
| | | | |
| 19 G 0 0 | 1 1 | 1.01 1.0 -.01 |I8 |
| 1 1 | 49 64 | 1.02 .12 1.2 -.07 | |
| 2 2 | 27 35 | 1.14 .15 1.3 .08 | |
| | | | |
| 24 H 0 0 | 32 42 | .88 .13 1.3 -.19 |I2 |
| 1 1 | 27 35 | .91 .16 1.7 -.14 | |
| 2 2 | 18 23 | 1.60 .14 1.0 .38 | |
| | | | |
| 2 I 0 0 | 2 3 | .15 .19 .7 -.19 |I7 |
| 1 1 | 3 4 | .93 .18 1.1 -.03 | |
| 2 2 | 72 94 | 1.09 .09 1.1 .15 | |
| | | | |
| 6 J 0 0 | 11 14 | .25 .18 .8 -.42 |I26 |
| 1 1 | 1 1 | .42 .1 -.09 | |
| 2 2 | 65 84 | 1.21 .09 1.0 .43 | |
| | | | |
| 12 K 0 0 | 5 6 | .20 .23 .8 -.29 |I25 |
| 1 1 | 23 30 | .77 .15 1.0 -.24 | |
| 2 2 | 49 64 | 1.28 .11 1.0 .37 | |
| | | | |
| 9 L 0 0 | 8 10 | .07 .14 .6 -.43 |I18 |
| 1 1 | 12 16 | .80 .15 .8 -.14 | |
| 2 2 | 57 74 | 1.25 .10 1.1 .42 | |
| | | | |
| 25 M 0 0 | 40 53 | .58 .10 .7 -.63 |I19 |
| 1 1 | 11 14 | 1.64 .14 .4 .31 | |
| 2 2 | 25 33 | 1.54* .13 1.3 .43 | |
| MISSING *** | 1 1#| 1.74 .10 | |
| | | | |
| 22 N 0 0 | 13 17 | .28 .20 .9 -.45 |I3 |
| 1 1 | 40 52 | 1.08 .10 .8 .03 | |
| 2 2 | 24 31 | 1.45 .16 1.1 .33 | |
| | | | |
| 27 m 0 0 | 52 68 | .81 .11 1.0 -.46 |I13 |
| 1 1 | 14 18 | 1.43 .14 .6 .22 | |
| 2 2 | 11 14 | 1.78 .15 .8 .37 | |
| | | | |
7
students who were higher on the logit scale did not successfully answer at least one of these
items. These items were answered correctly by most students but not by student one or student
12. The mean squared is also high for the items listed in table 10.4a which also confirms the
data doesnt fit the model.
Table 10.4a Example of items on the misfitting table with their item discrimination and
correlation
Item 7 8 14 17 22 24
Item Discrimination .10 .00 -.02 .19 .14 -.05
Biserial Correlation .17 .07 -.10 .26 .20 -.11
TABLE 10.4 Second Grade EOY Mathematics Benchmark ZOU286WS.TXT Aug 5 9:22 2017
Conclusion
When reviewing the Wright Map the mean of the items on the logit scale is set at zero (0) and the
mean of the students is approximately one (1) on the interval scale. The assessment was not too
easy or too difficult. The items between -1 and -2 could be removed or revisited. Students
between logit interval 2 and 3 should have access to items that are not on the test. Since this was
an end-of-the-year assessment the items from the next grade (third) could be added to the
assessment to give these students access to their abilities.
9
Each item assessed a specific standard for second grade mathematics. When viewing the data
ordered in excel the polytomous data is organized as a Scaleogram (Bond & Fox, 2015b, p. 24).
The more difficult items are to the right and the easier items are on the left. The items on the test
where valid and fit the model. Students who took the test fit the Rasch model as well. These
items/standards work well for second grade students after mastering the standards for the grade
level. In
References
Bond, T., & Fox, C. M. (2015a). Applying the rasch model: fundamental measurement in the
Bond, T., & Fox, C. M. (2015b). Applying the rasch model: fundamental measurement in the
CCSS mathematics resources - common core state standards (CA dept of education). (n.d.).
http://www.cde.ca.gov/re/cc/mathresources.asp
http://www.hmhco.com/