
How to Read the
CIRP Construct Percentage Reports 















CIRP
Constructs are designed to capture the experiences and outcomes institutions
are often interested in understanding, but that present a measurement
challenge because of their complex and multifaceted nature. To measure these
broad underlying areas more precisely, we use Item Response Theory (IRT) to
combine individual survey items into global measures that capture these
areas. CIRP Constructs are more than a summation of related items; IRT uses
response patterns to derive construct score estimates while simultaneously
giving greater weight in the estimation process to survey items that tap into
the construct more directly. This results in more accurate construct scores.
Constructs are particularly useful for benchmarking. They allow you to
determine if the experiences and outcomes for your students differ from your
comparison groups. Two sets of reports are generated for CIRP Constructs. The
Mean Report shows comparative information based on the mean score of a
construct. The Percentage Report shows comparative information based on the
percentage of students who score in the high, average, and low score group of
a construct. We suggest you use the report that best fits your needs as an
institution. 

CIRP Construct Definition – Summarizes the theoretical
rationale for creating the construct.
Comp 1 – The first comparison group is based on
your institution's type and control.
Comp 2 – The second
comparison group is based on a similar grouping of institution type and
control.
Statistical Significance – uses a proportional difference test to examine the
difference between the percentage of students in the high group for your
institution and the percentage of students in the high group in the
comparison group. Differences larger than what would be expected by chance
are noted with one, two, or three stars, which correspond to the three
standard levels of significance (*p<.05, **p<.01, ***p<.001).
Statistical significance measures the extent to which a difference is
occurring by chance, not the extent to which a difference is practically
important. Large sample sizes (like those in the comparison groups) tend to
generate statistical significance even though the magnitude of the difference
might be small and not practically important.














Academic SelfConcept is a unified measure of students’ beliefs about their
abilities and confidence in academic environments. 















Total 
Men 
Women 


Sample University 
Your Inst 
Comp 1 
Comp 2 
Your Inst 
Comp 1 
Comp 2 
Your Inst 
Comp 1 
Comp 2 


Total (n) 
1,361 
4,996 
14,835 
503 
1,586 
5,527 
858 
3,410 
9,308 


High
Academic SelfConcept 

43.6% 
33.2% 
30.3% 
59.6% 
43.0% 
36.7% 
34.3% 
28.7% 
26.5% 


Average
Academic SelfConcept 

30.3% 
33.9% 
39.9% 
23.5% 
31.6% 
39.1% 
34.3% 
35.0% 
40.4% 


Low
Academic SelfConcept 

26.1% 
32.8% 
29.8% 
16.9% 
25.4% 
24.2% 
31.5% 
36.2% 
33.1% 


Significance (based on High
score group) 
 
*** 
*** 
 
*** 

 

** 


Note: Significance *
p<.05, ** p<.01, *** p<.001 

























































































































































































Survey items and
estimation "weights":
Rate yourself on each of the
following traits as compared with the average person your age: 
For more information about IRT and the CIRP Construct
development process, see the CIRP Constructs Technical Report at
www.heri.ucla.edu 


* Academic ability
(3.01) 


* Selfconfidence
(intellectual) (1.51) 


* Drive to achieve
(1.18) 


* Mathematical
ability (1.14) 


Survey Items and Estimation "Weights" – The
survey items used in the creation of the CIRP Construct are presented in the
order in which they contribute to the construct along with the estimation
weights generated in IRT. Items that tap into a trait more effectively are
given greater weight in the estimation process. 
Charts – Provide a visual display of construct score
group percentages for your institution and two comparison groups. CIRP
Constructs have been scaled to a mean of 50 with a standard deviation of 10.
"Low" represents students who scored onehalf standard deviation or
more below the mean (less than 45). "Average" represents students
who scored within onehalf standard deviation of the mean (45 to 55).
"High" represents students who scored onehalf standard deviation
or more above the mean (higher than 55). 













