University of Arkansas Office for Education Policy

ACT scores are down, but it’s okay.

In The View from the OEP on September 7, 2017 at 6:24 am

This morning, ACT scores for the graduating class of 2017 were released.  Like we suggested in our blog yesterday, the scores were lower than the scores for every graduating class for over twenty years. While this may sound alarming, here at the OEP we are recommending that folks put the data in context and celebrate the good decisions Arkansas has made regarding ACT.

5yrs_17The statewide average composite score decreased from 20.2 for 2016 graduates to 19.4 for 2017 graduates.  As we mentioned yesterday, the decline is likely due to the fact that Arkansas tested 100% of graduates for the first time.  Declines were evident in every subject area:

  • English dropped from 19.8 to 18.9
  • Math dropped from 19.6 to 19.0
  • Reading dropped from 20.7 to 19.7 and
  • Science dropped from 20.2 to 19.5


Although there was an increase in the percentage of graduates tested, at least 90% or more of the graduating class has taken the ACT since 2013. Like we discussed in yesterday’s blog, some of the students who completed the ACT may be different than those who participated in prior years. These students likely struggle academically and/or don’t consider themselves college-bound.  To make the scores more comparable across the years, we suggest looking at the scores by academic preparation. Students who have taken ‘Core or More’ are those enrolling in more rigorous core content in high school (4+ English courses, 3+ math courses, 3+ social studies courses, and 3+ natural science courses).


Even when only comparing students who have taken similar academic preparation (Core or More), scores declined slightly across the state. The statewide average composite score for ‘Core or More’ graduates decreased from 20.9 for 2016 graduates to 20.4 for 2017 graduates.  Declines were again evident in every subject area:

  • English dropped from 20.7 to 20.2
  • Math dropped from 20.2 to 19.9
  • Reading dropped from 21.4 to 20.8 and
  • Science dropped from 20.8 to 20.4


One of the great things about ACT is that is taken by students throughout the US. In some states, however, only a small percentage of students take the ACT.  It isn’t appropriate to compare ACT scores for Arkansas with states that don’t test all their students. There are, however, 20 states that have tested 100% of students within the past two years, so we can compare the performance of Arkansas graduates to the graduates in those states. The states and the percentage of graduates tested since 2007 are presented in the table below. Years where 90% of graduates were tested are highlighted in green.

Note that Colorado, Illinois, and Mississippi have been testing all (or close to all) of their students for 10 years.  Michigan, who has also been testing 100% of graduates for nearly a decade switched to statewide SAT in the most recent year, which explains why only 29% of 2107 Michigan graduates completed the ACT.  We kept Michigan in the table, however, because they provide a great example of how changes in the percentage of students assessed can result in big changes to average ACT scores.


States Testing at Least 90% of High School Graduates, 2007-2017

Although the percentage of students tested impacts ACT scores, so does the demographics of the students.  State that serve a more economically disadvantaged population tend to have lower ACT scores than states that serve populations with less economic instability. We grabbed the percent of population that lives in poverty from, added it to the table of states that test all (or nearly all) of their graduates on ACT, and sorted it from least percentage of population in poverty to the greatest.  To this we added the average composite ACT score for the prior 10 years to illustrate the performance trends.


Poverty Rate (2014) and Average ACT Scores for States Testing at Least 90% of High School Graduates, 2007 to 2017

Of the states that tested at least 90% of high school graduates on the ACT, Wyoming had the smallest percentage of its population living below the poverty level at 11.2%.  Mississippi had the highest poverty rate at 21.5%.  In 2017, Wyoming had one of the highest ACT composite scores at 20.2, while Mississippi had one of the lowest at 18.6.  Arkansas has a greater poverty rate than most states at 18.9%, and an average composite score of 19.4.

Remember that Colorado, Illinois, and Mississippi have been testing all (or close to all) their students for 10 years. Although Colorado and Illinois have a smaller percentage of people living below the poverty line than Arkansas does, they can serve to illustrate what happens to ACT scores over time when 100% of students are tested each year. In Colorado, the scores have fluctuated slightly from 20.4 to 20.8, but have not consistently moved upward.  Illinois scores have also fluctuated slightly, from 20.5 to 20.9 until 2017, when scores rose but perhaps this was due to a reduction in the percentage of graduates completing the ACT (from 100% to  93%). Mississippi has a higher poverty rate than any other state that widely tests high school students on the ACT, and although lower than Colorado and Illinois, scores have also fluctuated slightly, from 18.4 to 19.0.

Kentucky and Alabama are the two states that are most similar to Arkansas in both ACT testing rates and poverty rates. Alabama tested 100% of graduates for the first time in 2015, and the statewide composite dropped from 20.6 to 19.1. The score stayed 19.1 for 2016, but crept up to 19.2 for the class of 2017.  Kentucky tested 100% of graduates for the first time in 2009, and the statewide composite dropped from 20.9 to 19.4. Over the eight years since the initial drop, however, Kentucky has made consistent gains in ACT scores, and has maintaining a high of 20.0 since 2015.  Although Kentucky has a greater percentage of people living below the poverty line than Arkansas, the average ACT score for the state is now higher than Arkansas’.

Remember Michigan?  They had tested 100% of students from 2008 to 2016, and had shown consistent improvement in average score over time.  Like with other states, there was some variation from 19.6 to 20.3, but in 2017 the score jumped to 24.1!  The reason- only 29% of students completed the ACT this year.

So What:

What these statewide scores over time illustrate several key points for Arkansas education stakeholders to consider.

  1. States experience a dip in scores when they begin to test all students.
  2. States that test a high percentage of graduates on ACT do not demonstrate large changes in average scores.  Most states hold relatively steady over time, with fluctuations of less than half a point.
  3. Poverty does not limit performance: Kentucky students are outperforming similar states.

Now What?

Arkansas need to keep testing 100% of graduates for the long haul.  We can’t just stop because we don’t like the scores.  Arkansas’ ACT scores will not increase, however,  unless educators and students do something differently!  This is a classic example of weighing the pig not making it fatter.


ACT is a meaningful test for students and parents, and student success on the ACT may soon become a part of how school quality is measured. While improvement may be difficult to see at the state level, changes can be implemented at the local level.  There are lots of resources available to support student success. Schools should be mindful to set realistic and meaningful goals for ACT improvement.

We LOVE that Arkansas is giving every student the opportunity to get a picture of his/her readiness for college and careers, and doing it early enough that students can use the information when making decisions. Although statewide scores decreased slightly, we think it is important to focus on the positive outcomes of the program rather than any decrease in scores. More students are getting more information about their achievement, and districts and schools have better information about the performance of their graduates.  This is a good thing, regardless of the statewide average scores.

What are your thoughts?


%d bloggers like this: