A Note of Caution about GED Test Score Changes

I just finished GED Testing Service’s Tuesdays for Teachers Webinar entitled Staying on Track and Hitting the Marks.  It was an upbeat session about the new changes to the GED test passing scores.  You can now pass each GED test with a score of 145 instead of the 150 that it was initially normed at.  This is being celebrated as “more inclusive” and “an opportunity”. 

I’ve been curious about the passing scores for the GED 2014 test since its creation.  On the 2002 series tests, teachers had access to the scoring charts for the official practice tests.  A passing score on the practice tests was usually between 60% and 70% correct if memory served.  I remember when I first learned how low the passing requirements were; I was a little shocked.  When I would tell my friends and family about my work, they too found it strange.

It’s gotten stranger.  As a part of the aforementioned webinar, I finally came across the information about the passing scores required after the original norming study when the passing score was set at 150.  Here they are reprinted from page 91 of the GED 2014 Technical Manual which is available at: http://www.gedtestingservice.com/uploads/files/01388a9ec8b540a6aa257f8727bc8830.pdf

Mathematical Reasoning: 43% of points required to score 150

Science: 45% of points required to score 150

Social Studies: 41% of points required to score 150

Reasoning Through Language Arts: 40% of points required to score 150

So we as educators are supposed to celebrate lowering the passing score of tests that already required less than 50% to pass?  Sure it’s more inclusive, but it’d be even more inclusive if we lowered the passing score to 125.  Of course, employers and colleges’ would stop accepting the GED then because it would be meaningless.  As it is, I know if I was an employer, I'd give you a nice hard sideways glance if you told me that you could pass the GED tests now by getting less than 40% of the points.  And if I was handing out diplomas at the county level, I'd feel very strange sending out diplomas for folks who got 1/3 of the answers correct.

My suspicion is that most teachers are totally in the dark on the actual passing scores for each test or how skewed the scale scores are.  Just last week I had a colleague tell me that they lowered the passing on the GED score from 75% to 72.5%, i.e. 150/200 to 145/200.  I had to try to gently explain that I thought those were scale scores and we don’t really know what a change of 5 points actually means in terms of number correct.

I've tried to keep my analysis here to a minimum and stick mostly to the facts.  I welcome further information, insights, or action on this topic.   

*Notes on edits.  I took out one sentence that was a leftover from an earlier draft.

Comments

After 80 views and 0 comments on my guarded comments from Tuesday about the strangeness of the GED tests' scoring system, let me try and express my worry as strongly as I can:

The 2014 GED tests were NOT designed to be High School Equivalency Tests at all.  They were designed and normed so a score of about 75% of the available points, a normal passing score for any assessment, would would yield a score of 170 and GED with Honors signifying College and Career Readiness.  Thus, they are best thought of as College and Career Readiness Tests.  However, since 2014 we've been using this same test to give out high school diplomas if people are able to get just more than 40% of the available points.  No teacher would administer a test with a 40% passing score in their classroom.  No employer would express confidence in a test that sets its passing score at 40%.  This is the educational equivalent of the DMV asking 15-year-olds to take the driver's license road test and awarding them their driver's permits so long as they don't get in an accident. 

The "good" news: I think most teachers and employers have no idea how the GED is actually scored. 

The bad news: These are numbers from the original Technical Manual of the 2014 GED test and the GED Testing Service just lowered the bar for high school equivalency by 5 scale score points.  At the risk of stretching my metaphor, now you can get in a fender bender and still get your permit.

Are we okay with this?

Source: GED Technical Manual page 91 http://www.gedtestingservice.com/uploads/files/01388a9ec8b540a6aa257f8727bc8830.pdf

Hi Josh, Thank you for sharing your concern here on LINCS. I am eager to hear from members on this issue. What do you all think about Josh's analysis? In most contexts, 40% would not be a passing score. Are there other relevant aspects of this issue to consider?

Cheers, Susan Finn Miller

Moderator, Assessment CoP

Josh,

I think you raise some good points here--especially now that the highest GED score level gets students college credit, I think it may be accurate to say that this is a "college ready" test. If you think about it, this means that if you score 175 or higher, you have shown that not only are you ready for college, but you have the skills and content knowledge of a student who has already taken a college-level class in that subject.

Imagine giving the AP exam to your HSE students--if they got 40-50% of the answers correct, you might be happy because it's a really hard test. If I went in to take a college-level science exam without ever having taken a college-level course and I got 40-50% of the points possible on the test, I would be pretty happy. I agree that if all the questions were written to gauge whether a student is at the HSE level, getting 40% of the answers correct would be concerning. On the other hand, since many of the items are actually meant to determine whether the student is at a college level, it would be irresponsible to require students to get a large number of these items correct in order to prove HSE equivalency.

As of this week, Pennsylvania is now going to begin accepting scores from both the HiSET and the GED. I looked through their website, and while I couldn't find current information about the percent of items test-takers need to answer correctly (in part because of the multiple iterations of the test--the technical document I found seemed to be from a previous version), it appears that at least in the beginning the passing rate was around 50-65% of possible points depending on the subject. I think that's for the minimum passing score in each subject, though--students need a slightly higher average score for the total battery, like the old GED.

I would be interested to see a comparison study--do students who take the GED, HiSET, and TASC get similar results? If you pass the HiSET, what are the odds you'll pass the GED or the TASC and vice versa?

In the end, I guess the only reason I would care what percent of questions students need to pass is because it contributes to the subjective sense of difficulty of the test. If a student knows that they have probably missed a large proportion of the questions, it may make them feel more pessimistic about passing. I do tell my students that some of the questions on the GED are designed to tell if they are college-ready, and so they can get those wrong and still be at the HSE level. Most of them appreciate knowing that.

I'm curious to hear what other people think about this issue--does it matter what percent of questions students need to get right?

 

Good morning... a few notes and information on some of the items raised in this thread:

The GED tests are scored using a very traditional scoring system used for most all large-scale assessments, based on Item Response Theory. This may not be a familiar system to some teachers, and it differs from what is used for classroom assessment, but it is a very commonly used system and has been for more than 50 years.

Thread comment: The 2014 GED tests were NOT designed to be High School Equivalency Tests at all.  They were designed and normed so a score of about 75% of the available points, a normal passing score for any assessment, would would yield a score of 170 and GED with Honors signifying College and Career Readiness. Thus, they are best thought of as College and Career Readiness Tests.

The tests were normed on a population of high school students with the passing standard set at a point where approximately 60% percent of students who received a high school diploma would have passed the GED test and 40% would have failed.  When the revised cut scores are taken into account, the percentages are approaching a point where 70% of students receiving high school diplomas would have passed the GED test and 30% would have failed, but continue on to earn a H.S. diploma. This pass/non-pass cut point has no reference to whether students are ready for college, but rather a reflection of the knowledge and skills of a typical graduating high school senior.

Thread Comment: However, since 2014 we've been using this same test to give out high school diplomas if people are able to get just more than 40% of the available points.  No teacher would administer a test with a 40% passing score in their classroom.  No employer would express confidence in a test that sets its passing score at 40%.  

These are assumptions rooted in how classroom assessments are done and extrapolating them to large-scale assessments and what someone would or would not expect or accept.  The reality is that most large scale assessments with multiple performance levels have similar percent-correct ratios to the GED test. And again, the tests are normed against the performance of graduating high school seniors. When discussing percentage of points earned to obtain a particular score, it must be remembered that the GED test now measures a broader range of content than in the past (more advanced problem solving and critical thinking skills, using evidence, etc.). Comparing percent of points earned on the 2002 Series with the same percent for the current test doesn't really provide any useful information, since the test construct between the two exams is very different. And, since the current test measures a broader range of content, it is only logical to assume that one might need a lower percentage correct to obtain a score at the minimum passing level. This is a characteristic of most large scale tests with multiple performance levels. 

 

Overall, we strive to ensure the GED test high school equivalency passing level and content is a reflection of the performance of today's typical graduating high school seniors. However, we all should encourage students to strive for the higher levels of performance, which is a huge benefit of the new GED test.  Yes, an adult can obtain a credential by demonstrating the minimum skill level that is comparable to a H.S. senior, but the test and the score report provide adults with a roadmap as to how they can demonstrate higher level skills -- and most will need those higher level skills to be successful in their lives and careers.

   

CT,

Thank you for the clarifications. 

I have some other questions for you.

1. Would it be fair to say that the GED test now measures: 1) high school equivalency, 2) readiness for college, and 3) readiness for careers?

Can you tells us:

2.  if/when the GEDTS anticipates having data needed to determine what cut scores indicate: 1) readiness for success in regular, not developmental, college courses, and 2) readiness for careers?

3. if the GEDTS has an operational definition of "college readiness" and "career readiness"?

Thanks.

David J. Rosen

djrosen123@gmail,com

 

CT,

Thanks for your response.  I view these forums as a safe place to try to think through issues of importance for our field.  And my last comment in this thread was obviously written hoping to generate more discussion.  Your response and some further thought really helped to clarify things for me.

Now, I think I can refine the two parts of my post that you commented on.  For the first, the statement that is most accurate about the design intent of the new GED Test is that it was not designed JUST as a High School Equivalency Exam.  It was designed to measure BOTH College Readiness as well as High School Equivalency.  The questions I’m interested in concern how this change in design affects our instruction, student advising, and the significance of the High School Equivalency Credential.  I don’t think I have nearly enough understanding of the design of the test at this point to even begin addressing any of those questions.  Before we can answer the, “Are we okay with this?” question from my previous post, we need to have a better idea of what exactly these passing scores mean.

My concerns about the 40%-45% passing scores, as you pointed out are rooted in my own assessment experience in the classroom and with the previous GED test. You emphasized how different this test is from both the previous GED tests and from classroom assessments.  Then you commented, “Since the current test measures a broader range of content, it is only logical to assume that one might need a lower percentage correct to obtain a minimum passing score.”  Of course, I understand this.  I was expecting a lower passing score, my reason for starting this thread was that 40% to 45% passing scores were MUCH lower than I would have expected.  Following that, you said, “This is characteristic of most large scale tests with multiple performance levels.”  I was unsure if you meant “this characteristic” to be interpreted simply as lower passing scores for lower performance levels or more broadly that passing scores in the 40%-45% range a common on other similar assessments.  In either case, what I think I’m suffering from is a lack of a frame of reference for the scoring system.  Since, as you’ve pointed out, our knowledge of classroom assessments and the 2002 GED test series doesn’t serve us well here, it would be very helpful to use another test that is a more appropriate frame of reference for understanding the 2014 GED.  I am not very familiar with the standardized testing landscape as a whole, so I don’t have another “large scale test with multiple performance levels” in mind.  Which test(s) would you suggest as a good frame(s) of reference for understanding the scoring system for the different performance levels of the 2014 GED Tests?  


Also, with the recent score changes, the passing scores are no longer at 40%-45%.  When does the GED Testing Service plan to update the 2014 GED Test Technical Manual to reflect the score changes?  And could an announcement be made here on LINCS when it’s available?

Hi Josh,

I checked in with GED Testing Service's Vice President of Assessment Services, Martin Kehe, for answers to your additional questions. Here's the scoop.

You asked for example assessments that might have a similar structure

Please refer to any state K-12 assessments that include multiple performance levels, the Smarter Balanced assessments, the PARCC assessments, NAEP and so on. All of these tests have multiple performance levels where attainment of a lower performance levels would be at a lower "percent correct" value.  These tests are not like classroom tests where there is a limited selection of content being assessed and one might expect someone to attain 70% or 80% of that content in order to pass.  Instead, the GED test and others like it sample content from a broader domain and use the responses to sort candidates into the various performance levels, based on the difficulty and complexity of the content.

People are sometimes concerned that someone might be able to "guess their way to proficiency."  Of course, this is always a possibility with any assessment.  However, the chances of that happening with a test like the GED test are extremely remote.  All of the responses to our multiple choice items are plausible, and the incorrect responses represent common mistakes made and misunderstandings held by test takers who lack proficiency.  In addition, the technology enhanced items represent a vast array of answer choice combinations, making it nearly impossible for someone lacking proficiency to guess a correct answer.

You asked about our technical manual

We currently have a technical supplement on our web site that addresses the performance level adjustments that were made.  The official technical manual is on a schedule of updates that are made once per year and encompass all of the additional technical information updates for that year.  The next update will be released this summer. You can imagine the difficulty in making frequent updates to the manual, as technical information about the assessment is being added to every day with each and every additional test taker.  In the interim between updates to the manual,  and significant changes are documented in special postings  of documentation to our web site,  which are later incorporated into the full manual as part of the next scheduled update.

I hope this helps!

Have a great day,

Jessica