top of page

Big Changes in the Rubric for Significance Tests

Updated: Sep 12, 2023

One of the biggest reasons we love the AP Statistics Exam is that you never quite know what is going to show up (especially on the Investigative Task!). But there is one question you can always expect to see:


A SIGNIFICANCE TEST!


Every year. Since the the first exam in 1997. And in the many years since the first exam, the rubrics used to grade these significance tests have been very consistent. Until now. The 2019 AP Statistics Exam rubric for question #4 revealed some big changes.


1. The significance test is now graded in 3 sections

In the past, the rubrics for significance tests were always graded in 4 sections (which makes sense because they are graded out of 4 points total). The 2019 rubric shows that they will be graded in 3 sections (changes are in bold)


Section 1

  1. Hypotheses

  2. Identify parameters as proportions

  3. Clearly define parameters in context

  4. Identify test by name or formula

Section 2

  1. Random condition

  2. Normal condition

  3. 10% condition

  4. Test statistic

  5. P-value

Section 3

  1. Compare P-value to alpha

  2. Correct conclusion consistent with the alternative hypothesis

  3. Conclusion is in context

By grading the significance test in 3 sections, this opens up the opportunity for the College Board to ask a follow up question after the significance test. Check this post for some potential follow up questions that you might see.


2. Clearly define parameters in context (contextual subscripts are not enough!)

In the past, using a contextual subscript p(sub14) was sufficient for identifying parameters. No longer! In the 2019 rubric, the College Board is making it clear that these parameters must be defined in words. So here is what is expected:

Let p(sub14) represent the proportion of the population of kochia plants in the western United States that were resistant to glyphosate in 2014

3. 10% condition

While I teach this condition to my students every year, I always reveal at the end of the year that they won't need this for the AP Exam...and most for them stop writing it! Bummer for my 2019 students. So from here forward, if the context shows sampling without replacement, students should be checking the 10% condition.

it is reasonable to assume that the sample sizes are less than 10 percent of the respective population sizes.

Is it fair that College Board abruptly made these changes?

So some might say that it was quite rude of the College Board to drop these new changes without any heads-up. The reality is that they did this in a very fair way. They designed this particular rubric such that a student who didn't contextually define parameters or check the 10% condition could still get full credit.

  • Section 1 is scored Essentially correct (E) if the response satisfies components 1 and 4 AND at least one of the remaining components:

  • Section 2 is scored Essentially correct (E) if the response satisfies components 1 and 2 AND at least two of the remaining components:

Moving forward, I don't believe these rubrics will be written "softly" like this one. All of the components will be required for full credit. So start training your students NOW!

12,767 views12 comments

Recent Posts

See All

12 Comments


Luke Wilcox
Luke Wilcox
Mar 05, 2020

April...according to the rubric "The two-sample z-test for proportions is identified by name or formula." So students can simply name the procedure without stating the formula.

Like

april.clark
Mar 05, 2020

If students name the test, do students have to show the necessary formulas with plugged in values to calculate the test statistic or can they just record it from the calculator?

Like

Luke Wilcox
Luke Wilcox
Feb 11, 2020

Stacey...you are correct. Most of the time, an experiment utilizes volunteers rather than a random sample. In this case, there is no need to check the 10% condition.

Like

sstacey
Feb 10, 2020

"So from here forward, if the context shows sampling without replacement, students should be checking the 10% condition."


I am a relatively new AP Statistics teacher and still have a lot to learn so please correct me if I am wrong here but as for what I have seen in the past, most of the time we are sampling without replacement in the Free Response questions I have seen (on both ST's and CI's) on the exam. There are some examples in which there is an experiment in which subjects are randomly assigned to treatments. I believe that in the case of an experiment we would not need to check the 10% condition. I have tried to look for previous…

Like

Unknown member
Feb 10, 2020

"Is it fair that the College Board abruptly made these changes?" It's great that they structured last year's rubric in a way that did not penalize students who did what teachers have been teaching for years. What is not great is that, if I had not seen this post on FB, I don't know that I every would have heard about this. Perhaps it is in the new CED binder, but, while I look things up in there, I certainly haven't had the opportunity to read it cover to cover.

Thanks for posting - this is very helpful.

Like
bottom of page