Search This Blog

Friday, January 24, 2020

Midterm Exam Reflect (and Tuesday's HW)

First off, here's Tuesday night's homework if you want to get ahead:

  • For #'s 11 and 12 you just have to provide the name of the sampling method (no need for a sentence)
  • For 13a name the method (again)
  • For 13b I need some WELL WRITTEN, NOT LAZY sentences here--think realistically--what might be a problem with this sampling design?
    • Remember, the goal in taking any sample is to collect a sample that represents the population
  • For 15a name the sample
  • For 15b write me a sentence, thinking similarly to 13b (getting us to consider bias)
  • For 13c:
    • Sampling Frame = all people/things that could have been sampled
    • Sample = the people/things that we actually collected data from
    • Population = the "whole group" of people/things we are trying to draw conclusions about based on our sample
  • For 27a you need to write a detailed procedure
    • This is extra important because many of us underperformed on the midterm in describing how to randomly assign treatments; this SRS is essentially the same way you randomly assign treatments
  • For 27b you need to write another procedure!
  • Always check your work!
Alright, now let's get into that midterm and look at lots of data:

How was the exam scored? 
  • I used historical data to determine "cut scores"--what raw % corresponded to what AP score.
  • Based on these cut scores I determined if you would be projected to earn a 1, 2, 3, 4, or 5
    • 1 = F
    • 2 = D
    • 3 = C
    • 4 = B
    • 5 = A
  • Then, I listed all scores in order to start to determine what would end up in PowerSchool (see the link).
  • Lastly, if your score is a 79, you're projected to get a 3 but you know you're close to a 4. If you earned a 90 you earned a projected 5 but you know you were close to a 4, and so on.
  • Here's all the data, cut scores, and a conversion chart:
Want to start off Q3 with 3 extra stamps?
  1. Open the 2019-20 Midterm Scoring and Data Link
  2. Calculate and interpret the z-score for your RAW SCORE.
    • You'll need to work backwards from the conversion chart; find your curved score in PowerSchool, then use the table (colored in green/yellow/red) to find your raw score
    • Use the table of summary statistics for the mean and standard deviation for calculating the zscore
  3. You ONLY EARN STAMPS IF you CORRECTLY CALCULATE and INTERPRET the z-score (in context). 


My Takeaways:
  • WE HAVE WORK TO DO.
  • To me, there is clear evidence that we did not study/prepare well for this test. The free response, in particular, was not a hard section in comparison to what we’ve done in class. Many of these questions were directly based on what we learned in class; we HAVE to get those right.
  • Please read this!


No comments: