Search This Blog

Monday, October 29, 2018

QUIZ TOMORROW!


Today's Class Recap:

Monday Homework: 

1.) Correct your free response from class using the answers below.
2.) Start counting your stamps--you can turn them in tomorrow, as we won't have a stamp tomorrow (quiz to start class)
3.) Start studying your vocab!
  • Here's what you need to score your free response from class:
    • Use a different color pen/pencil to make any corrections or add anything that was missing! This is review for our quiz!
    • Study more! You can use the key from Friday's blog post to check your Thursday hw responses as well!
    • Classwork Key:
      • a.) predicted % weeds killed =  - 20.5893+ 24.3929(# tsp. weed killer)
        • Common mistake: remember to define your variables!
      • b.) Slope:
        • According to the LSRL, for each additional 1 teaspoon of weed killer (used), the predicted % of weeds kills increases by roughly 24.3929%.
      • b.) Y Intercept
        • Based on the model, if no weed killer is used (# of teaspoons of weed killer = 0), the predicted % of weeds killed is -20.5893%. This is unrealistic because the % of weeds killed cannot be negative.
          • *Someone might say this is realistic, but since it isn't immediately clear you'd still want to comment on this--one might say, "This is realistic, as a negative % of weeds killed suggests more weeds grew."
        • Common mistakes: 
          • For both of these interpretations be sure to pay attention to context! Write about the "% of weeds killed" and the "# of teaspoons of weed killer used," not just "% killed" and "# tsp.!" 
          • Remember to use "Based on the data/according to the model" AND "predicted"
      • c.) Here's the thought process....
        • First, we have to think about how the only given information (a residual plot) connects to whether a prediction is too large/too small....
        • Then, we have to notice that the horizontal axis on the residual plot is "predicted," so we can't look for 2.6 on that axis--we need a predicted value
        • Predicted % killed = -20.5893 + 24.3929(2.6) = 42.83
        • Now, we can't find the exact residual for a predicted value of 42.83, but we can see the residuals around this value--based on this info/pattern, would you expect the residual for a prediction of 42.83 to be positive or negative? Why?
        • Finally, now we know if we expect the residual to be positive or negative, so let's connect that to the original question--is the prediction expected to be too large or too small?
        • Here's a sample response:
          • "I would expect the predicted % of weeds killed for 2.6 teaspoons to be too large. Based on the pattern/shape of the residual plot, I would expect the residual for 42.83% killed to be negative, which means my prediction is expected to be an overestimate (too large). OR....
          • "I would expect this prediction to be too large. We predict 42.8% of weeds to be killed with 2.6 tsp of weed killer, and the residuals around 42.8% are all negative, so I'd expect the residual for this prediction to be negative--which means the prediction is too large."
        • What is required for your written response:
          • Do you expect the prediction to be too large or too small?
          • Why?
            • Reference predicted value is 42.8%
            • Comment on the fact that we EXPECT the residual for this prediction to be negative
            • Comment on how we know this--because the residuals around 42.8% are negative or because of the pattern in the residual plot suggests this residual would be negative
          • Context
      • d.) A linear model is not appropriate for the # teaspoons v. % of weeds killed because the residual plot has a clear pattern/curve. 
        • Common mistake: be sure to say there is a pattern IN THE RESIDUAL PLOT
        • If a student states "there is a pattern" without referencing "the residual plot" he/she would earn a 0--because if there is a pattern in the scatterplot, that's good!
          • If a linear model were appropriate then we would want to see a pattern in the scatterplot, we'd want to see "roughly linear' shape!
      • e.) Based on the data, 97.2% of the variability in the % of weeds killed can be explained by variability in the number of teaspoons of weed killer used.
        • Common mistake: if you do not include the "variability in" number of teaspoons or % weed killed then you have written a false statement and your response is incorrect (no points)
      • Extra Question: Find the correlation and describe the strength and direction.
        • r = sqrt(0.972) = +/- 0.9859; r is positive since the slope is positive
        • r = + 0.9859
          • The correlation between the # of teaspoons of weed killer and the % of weeds killed is strong and positive. 
STAMPS ARE DUE WEDNESDAY!
  • Gather all your stamps--you can leave them on the full paper, or rip off the little stamp, whatever is easier
    • You will get all of these back, so if you have notes on a paper with a stamp, don't worry, you'll get it back
    • If you count your stamps in class you will be given a zero--this is your responsibility OUTSIDE the classroom
  • Count all your stamps--double check!
    • I will choose 5 random names from each class and double check those students' stamps; if your total is incorrect, you will earn a score of 0
  • Fasten all your stamps--put them in a ziploc bag, put them in an envelope, staple them all together, or maybe glue/tape all the cut out stamps to one paper--whatever works best for you
  • Write your name and the total on the front of your stamps
  • Turn this in by Wednesday!
  • Stamps count as a classwork grade!
    • For each class I will find Q3 and use this to determine how many points the stamp grade is out of, so I can't tell you how many you need....
    • It doesn't really matter how many you need anyway, you just have to turn in all your stamps!
Chapter 8 Vocab Quiz Wednesday!

  • Residual (what does it measure? how is it calculated? how is it represented visually on a scaterplot?)
  • Actual Value (y): based on given data or the point in a scatterplot
  • Predicted Value (y-hat): generated by substituting into the LSRL equation or looking at the line of best fit
  • Interpret Slope
  • Interpret Y Intercept
  • Coefficient of Determination = R^2
  • Interpret R^2
  • Correlation
  • Underestimate (connect to residuals)
  • Overestimate (connect to residuals)
  • Linear Model
  • Least Squares Regression Line (LSRL)
  • "Is a linear model appropriate?" (what do we check?)
  • Explanatory Variable
  • Response Variable
  • Scatter plot
  • Lurking Variable
Posted by at

No comments:

Post a Comment

Subscribe to: Post Comments (Atom)