Hi,
I’d like to check whether there’s a possible mistake in the Graded Lab – Exercise 4 (Cut proportions), specifically Step 4.
What I noticed:
In cell E15, the provided formula for calculating the p-value is:
=NORM.S.DIST(D15)
But from what I understand, this appears to be a one-tailed test formula, and we’re performing a two-tailed test here — since the alternative hypothesis is:
H₁: p ≠ 0.67
Could it be that the correct formula should be:
=2 * (1 - NORM.DIST(ABS(D15), 0, 1, TRUE))
This would give a two-tailed p-value, which seems to match the hypothesis we’re testing.
My current values:
- Sample proportion (p̂):
0.6632
- Hypothesized proportion (p₀):
0.67
- Test statistic (z):
-1.0226
p-value using =NORM.S.DIST(D15)
: 0.1533
p-value using the adjusted formula: 0.3065
I might be wrong, but this looks like it could be an error in the lab. Could you confirm whether this is indeed a two-tailed test and if the p-value formula should be adjusted?
Thanks again for this excellent specialization. I really appreciate the clarity of the teaching and want to help fine-tune any early glitches!
Christophe
Hi,
You’re absolutely right to question the formula — great critical thinking!
Yes, this is a two-tailed test, since the alternative hypothesis states that the proportion is not equal to 0.67 (H₁: p ≠ 0.67). In a two-tailed test, we need to consider both extremes of the distribution, so the p-value should reflect the area in both tails.
The formula currently used in cell E15
(=NORM.S.DIST(D15)
) only calculates the left-tail probability, which would be appropriate for a one-tailed test. However, for a two-tailed test, the correct formula should be:
=2 * (1 - NORM.S.DIST(ABS(D15)))
This adjustment gives us the proper two-tailed p-value, based on the absolute value of the test statistic.
That said, even with the corrected formula, the decision remains the same. The resulting p-value (~0.31) is still much higher than the significance level of 0.01, so we fail to reject the null hypothesis — there’s not enough evidence to conclude that the proportion of ‘Premium’ or ‘Ideal’ cut diamonds differs significantly from 67%.
Thanks for spotting this — I guess it was easy to miss this since it was pre-filled cell.
1 Like
Thank you @lukmanaj for your answer!
I really appreciate the clarification.
I personally enjoy checking for overall consistency in results, and I had a feeling something was slightly off with the value in that particular cell — even if the final interpretation (fail to reject the null hypothesis) remains unchanged.
That said, I believe it’s important to maintain rigor in statistical reporting, especially in a graded context.
I strongly encourage the @Community-Team to consider updating this exercise to include the correct two-tailed p-value formula in Exercise 4 – Step 4, as it currently reflects a one-tailed test structure (=NORM.S.DIST(D15)
) which doesn’t match the stated hypothesis (H₁: p ≠ 0.67
).
Christophe
Hi Christophe! Thank you for the feedback! We’ll look into it this week and update the spreadsheet if necessary.
1 Like