KA P/S Passage Question

[betterfuture]

Could someone help me look at this question. I can't understand how they concluded the relationship to be "strong." Shouldn't there be a numerical value to indicate if something is strong or not? The link for the passage is:

https://www.khanacademy.org/test-pr...-tut/e/counting-systems-and-the-piraha-tribe-

---

Members don't see this ad.

 

[theonlytycrane]

I agree with what you're saying, but take the KA passages with a grain of salt. There are a lot of mistakes and not fully fleshed out questions. In this specific question, choices b-d make statements that don't depend on data from the figure at all.

 

[betterfuture]

Thank God someone agrees with me about these graphs! So I am trying to get better at data/graph interpretation. If the graph/data presents information to you about something, you should only make conclusions based on the graph, correct?

For instance, if there was a graph showing how attending churches on a weekly basis reduces murder rates, we shouldn't conclude from the data something like "Going to church is the best method to reduce murder" or "Attending social activities can decrease the likelihood of engaging in criminal activities", right? Because these things aren't measured or directly shown from the data, so we can't make conclusions about something other than attending church and murder rates. Right?

 

[theonlytycrane]

Yep! If it's like: p < .001

(a) Attending church is correlated with lower murder rates
(b) No murderers go to church
(c) If everyone went to church there would be no murderers
(d) Correlation does not imply causation

(b) and (c) don't even relate to the study you mentioned above. (d) is a true statement, but even more unrelated. Practice the process of elimination techniques

 

[betterfuture]

Yeah. I can see from the answer choices, how all of them except (a) are wrong answer choices. By the way, I am trying to get the p value down.

So if someone conducts an experiment about murder rates and going to church and then they give you a p value.

• If p<0.05 (p value is lower than 0.05) that means the data is significantly different. Which means, the null hypothesis must go, and the alternate hypothesis must hold true? Correct?

 

[theonlytycrane]

Yep. And if p >= .05 then there is no statistical significance between the data.

 

[betterfuture]

If the p is low, the null must go!

Btw, I should add, the low p value does not add support for the alternate hypothesis. Only that it rejects the null hypothesis. I thought I should clear that up!