vignette edits

[jgabry]

I realized we didn't have an open issue for this. There are still a few remaining comments on the vignette pull request that need to be addressed even though it was merged. These include at least the following (links to the relevant review comments in parentheses):

• clarifying how the weights are used (link)
• updating the survey package syntax used (link)
• rewording the section on OOP and functions vs methods (link)
• deciding which terminology to use (link1, link2)

There may be other items or I may have missed when some of these things were already done, so I'm not sure how accurate the list is. Feel free to edit it.

[lauken13]

Terminology wise, what terms are still undecided?

I think we should use workflow over grammar, but I couldn't find other concerns? (the link is just to the vignette comments so it was hard to find).

[jgabry]

(the link is just to the vignette comments so it was hard to find).

weird, for me the links take me to the particular comments about the terminology in different spots in the vignette. not sure what's up with that. the terminology brainstorming was in reference to the following:

"Get the estimated value based on the post-stratification dataset."

We need to say what "estimated value" is referring to, but we didn't decide on terminology for that. In the particular case of a binary response variable we could say "estimated probability of the outcome variable" or something similar to that.

6. Get the estimated value of the response variable's probability based on the post-stratification dataset.
7. Aggregate the estimated value to some level of instance, for example for state/province level.
8. Visualize the estimated probability.

In 6 it looks like we're actually already using "response variable's probability" so we should decide if we like that and if so also use it in the other case.

In 7 we need to decide what to replace "level of instance" with (unless that's a real expression I'm not familiar with, in which case maybe I just need to learn it)

[lauken13]

Hmm I use outcome a lot. So it would be

Predict the probability of the probability of the outcome based on the post-stratification dataset
Estimate the outcome to some level, for instances stat/province level
Visualize the probability of the outcome.

For non-binary variable language it would just become "the mean of the outcome" etc. What do you think?

[lauken13]

@Dewi-Amaliah can you please change to probability of the outcome? Also, these three sentences read a little clunky, suggest the following:

6. Predict the probability of the outcome using the post-stratification dataset.
7. Estimate the probability of the outcome at a level of interest, for example, state/province level.
8. Visualize the probability of the outcome

[Dewi-Amaliah]

Hi Lauren,

I have changed the wording as you mentioned above.