Summarize to find the average of different factors

[aabbccwyt]

Hi. Another question about the polling analysis. 2016 General Election Polls: Trump vs. Clinton vs. Johnson vs. Stein | RealClearPolling
I'm now trying to calculate the average margin of error of each polling websites to find out the confidence that a poll result would reflect the result. Here is what I have.

poll %>%
select(Poll, MoE) %>%
group_by(Poll) %>%
summarise(avg = mean(.data[[poll]], na.rm = TRUE))

And the error comes out:

Error: Must subset the data pronoun with a string

This should be pretty simple but I kind of stuck in summarise part. Thank you in advance for any hints.

[nirgrahamuk]

just guessing because I can't run your code on your data.
but does this work better?

poll %>%
  select(Poll, MoE) %>%
  group_by(Poll) %>%
  summarise(avg = mean(MoE, na.rm = TRUE))

[aabbccwyt]

@nirgrahamuk

poll %>% select(Poll, MoE) %>% group_by(Poll) %>% summarise(avg = mean(MoE, na.rm = TRUE))

I appreciate your reply.
It doesn't work and ends up with another error

argument is not numeric or logical: returning NA

The data is similar to the first polling data in the link.

[nirgrahamuk]

well, if you scraped the table, then MoE values like -- will have caused your MoE to be a character type, for which there is no concept of mean(). therefore consider cleaning your character variable and converting it to a number. readr::parse_number() may be useful