Computing Mean Score - ignoring certain values without excluding participant/row altogether

[Emily1313]

I’m working with a spreadsheet containing ‘99’ values – these are not missing values, but rather, where participants have chosen that a particular item in a scale is not applicable to them.

I’ve been asked to create Mean scores for several scales including some of these 99 (not applicable) values. At present, the 99s are treated as numbers and so obviously impact the mean – but if I change them to be missing values, then a mean score will not appear at all for that participant. Instead, what I’d like is for Jamovi to calculate the mean score for these participants excluding any 99s but still including all other values for that scale.

Is there a workaround for this? I struggle with correct language for forumulas for Jamovi – i.e. how to word things in the formula box – and struggle to find any help on specific issues such as these.

Please let me know if this doesn't quite make sense and I can clarify

[jonathon]

i *would* treat the 99 as a missing value, and then use the ignore_missing=1 argument to the MEAN() function.

cheers

[Emily1313]

Hi Jonathon, thanks for your suggestion. Just a couple of follow-up questions:
a) I'm not sure how this formula would work. Would I need to change all the 99s to 1s for it to work? Would it ensure that a mean based on all other scores and excluding the 99s (i.e. not substituting them for 0)?
b) There are also actual missing values in the data - is there a way to still have no score for the missing values, but a score for the 99 values?

[Emily1313]

In other words:
1. What does the formula "ignore_missing=1" actually do to the data? Is the mean calculated without the missing value, or is the missing value substituted for a value of 1? Sorry, as mentioned I don't understand the language of this software. I'm very new to data analysis and Jamovi is the first software I've ever used - so I'm not familiar with this kind of language generally.
2. Is there some way I can calculate the mean score so that it still provides a mean for the 99 scores but does not provide a mean score for actual missing data?

[MAgojam]

In other words:
1. What does the formula "ignore_missing=1" actually do to the data? ....

Do you think this screenshot helps you understand?

screenshot.png (127.6 KiB) Viewed 1794 times

Cheers, Maurizio

[jonathon]

so in statistics, we treat a missing value as "don't know" ... so if you take the mean of three values, and one of the values is "don't know", then you "don't know" the mean of those three values either.

ignore_missing=1 is where you say to ignore the missing value, so if one of three values is missing, it will give you the mean of the remaining two.

but take a look at mz's screenshot above ... see how his 99's are grey and a bit smaller? this indicates that they are being treated as missing values. if you double click on a column header, you'll find a spot where you can designate missing values.

2. Is there some way I can calculate the mean score so that it still provides a mean for the 99 scores but does not provide a mean score for actual missing data?

so if i understand you correctly, you might have two columns. the first column might have the values 1,2,3,missing,5 and the second column 3,4,5,6,99 ... and you want to compute the means 2,3,4,missing,5. (i.e. row 4 has a mean of 'missing' because one of it's values is missing ... but row 5 has a mean of 5, because the 99 is "ignored").

if that's what you want ... i'm not sure i can think of a way to do this ... maybe someone else can suggest something.

jonathon