Possible Bug when Inserting Checkboxes

When inserting checkboxes, the empty boxes don’t automatically register as FALSE the way checked boxes would register as TRUE. You have to check and uncheck each box for it to be registered as FALSE in the formula bar. As a result, there are output errors when using certain functions such as the AVERAGEA function. That doesn’t seem to be by design but rather a bug, right?
cc: @Adam_Steinfurth

Hello @Carlos_Kay , Welcome to the community.

The checkbox today behaves pretty much like a cell filled booelan value - TRUE, FALSE or blank. But the problem is that there is no visual distinction between FALSE and blank. Also once its checked and unchecked, it reverts to FALSE and not blank.

So, I agree there is some potentially confusing behaviour here.

Allow us to brainstorm internally; I’ll update this thread.

Uhm, my understanding is that as long as there is a visual checkbox (checked or unchecked) then blank isn’t an option. For context, I’m basing my claim on checkbox behaviour in Google Sheets. Thank you.

“… as long as there is a visual checkbox (checked or unchecked) then blank isn’t an option”
Yes and probably no.

I’m aware of the Google Sheets behaviour - and I think it makes sense for a range of check-box cells. Google Sheets also de-formats ( takes out the checkboxes themselves) if you select a range and press Delete - We don’t do that and I like our behaviour better.

But at a Column level setting, it probably makes sense to have a blank ( or uninitialized) state. Imagine you are collecting Form entries from people in your company. And a there is a Opt-in/out ( for some activity) Column checkbox.
As your form entries start arriving as rows, they’ll be either on checked or un-checked

But for the blank rows in the sheet towards the bottom which do not correspond to a form-entry yet, the check-box column needs to have a blank state. Else if you want to calculate the % of people who have “opted in” using either =AVERAGE(H5:H) or… =AVERAGEA([Opted_In]) (soon to be released Named Columns support in Formulas ), the blank ones would contribute towards FALSE and result in incorrect results.

I think the solution may be a distinct visual representation for the blank/uninitialized checkbox cells, at least for checkbox Columns.

Thank you. I see what you’re saying. Maybe the distinction in visual representation is the way to go.

Missed out on updating this thread. Logged defect #004015 to track this.