Please or to access all these features

Add post

Watch this thread

Save thread

Start a new thread

Flip thread

Hide thread

My feed

Active Unanswered threads

Getting started FAQ's

Unanswered threads Acronyms Talk guidelines

Hide shortcut buttons

Talk

Chat

Join the discussion and chat with other Mumsnetters about everyday life, relationships and parenting.

Flip

Original poster

You brainy lot, is there any way of getting this info into a graph please?

24 replies

Brainfry · 31/01/2021 22:32

DD has to create a graph showing this attached information. I’ve racked my brains on how to show all of this info a graph form, it’s easy if she didn’t have to include the %age of respondents per salary bracket, I suggested a stacked bar chart.

Any ideas please?

You brainy lot, is there any way of getting this info into a graph please?

OP posts:

ragged · 31/01/2021 22:38

clustered columns?

Cluster by salary range.
One color for each column heading

I always want clustered stacked columns but doesn't seem to a way to do them in Excel & I'm too lazy to learn any other software.

Original poster

Brainfry · 31/01/2021 22:39

Ah that might work, she will have to hand draw it then.....thank you!

OP posts:

Camomila · 31/01/2021 22:39

I like grouped bar charts myself, I use SPSS but here's a hopefully useful link for how to do it in excel.

superuser.com/questions/1099503/creating-a-grouped-bar-chart-from-a-table-in-excel

Magpiecomplex · 31/01/2021 22:40

Two stacked bars, side by side, showing the number (not percentage) of respondents who do want an incentive in one, divided by income bracket, and the number who don't in the other?

Magpiecomplex · 31/01/2021 22:42

@Camomila I'm an R user myself - I don't like it but I must admit it runs rings around Excel!

Strike000 · 31/01/2021 22:45

Have the salary ranges on your x axis. Have two y axes. One y axis set from 0-30 and the other y axis set from 80-100. Put the percentage of respondents as bars. Put the other two as lines. ?

Camomila · 31/01/2021 22:49

I've used R a bit at uni, but my old laptop couldn't really handle it. Now I'm wfh I have a shiny new laptop but no need to do any statistics!

Original poster

Brainfry · 31/01/2021 22:50

Strike, thanks that could work! Cheers

OP posts:

ragged · 31/01/2021 22:55

Ah, just looked better, you can do stacked columns for the fin/non-fina incentive info (since always adds up to 100%), and put the number of respondents in the x-axis label to complete the information presented,. Very simple. eg:

£20k, n=30
£20-£30k, n=19 etc. for the x-axis labels

Dadalus · 31/01/2021 22:55

I can imagine it as a pie chart with segments being salary brackets, and each segment having an inner and outer ring whose sizes are based on the proportion who wanted the financial incentive or not. I think that would be easy to understand visually, but no idea if that's possible in Excel.

Rina66 · 31/01/2021 23:00

Sorry to be of no help on the graph but should the % of respondents add up to 100?

MixedUpFiles · 31/01/2021 23:06

Put the salary ranges on the x axis. Create columns on the y axis that represent the population percentage. Then superimpose a line graph that shows the percentage of each salary range from the third column of data. You don’t need to do anything with the 4th because it is just 100-the 3rd.

ErrolTheDragon · 31/01/2021 23:06

How about : for each salary range a column of full height split by the financial/nonfinancial incentive (so the 66 is narrowest.

I think this representation ends up with the areas of blue/red proportional to the number who want financial/none.

(Because the numbers wanting non financial are much smaller she might want to start the y axis from maybe 75 not zero.)

LemonadeFromLemons · 31/01/2021 23:27

I’d have the salary ranges on the x axis. Do a combo chart with the percentage wanting/not wanting fin incentive as a stacked bar and the percentage of respondents as a line.

ErrolTheDragon · 31/01/2021 23:27

Not sure I described that well... stacked bar chart with the width of the bars proportional to the percentage of respondents for that bar.

ErrolTheDragon · 31/01/2021 23:31

Or simpler, just plot the percentage who want non-financial with proportional width columns.

visitorfromtheplanetzog · 31/01/2021 23:35

Erm... I hope you don't mind me mentioning it. There's one piece of information missing from those columns of data.

The actual number of people in each salary range. Without that information, will interpretation of the results be meaningful? There might be hundreds of people in the lower salary brackets and only a handful at the top.

ErrolTheDragon · 31/01/2021 23:43

@visitorfromtheplanetzog

Erm... I hope you don't mind me mentioning it. There's one piece of information missing from those columns of data.

That might matter if there was a systematic bias in how likely people in each group were to respond, it's probably meant to reflect the numbers in each group.

SpamIAm · 31/01/2021 23:45

Am I the only person who wouldn't put percentage of respondents in each age bracket on the same graph as the other data then..?

ErrolTheDragon · 31/01/2021 23:48

@SpamIAm

Am I the only person who wouldn't put percentage of respondents in each age bracket on the same graph as the other data then..?

If the question is 'create a graph', then i assume one graph is what's wanted. I think my suggestion would do it quite efficiently and obvious to interpret.

Magpiecomplex · 01/02/2021 09:57

I like your approach Errol. My friendly statistician doesn't recommend superimposed graphs with two different y axes as they tend to get messy and difficult to interpret, if not misleading.

And I agree with Visitor that number of respondents would be helpful, since percentages is losing the context here.

Brain - your daughter could calculate the percentages as a proportion of the whole, so the 100% of people

cheeseybean · 01/02/2021 10:01

What's jumping out to me is where would someone earning £30500 go?

The < and - signs in the first column make for inaccurate bounds

cheeseybean · 01/02/2021 10:04

< 20k
20k

ErrolTheDragon · 01/02/2021 10:15

@cheeseybean

What's jumping out to me is where would someone earning £30500 go?

The < and - signs in the first column make for inaccurate bounds

Yes, the data ranges aren't correctly defined but I think for the purpose of this exercise that's maybe a bit pedantic.

Flip