Is it possible to type a contingency table direct into jamovi to do an indepedent samples test e.g.
use:
1 2 3 4
Moderate 15 32 18 5
Mildly severe 8 29 23 18
Severe 1 20 25 22
to produce the chi-square stats
Rather than the disaggregated data;
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Moderate
1 Mildly severe
1 Mildly severe
1 Mildly severe
1 Mildly severe
1 Mildly severe
1 Mildly severe
1 Mildly severe
1 Mildly severe
1 Severe
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Moderate
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Mildly severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Severe
2 Moderate
2 Moderate
2 Moderate
2 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Moderate
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Mildly severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
3 Severe
4 Moderate
4 Moderate
4 Moderate
4 Moderate
4 Moderate
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Mildly severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
4 Severe
conduct independent samples test direct from contingency tab
Re: conduct independent samples test direct from contingency
Hi, @vanagpau.
See the attached screenshot, it should answer your question. Cheers,
Maurizio
See the attached screenshot, it should answer your question. Cheers,
Maurizio
-
- Posts: 3
- Joined: Sun Jul 21, 2019 2:32 pm
Re: conduct independent samples test direct from contingency
I have a similar question regarding independent samples t-test [ISt] and anovas. Not starting a different topic. For both the ISt and ANOVA, the grouping variable is a must. This is the same in SPSS and other software offering drag and drop stats for non-statisticians such as I.
Trouble is this. I am a surgeon and most of the data I receive is usually in the following format. For eg.Groups A and B with a weight of pts at 0 and 6 months.
GBW0 GBW6 GAW0 GAW6
5.8 9 10.75 58
6.8 9.4 11.15 67
9.5 11.5 12.7 86
7.6 9.5 11 70
5.9 7.75 9.4 65
6.1 7.95 9.6 60
7 8.5 10.5 62
7.4 9 10.68 65
6.1 7.5 9.54 61
6.5 8.75 9.78 70
8 9.75 10.85 69.5
8.6 10.4 11.75 79
5.9 7.95 9.69 67
7 8.85 9.79 61
6 8.25 9.47 67
8.5 10.75 11.65 79.5
Now a paired sample is easy as I can just select the needed variables. But when I try to do independent samples [ e.g. GBW0 vs GAW0], I have to add a grouping factor. This means I have to re-write the columns and add new variables etc. It is tedious as datasets are longer [ the example is having 26 such variables and 300 or so observations, all numerical!]. Is there any way you could introduce an option to select two numerical columns for independent samples t-test [ and 3 or more for ANOVA] without the need of a grouping variable?
Bluesky is the only one I found which offers "two independent numeric variables ' option for the independent t-test. For ANOVA, I haven't come across this option anywhere. Nothing compares to ease and output clarity of Jamovi and if you could add this functionality, it would be awesome as this hinders almost all the clinical data for papers that I come across.
Trouble is this. I am a surgeon and most of the data I receive is usually in the following format. For eg.Groups A and B with a weight of pts at 0 and 6 months.
GBW0 GBW6 GAW0 GAW6
5.8 9 10.75 58
6.8 9.4 11.15 67
9.5 11.5 12.7 86
7.6 9.5 11 70
5.9 7.75 9.4 65
6.1 7.95 9.6 60
7 8.5 10.5 62
7.4 9 10.68 65
6.1 7.5 9.54 61
6.5 8.75 9.78 70
8 9.75 10.85 69.5
8.6 10.4 11.75 79
5.9 7.95 9.69 67
7 8.85 9.79 61
6 8.25 9.47 67
8.5 10.75 11.65 79.5
Now a paired sample is easy as I can just select the needed variables. But when I try to do independent samples [ e.g. GBW0 vs GAW0], I have to add a grouping factor. This means I have to re-write the columns and add new variables etc. It is tedious as datasets are longer [ the example is having 26 such variables and 300 or so observations, all numerical!]. Is there any way you could introduce an option to select two numerical columns for independent samples t-test [ and 3 or more for ANOVA] without the need of a grouping variable?
Bluesky is the only one I found which offers "two independent numeric variables ' option for the independent t-test. For ANOVA, I haven't come across this option anywhere. Nothing compares to ease and output clarity of Jamovi and if you could add this functionality, it would be awesome as this hinders almost all the clinical data for papers that I come across.
Re: conduct independent samples test direct from contingency
hi,
so in this data, each row represents two participants? that's a little awkward. there's some principle where independent measurements shouldn't inhabit the same row - but i appreciate this was not be your choice.
i would be reluctant to add support for this, just because it complicates the ttest UI, and it seems like more of a data rearrangement problem, and i would rather handle it that way. although i appreciate our data restructuring facilities are a bit scant at present.
on solution would be to use the computed variable system like so:
but i appreciate this is a little awkward. you have to manually increase the number of rows for example.
cheers
jonathon
so in this data, each row represents two participants? that's a little awkward. there's some principle where independent measurements shouldn't inhabit the same row - but i appreciate this was not be your choice.
i would be reluctant to add support for this, just because it complicates the ttest UI, and it seems like more of a data rearrangement problem, and i would rather handle it that way. although i appreciate our data restructuring facilities are a bit scant at present.
on solution would be to use the computed variable system like so:
but i appreciate this is a little awkward. you have to manually increase the number of rows for example.
cheers
jonathon
-
- Posts: 3
- Joined: Sun Jul 21, 2019 2:32 pm
Re: conduct independent samples test direct from contingency
Thanks. That helps provided one knows coding R! Anyways, is there any way to create a new categorical variable with specified values in rows. for eg rather than typing individual A and B in the group variable above, could you specify in syntax such as ROW ()< 61 then Group=A , ROW (61,123) Group = B etc? I know this sounds silly but would be a cool trick to know! Thanks again.
Re: conduct independent samples test direct from contingency
hey,
> That helps provided one knows coding R!
there's actually no coding in R here, it's just using the simple spreadsheet formulas. if you open that .omv file i attached above, you can double click on the column headings and it shows you the formula used to generate it (and you can adjust it if need be).
here's an article on computed variables: https://blog.jamovi.org/2017/11/28/jamovi-formulas.html
so in the attached .omv file, i think i used the formula `IF(ROW() <= 16, 'A', 'B')`
take the time to wrap your head around the computed variable system, because there's a lot of nifty things you can do with it.
cheers
jonathon
> That helps provided one knows coding R!
there's actually no coding in R here, it's just using the simple spreadsheet formulas. if you open that .omv file i attached above, you can double click on the column headings and it shows you the formula used to generate it (and you can adjust it if need be).
here's an article on computed variables: https://blog.jamovi.org/2017/11/28/jamovi-formulas.html
so in the attached .omv file, i think i used the formula `IF(ROW() <= 16, 'A', 'B')`
take the time to wrap your head around the computed variable system, because there's a lot of nifty things you can do with it.
cheers
jonathon