First steps with my study

[gvt]

Hello,

now I´m doing my first study and I have some questions. I checked that all data for my variables are non parametric.

In my study I measure the power of bike 6 times in 6 different settings, using 3 differents methods and in 2 conditions, in and out. I have the data with this columns (Settings, Condition, Method A, Method B, Method C).

I need to find differences in:

1- Condition
2- Settings per method
3- Intra setting
4- Methods
5- Methods per condition

I have the data correctly imported or is required to transform it?

[reason180]

I would need more information about the research design in order to provide useful commentary.

-Is there just one bike, or several?
-Is "bike" the unit of analysis?

Assuming that "bike" is the unit of analysis:
-Does "method" vary within or between (across) bikes?
-Does "condition" vary within or between bikes?

[gvt]

Only one bike.
The unit of analysis is power.

Hope this infographic help.

Captura de pantalla 2023-09-01 a las 23.36.08.png (270.48 KiB) Viewed 2547 times

[reason180]

I think I would need to see a spreadsheet of either the actual data, or similar, made-up data.

[gvt]

Captura de pantalla 2023-09-02 a las 19.35.33.png (66.47 KiB) Viewed 2505 times

Only one cyclist, doing 72 trials, 6 in each 6 different settings, in 2 conditions (in and out), and we calculate power with 3 different methods.

[reason180]

It seems to me that your unit of analysis is "Trial" ("Test") rather than the traditional, "Subject" (i.e., "participant"). Setting varies between trials, as does Condition. Power-Type (A, B, C) varies within-trials and thus constitutes a repeated-measures factor. The data set is taken to be a good-enough approximation to a random sample the population of responses this particular biker might have have produced.

One commonly used method to analyze this type of data set is Repeated-Measures ANOVA. You data are arranged properly for conducting Repeated-Measures ANOVA. For example, see https://danawanzer.github.io/stats-with ... anova.html

[gvt]

ok! I did with this. But I have some doubts:

1- To calculate Kolmogorov is required to do or choose T student, but theoric say that before nothing is required to check normality no?

2- I did reapeted measure ANOVA (Friedman) and checked that we have significative differences. Choose with the same step to show Durbin-Conoves posthocs to see where are the differences between A, B and C power model. But to check significative differences for power model "in VS out" is correct to choose one way ANOVA (Kruskal-Wallis)? I did on the same step compatives pairs DSCF.

3- I did the same step, one way ANOVA (Kruskal-Wallis) with compatives pairs DSCF to check differences for power models for "settings". It´s correct?

[reason180]

Neither a Student's t test nor a one-factor ANOVA seems appropriate to your data set.

You have a multifactor ANOVA.

The design is a 2 (Condition: out or in) by 6 (Setting: one through six), by 3 (Power Type: A, B, or C) repeated measures ANOVA. Condition and Setting are non-repeated-measures factors. Power Type is a repeated-measures factor.

jamovi's repeated-measures routine includes options to assess violation of the equal variance assumption and the normality assumption, and includes options for post-hoc tests of differences between means.

[gvt]

Then I have some doubts with it. Jamovi only show 2 options for non parametric ANOVA, one way Kruskal-Wallis, and reapeted measures Friedman...

[reason180]

Sorry, I don't follow your description. You say that "I checked that all data for my variables are non parametric." However, being "parametric" or "non-parametric" isn't a property of a variable, it's a property of the analytic approach you choose to take. As described, I don't think your data or research questions are compatible with any standard non-parametric approach. For example, I don't think there's any widely-accepted method to examine an "ANOVA interaction" non-parametrically.