it's been a while ... help choosing correct test

[NBrand]

Hello, it has unfortunately been a long time since I've had the chance to use my college stats skills and it turns out I'm SUPER rusty.
I need help choosing the correct test to accurately compare some data.
I want to compare attendance between two time periods.
ex. attendance = 14 during a 6.5 week period, while in a separate period of only 6.25 weeks, attendance was 7 times
Any suggestions?

thanks in advance!

 

[hlsmith]

Please describe data in more detail. Does 6.25 represent an average for a group? How many people are in the group? Are the exact same people in both periods? Etc.!

 

[NBrand]

This is one person's attendance during two periods of time.
During the first period of time which is 6.5 weeks, the individual attended 14 times.
During the second period of time which is 6.25 weeks, the individual attended 7 times.

 

[gw1]

Hey, what's the research question?

 

[NBrand]

I'm just looking for a more nuanced way to compare the same individual's attendance between two unequal periods of time than simply just calculating the averages of both.
I seem to remember that there's a way (maybe this isn't the right word) normalize the data because the time periods are different.

 

[Karabiner]

You standardize them if you calculate number of attendances / time.

You could additionally calculate largest time gap between 2 attendances, if that contains something of interest.

Do you really want to perform a statistical test of significance? That could be a one-sample Chi² test.
Using your example: you observed 21 attendances. According to the respective length of the periods,
your Null hypothesis expects that 50.98% of attendances were in the first period, an 49.02% in the
second period.

With kind regards

Karabiner

 

[NBrand]

You standardize them if you calculate number of attendances / time.

You could additionally calculate largest time gap between 2 attendances, if that contains something of interest.

Do you really want to perform a statistical test of significance? That could be a one-sample Chi² test.
Using your example: you observed 21 attendances. According to the respective length of the periods,
your Null hypothesis expects that 50.98% of attendances were in the first period, an 49.02% in the
second period.

With kind regards

Karabiner

You know ... I was leaning toward

.
Thank you. I'll play around with that.

 

[gw1]

You standardize them if you calculate number of attendances / time.

You could additionally calculate largest time gap between 2 attendances, if that contains something of interest.

Do you really want to perform a statistical test of significance? That could be a one-sample Chi² test.
Using your example: you observed 21 attendances. According to the respective length of the periods,
your Null hypothesis expects that 50.98% of attendances were in the first period, an 49.02% in the
second period.

With kind regards

Karabiner

? Didn't the individual visit ~48% less often in the shorter period
((6.25/7) - (6.5/14)) / (6.25/7).

 

[gw1]

That could be a one-sample Chi² test.

You know ... I was leaning toward

.

How can Chi² be suggested when no explicit response (dependent variable) is stated? No dependent variable is explicitly stated, no research question is stated, no independent variables are explicitly stated. Plus the sample size is also 1? How can any meaningful statistical analysis be done with sample size 1 and no explicit response or research question?

 

[Karabiner]

How can Chi² be suggested when no explicit response (dependent variable) is stated?

Well, if my suggestion isn't able to answer the OP's research question, he can just ignore it.

Plus the sample size is also 1?

The sample size in the example was 21 (events).

With kind regards

Karabiner

 

[priyac1987]

If all the data is numeric in both the datasets (compliant and non-compliant) then you must apply T-Test (https://www.projectguru.in/how-to-perform-t-test-statistics-in-spss/)

If the data is categorical then you must apply chi square test (https://www.projectguru.in/chi-square-test-spss/)