Map series of repeated measures to calendar using start date

[ladyc]

Howdy,

I have a dataset with the following variables:

Subject_ID
StartDate
W1
W2
W3
...
W10

Each individual (Subject_ID) was measured weekly for 10 weeks (resulting in 10 repeated measurements, the W1-W10 variables). The subjects were enrolled into the study on a rolling basis, with the date of their first measurement saved as StartDate. Thus, W3 was collected on StartDate + 2 weeks. We have an interest in quickly seeing whether some odd blips seen in the repeated measures are related to outside factors that came and went over time. Is there an easy way to transform/manipulate this data with SPSS so we can graph the W1-W10 data for all of the subjects against the actual date corresponding to each measurement? I could do it manually in excel, etc, but it would take quite a while. I have coding experience, but I'm new to SPSS syntax. I'm also not certain there isn't a way to do it through the GUI that I just haven't figured out. Does anyone have any ideas about how to go about this?

Erin

[Bruce Weaver]

Howdy.  :-|

If you restructure your data to a LONG format (with multiple rows per Subject_ID), you could estimate your model via MIXED, and include number of days from the minimum StartDate value (or some other convenient variable reflecting actual date) in your analysis.  Singer & Willett's book is a great resource for these types of models--and you can see worked examples from it here:

http://www.ats.ucla.edu/stat/examples/alda/

HTH.

ladyc wrote

Howdy,

I have a dataset with the following variables:

Subject_ID
StartDate
W1
W2
W3
...
W10

Each individual (Subject_ID) was measured weekly for 10 weeks (resulting in 10 repeated measurements, the W1-W10 variables). The subjects were enrolled into the study on a rolling basis, with the date of their first measurement saved as StartDate. Thus, W3 was collected on StartDate + 2 weeks. We have an interest in quickly seeing whether some odd blips seen in the repeated measures are related to outside factors that came and went over time. Is there an easy way to transform/manipulate this data with SPSS so we can graph the W1-W10 data for all of the subjects against the actual date corresponding to each measurement? I could do it manually in excel, etc, but it would take quite a while. I have coding experience, but I'm new to SPSS syntax. I'm also not certain there isn't a way to do it through the GUI that I just haven't figured out. Does anyone have any ideas about how to go about this?

Erin

[Bruce Weaver]

p.s. - You can use AGGREGATE with no break variable to get the minimum value of StartDate, and the DATEDIFF function to get number of days from that minimum StartDate.

Bruce Weaver wrote

Howdy.  :-|

If you restructure your data to a LONG format (with multiple rows per Subject_ID), you could estimate your model via MIXED, and include number of days from the minimum StartDate value (or some other convenient variable reflecting actual date) in your analysis.  Singer & Willett's book is a great resource for these types of models--and you can see worked examples from it here:

http://www.ats.ucla.edu/stat/examples/alda/

HTH.

ladyc wrote

Howdy,

I have a dataset with the following variables:

Subject_ID
StartDate
W1
W2
W3
...
W10

Each individual (Subject_ID) was measured weekly for 10 weeks (resulting in 10 repeated measurements, the W1-W10 variables). The subjects were enrolled into the study on a rolling basis, with the date of their first measurement saved as StartDate. Thus, W3 was collected on StartDate + 2 weeks. We have an interest in quickly seeing whether some odd blips seen in the repeated measures are related to outside factors that came and went over time. Is there an easy way to transform/manipulate this data with SPSS so we can graph the W1-W10 data for all of the subjects against the actual date corresponding to each measurement? I could do it manually in excel, etc, but it would take quite a while. I have coding experience, but I'm new to SPSS syntax. I'm also not certain there isn't a way to do it through the GUI that I just haven't figured out. Does anyone have any ideas about how to go about this?

Erin