Do If statement needed?
Locked
 
|
I am trying to figure out a way to include data from selected variables
from an initial clinical review form. I have a determination of need More info Need BH services No needed services I want to include information from the 'More info', in the initial clincial review, which requires another response when they get more info, to reflect either Need BH services, or no needed services from an updated form. But this information will be included in the initial clinical review from rather than the updated clinical review form - that information will be analyzed separately. Do I use some sort of DO IF statement to get all the 'more info' data and that following response? I hope someone can help me. Thanks. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
Judy,
It may be just me but I couldn't understand what you wanted. It almost sounds to me like you aren't talking about datasets at all but rather the (paper) forms on which data is being/has been collected. In fact, there seems to be two forms, an initial and an updated clinical review form, that have some sort of workflow relationship to one another. Perhaps you could rework your description and illustrate the relevent aspects of your data set with sample data that illustrates the issue you need help with. Gene Maguin >>I am trying to figure out a way to include data from selected variables from an initial clinical review form. I have a determination of need More info Need BH services No needed services I want to include information from the 'More info', in the initial clincial review, which requires another response when they get more info, to reflect either Need BH services, or no needed services from an updated form. But this information will be included in the initial clinical review from rather than the updated clinical review form - that information will be analyzed separately. Do I use some sort of DO IF statement to get all the 'more info' data and that following response? I hope someone can help me. Thanks. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
Judy,
Let's back way, way up and start with some very basic things. First, what is the structure of your data? Show us a some sample data, along with the variable names, that illustrates the problem. The crosstabs doesn't help very much. The only thing I've learned is the the data are all in one file. You also appear to have very long variable names which is going to make the example data a problem because of line wrapping. Perhaps, drastically shorten names. This is one specific thing to do to help us. Issue 2. It sounds like people fill out some kind of form at time 1. Then additional forms are filled out at later points in time. I guess the reason for the additional forms are that the initial (Time 1) form was either incompletely filled out, for whatever reason, or that the Time 1 data is no longer accurate and the new form has updated information. It furthermore sounds like every form, whether original or an update, has been entered in the same file. You know there has been an update because there must be a variable that has 'update' as one of its values. There also ought to be couple of other variables as well, like an id number and a form completion date. Is all this is true? If so, then you have an database update problem. I don't know very much about this kind of problem, others know a lot more, but I'd guess that normally the Time 1 data are put into one file, the master file, and the updates are put in another (the transaction file) and an update command is used to merge the two files, i.e., to update the master file. If all this is really true, then one solution is to take the file apart and separate it into Time 1 records that are the master file and all the rest into an update file. Look at the Update command. However, you've got to split the current file. The other alternative is to do all your work in the current file. Suppose the file is sorted by id number and form date. So, if there is two records with the same id number, the second one will be the more recent (the 'update') form and you want to transfer data from that back to the original form. There are two methods for this. One is to do a casestovars command to combine all the records on to a single record, then write do repeats to update the original values and finally delete the 'update' variables. Doable but tough. The other method is to use a lag function to look backwards from the original form to the first update and pull data from the first update forwards to the original form. Accomodations can be made for multiple updates per id. Either method is going to be hard to do. One way is to use a do repeat structure(s) with lag functions to >>Sorry, to be more explicit, here's a crosstab of the data I have. See under 'More Info', there are multiple reviews. I need to know the outcome of those multiple reviews. For example, there are 2 updates of "Reveiw #2 of 4". So what are the outcomes of those 2 updates? in other words, when 'More info' is filled out on a form, another form is filled out completing the information missing from the 'more info' form. I need to know how to extract that information, or if there was even any other forms filled out. So I thought a do if statement with the more info as the qualifier would give me results of other forms - if cbhc_determination eq 'more info', then, ???? Does that make more sense? dbo_clinical_review#caption cbhc_determination Initial Update Total review_number Review #1 of 1 1 1 Total 1 1 More info review_nu Review #1 of 1 62 0 62 Review #1 of 2 38 0 38 Review #1 of 3 2 0 2 Review #1 of 4 2 0 2 Review #2 of 2 0 6 6 Review #2 of 3 0 1 1 Review #2 of 4 0 2 2 Review #3 of 3 0 1 1 Review #3 of 4 0 1 1 Total 104 11 115 Needs BH Services review_number Review #1 of 1 922 0 922 Review #1 of 2 122 0 122 Review #1 of 3 7 0 7 Review #1 of 4 7 0 7 Review #2 of 2 0 156 156 Review #2 of 3 0 12 12 Review #2 of 4 0 5 5 Review #3 of 3 0 14 14 Review #3 of 4 0 9 9 Review #4 of 4 0 7 7 Total 1058 203 1261 No BH Services review_number Review #1 of 1 234 0 234 Review #1 of 2 9 0 9 Review #1 of 3 2 0 2 Review #2 of 2 0 14 14 Review #3 of 3 0 2 2 Review #3 of 4 0 1 1 Review #4 of 4 0 1 1 Total 245 18 263 Unable to Complete review_number Review #1 of 1 6 0 6 Review #1 of 2 1 0 1 Review #2 of 2 0 1 1 Total 7 1 8 -----Original Message----- From: Gene Maguin <[hidden email]> To: [hidden email] Sent: Wed, Feb 10, 2010 12:24 pm Subject: Re: Do If statement needed? Judy, It may be just me but I couldn't understand what you wanted. It almost sounds to me like you aren't talking about datasets at all but rather the (paper) forms on which data is being/has been collected. In fact, there seems to be two forms, an initial and an updated clinical review form, that have some sort of workflow relationship to one another. Perhaps you could rework your description and illustrate the relevent aspects of your data set with sample data that illustrates the issue you need help with. Gene Maguin >>I am trying to figure out a way to include data from selected variables from an initial clinical review form. I have a determination of need More info Need BH services No needed services I want to include information from the 'More info', in the initial clincial review, which requires another response when they get more info, to reflect either Need BH services, or no needed services from an updated form. But this information will be included in the initial clinical review from rather than the updated clinical review form - that information will be analyzed separately. Do I use some sort of DO IF statement to get all the 'more info' data and that following response? I hope someone can help me. Thanks. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Do If statement needed?
|
|
I work in drug and alcohol research. In my main data base I have a
variety of psychometric tests for clients. These are repeated up to 12 times. I have written syntax that selects the earliest test and the latest test in each series (usually for a 6-month period, but I can make it whatever I like), and compares them using paired T-tests. This shows average improvement over time for each test. This system works well. Dr Alex Davidson MRSNZ Research Analyst Odyssey House Trust Private Bag MBE M230 Auckland 1140 Tel: +64 9 623 1447 Fax: +64 9 623 9151 Web: www.odyssey.org.nz ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |