Find the right analysis
Locked
 
|
Hi everyone,
Based on the below variables I would like to understand which analysis could I use to find some kind of relationship. I thought about Canonical correlation analysis and used Discriminant analysis (since I don't have Manova), but I couldn't find any significant relationship. The first question (4 drugs) is how many patients are receiving each drug: 30 patients received drug 1 30 patients received drug 2 10 patients received drug 3 30 patients received drug 4 The second question (13 attributes) rate the importance when choosing a drug (any) (1 not important - 7 extremely important). The Third question (13 attributes) rate the importance when choosing drug 1 (1 not important - 7 extremely important). This question was asked randomly using a rather complicated routine, since some doctors have the same number of patients within different drugs (that's why there are blanks). I've attached a sample: sample.txt Thanks in advance, Mils doctor q11_1 q11_2 q11_3 q11_4 q14a_1 q14a_2 q14a_3 q14a_4 q14a_5 q14a_6 q14a_7 q14a_8 q14a_9 q14a_10 q14a_11 q14a_12 q14a_13 q15a_1 q15a_2 q15a_3 q15a_4 q15a_5 q15a_6 q15a_7 q15a_8 q15a_9 q15a_10 q15a_11 q15a_12 q15a_13 asd 30 30 10 30 4 4 4 4 4 4 5 6 5 6 6 5 6 qwe 8 10 2 6 6 7 5 4 6 4 7 6 6 6 6 6 7 zxc 8 8 3 1 7 7 6 6 5 1 6 5 5 7 7 6 7 6 6 5 5 5 1 6 4 7 7 7 6 5 wer 20 25 10 15 6 6 6 6 6 4 7 5 6 6 7 6 5 sdf 15 15 10 20 5 5 5 5 5 5 5 4 5 5 4 4 5 5 4 4 4 4 2 5 5 4 7 5 5 5 xcv 5 5 20 10 7 7 6 6 7 7 6 7 7 7 7 7 7 ert 40 30 10 20 6 6 5 4 6 3 4 3 6 2 6 5 4 6 6 5 4 6 2 6 5 4 6 6 4 4 dfg 7 5 2 4 6 6 5 4 6 5 5 4 5 4 7 6 5 5 5 5 4 5 3 6 5 5 6 5 6 5 cvb 15 16 0 10 5 6 5 5 6 6 6 5 6 6 6 5 7 tyu 5 45 40 40 6 4 7 6 4 5 6 4 3 4 7 7 6
mils
|
Administrator
|
This is terribly unclear as stated. Please clarify?
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
Ok, lets see if I can explain myself better:
I would like first to understand if there is any relationship between the number of patients that are currently receiving drugs (q11_1, q11_2, q11_3, q11_4) and the reasons when choosing drugs (q14a_1, q14a_2, q14a_3, q14a_4, q14a_5, q14a_6, q14a_7, q14a_8, q14a_9, q14a_10, q14a_11, q14a_12, q14a_13). Second I would like to understand if there is any relationship between the number of patients that are currently receiving drug 1 (q11_1) and the reasons when choosing drug 1 (q15a_1, q15a_2, q15a_3, q15a_4, q15a_5, q15a_6, q15a_7, q15a_8, q15a_9, q15a_10, q15a_11, q15a_12, q15a_13). And also for the rest (drug 2, 3 and 4) which I haven't included any data. Does that make any sense now? Thanks in advance! Mils
mils
|
|
|
There seems to be something quite odd about this--or it's just that I don't understand. So let's back up a bit to understand the story. You have 100 people, all of whom are receiving one of four drugs and you know which of the four drugs each person is receiving.
Did each person have a choice receiving each of the four drugs and they chose the one they are taking or did a unique somebody for each person decide which of the four drugs the person would receive or did a (as in one) somebody decide which drug each person would receive? Next, who did the ratings? The drug receivers or the drug chooser/choosers? The question 2 ratings. Did the raters a) rate their drug choice relative to the other three possible drug choices, b) did they rate each eligible drug choice or c) something else. The wording of the rating task is critical to understand. The question 3 ratings is not clear. Did they a) rate the drug that the person was receiving and not any of the other alternatives or b) something else? Let's say is was a). For the people receiving drug 1 what was drug 1 rated relative to? Gene Maguin -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of mils Sent: Friday, August 16, 2013 5:40 AM To: [hidden email] Subject: Re: Find the right analysis Ok, lets see if I can explain myself better: I would like first to understand if there is any relationship between the number of patients that are currently receiving drugs (q11_1, q11_2, q11_3, q11_4) and the reasons when choosing drugs (q14a_1, q14a_2, q14a_3, q14a_4, q14a_5, q14a_6, q14a_7, q14a_8, q14a_9, q14a_10, q14a_11, q14a_12, q14a_13). Second I would like to understand if there is any relationship between the number of patients that are currently receiving drug 1 (q11_1) and the reasons when choosing drug 1 (q15a_1, q15a_2, q15a_3, q15a_4, q15a_5, q15a_6, q15a_7, q15a_8, q15a_9, q15a_10, q15a_11, q15a_12, q15a_13). And also for the rest (drug 2, 3 and 4) which I haven't included any data. Does that make any sense now? Thanks in advance! Mils ----- mils -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Find-the-right-analysis-tp5721584p5721600.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== 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 |
|
|
Hi Eugene,
I have 100 doctors who completed the survey. Each doctor is asked: Q11.Thinking of all your patients with Psoriasis in your care who is currently receiving a biologic, how many are receiving each of the following drugs? Q14. Below are several criteria that may or may not be important when choosing a biologic for Psoriasis. For each item (attribute) please rate its importance by using a scale of 1-7 where 1=not at all important and 7=extremely important. Q15. Based on your experience and what you have read or been told about each biologic, please use a scale of 1-7, to indicate what you think of the performance of each of the following drugs on each attribute for the treatment of Psoriasis. (This question is using a rather complicate filter and it will show those drugs with a high number of patients (q11). This filter will take into account those drugs with the same number of patients at q11 and show just one. Please, let me know if this is making sense now... Joan
mils
|
|
|
I understand better. Thank you. Ok, here's now how I envision the survey working. Q11_x to Q11_y is the number of persons being treated with each drug.
Q14_a to Q14_b is selection criteria list that is presented once for each respondent and questions are not keyed to whatever was said for Q11. Q15_c to Q15_d is presented once for each drug whose Q11 number of patients is greater than the threshold. Some docs might see the Q15 set once while others see it 3 or 4 times depending on their prescribing patterns and number of patients seen. >> I would like first to understand if there is any relationship between the number of patients that are currently receiving drugs (q11_1, q11_2, q11_3, q11_4) and the reasons when choosing drugs (q14a_1, q14a_2, q14a_3, q14a_4, q14a_5, q14a_6, q14a_7, q14a_8, q14a_9, q14a_10, q14a_11, q14a_12, q14a_13). Keep in mind that others, I'm sure, have much more experience with analyzing this sort of data, and I hope you from them. I think I might begin with the Q11 data and convert the number of patients to percents and then look at the correlations between the Q14 item set and prescribed percentage share on a drug by drug basis. You could also convert the within doc percents to within doc ranks and use the ranks. There are some issues with using percents because they are ratios. I don't work with them so others know more than me. This could wind up with a multiple regression. >> Second I would like to understand if there is any relationship between the number of patients that are currently receiving drug 1 (q11_1) and the reasons when choosing drug 1. From what I remember of your original message, I don't think your data are organized correctly because I think the question here is whether the reason x rating differs across the four drug groups that you illustrated in your original message. This could be a series of univariate anova or a multivariate anova. I'm assuming here that the four groups you illustrated are defined by the highest percentage prescribed drug. I hope all this is helpful but I don't know that it is because I may not understand the data well enough or overall goals. Gene Maguin -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of mils Sent: Friday, August 16, 2013 11:50 AM To: [hidden email] Subject: Re: Find the right analysis Hi Eugene, I have 100 doctors who completed the survey. Each doctor is asked: Q11.Thinking of all your patients with Psoriasis in your care who is currently receiving a biologic, how many are receiving each of the following drugs? Q14. Below are several criteria that may or may not be important when choosing a biologic for Psoriasis. For each item (attribute) please rate its importance by using a scale of 1-7 where 1=not at all important and 7=extremely important. Q15. Based on your experience and what you have read or been told about each biologic, please use a scale of 1-7, to indicate what you think of the performance of each of the following drugs on each attribute for the treatment of Psoriasis. (This question is using a rather complicate filter and it will show those drugs with a high number of patients (q11). This filter will take into account those drugs with the same number of patients at q11 and show just one. Please, let me know if this is making sense now... Joan ----- mils -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Find-the-right-analysis-tp5721584p5721617.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== 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 |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |