FW: New to SPSS (Multiple responses)
Locked
FW: New to SPSS (Multiple responses)
 
|
For information to other listers for answers on research questions. There is now a new file with variables from both time 1 (t1.1 to t1.22) and time 2 (t2.1 to t2.22). I also generated an ID for each case (not really necessary, but just habit for me: compute id = $casenum, then dragged it to the top of the file) changed the original varnames (see table below for an indication of which variable is which: sorry but I don’t have time to generate shorter labels from the original enormously long ones. Bring back the days of max 20 characters!) merged the files for time 1 and 2 and created new variables diff1 to diff22 combining the responses. The 170 cases are the same at both times and are in the same order in both original files, therefore safe to merge. Roz now needs advice and assistance on appropriate statistical analyses and tests. To be honest, I’m not sure we can go much beyond descriptive analysis identifying which variables exhibit which patterns in new variables diff1 to diff22, particularly when they have changed between time points. diff1 to diff22 coded as: Value Time1 Time2 11 = Yes Yes 12 = Yes No 21 = No Yes 22 = No No John F Hall (Mr) Email: [hidden email] Website: www.surveyresearch.weebly.com From: Mohamed Danial [mailto:[hidden email]] Hi Mr Hall Thanks so much for helping me do this. I think it will take ages for me to do what you did to the file. Appreciate your help. I hope there is someone who could help me esp a statistician. I have posted my response to all the questions asked. Hopefully, the response will give a clear picture of what I am trying to do and what I am looking out. Thank you. Roz From: John F Hall <[hidden email]> Managed to add a serial number and create a combined data set roz1.sav (attached). 170 cases, 22 variables with Yes/No at time 1 (t1.1 to t1.22) repeated at time 2 (t2.1 to t2.22), then combined to yield 11 = y/y 12 = y/n 21 = n/y and 22 = n/n in var diff1 to diff22. [Quick check extract below] t1.1 t2.1 diff1 1 2 12 1 1 11 2 1 21 1 2 12 2 2 22 2 2 22 2 1 21 2 1 21 2 2 22 2 2 22
As David says, “What are you attempting to achieve with this data?“ Who are your respondents and why were they asked these questions? You’ll need a statistician to provide appropriate tests to see if there are any significant differences between answers at time 1 and at time 2. Over to you. John F Hall (Mr) Email: [hidden email] Website: http://surveyresearch.weebly.com/ From: John F Hall [[hidden email]] Roz I’ll have a look at these, but I can basically only do the mechanics of SPSS and some thinking about your research question(s). I’m not a statistician, but others on the list are (and still answering in the middle of August!) so I’ve copied my reply to the list (without the files, since it won’t accept attachments or embedded links). Hopefully one or two listers will address themselves to the inferential statistical questions. You have the same 170 cases and the same 22 variables in each of two files representing Time 1 and Time 2. I assume the cases are in the same order in each file. Your variable names are very long (typical in some beginners’ work) and can be replaced by something much shorter, since the same info is already included in the variable labels. The data needs tweaking to change the variable names for time 2 and then the files need to be merged so that all the data are in the same file. It’s also advisable to create a new variable, SERIAL, for each case. I suggest changing the variable names to something reflecting the sequence of items at each time (eg pf1.1 to pf1.22 and pf2.1 to pf2.22) and adding a serial number to each case. This can be done simply. I’ll get back to you after lunch with a new version of your file and the syntax which creates it. John Email: [hidden email] Website: http://surveyresearch.weebly.com/ John Hall From: Mohamed Danial [[hidden email]] Hi Mr Hall Thanks so much for helping. Really need help desperately. I have attached the two SPSS files. PF1 is the first reflection from students transfered into quantative numbers and PF2 is the second one. I guess i need to explain what I have done. What I did is that students wrote a reflection based on a peer feedback activity. I read these reflections and counted the number of times they said e.g. they have learnt, they received assistance etc. Then I have another feedback activity and the students wrote their reflections and I counted the number of times the students said they have learnt etc etc. My Sup believes that numbers from qualitative data adds value and I agree with him. We cant perform the Mcnemar Test because the number of Yes/Yes, No/No and Yes/No and No/Yes are way too small. So I came across multiple responses from the book Surveys in Social Research by David De Vaus which said multiple responses. So I tried, but I have no idea what the results show and my Sup has no idea either. YEs these figures are for peer feedback 1 and 2 and the same population at two different times. The first peer feedback was taken on Week 3 and the second on Week 5. My concern is that if I just substract the percentage, is it valid? For e.g if i use the phase 'significant increase', my Sup just shot me down because he said I cant say that because it is not tested. What test am i suppose to do? I am getting confused, very confused. Appreciate your help. I will look at the tutorials, I have been looking at a number of multiple reposnses youtube video, yeah understood the concept but how do i interpret them? Thanks thanks for the help Roz From: John F Hall <[hidden email]>
|
Re: New to SPSS (Multiple responses)
|
|
Matthew Thanks for this reply, but it’s Roz’s project. I just did a “rescue job” using some simple mechanics on the SPSS files in response to a plea for help via the list. The values in diff1 to diff22 are a combination of values (1 = Yes, 2 = No) on the original 22 variables at time1 (t1.1 to t1.22) and time2 (t2.1 to t22). I’m no statistician and so can’t offer further advice. There are no missing values, so a simple recode of 2 = 0 for t1.1 to t2.22 would work for your suggested tests: the combinations using: do repeat x = t1.1 to t1.22 /y = t2.1 to t2.22 /z = diff1 to diff22. compute z = x*10+y. end repeat. . . would then yield values 0,1,10 and 11. I can use mult resp to produce dichotomous tables, but I need to generate them without Roz’s original labels. Not sure if changing the settings works with mult resp. Would it help if I sent you the *.sav file I created (off-list)? John F Hall (Mr) Email: [hidden email] Website: www.surveyresearch.weebly.com John From: Poes, Matthew Joseph [mailto:[hidden email]] Hi John and Roz, I’d be willing to help if I could better understand what it is you are trying to achieve, and what the issues are. People responded yes or no to a set of questions, and as a result you have a set of yes/no combinations? You want to see if these combinations changed in some meaningful way from pre-test to post-test, correct? The multiple response piece of SPSS is for tables, so obviously that’s not an option. What needs to happen would be variables which reflect the categories of possible combinations. What is the diff score for? In general I don’t use or like diff scores in statistics. It’s an under-powered approach that makes improper assumptions for change, there are better approaches. In this case, I’m just not sure why you wouldn’t use a McNemar test for 2 related samples, and use an exact statistic to handle the power/assumption issue associated with the small sample within groups. Another possibility would be to convert the yes/no responses to 1’s and 0’s, and use them to create ratios. The ratio’s could reflect meaningful categories of yes responses (assuming that is the desired outcome of interest, more yes responses). Then it can be treated as a linear outcome, and you could test if the ratio changed significantly time 1 to time 2. In that case, what I would do is just treat them as a paired samples t-test. Compare the mean % yes responses in T1 to that of T2. If there were demographics and such, just switch to a 2 step regression, and use the T1 % yes responses as a covariate in the first step, T2 as the DV, and then the demographics or other predictors as additional covariates/predictors added in T2. By having the T1 percent yes responses at the first step, you essentially set the intercept to that of the time 1 score, and test if the t2 is significantly greater than t1. So by step 2 you are predicting the variation in T2 scores, accounting for where the people started at T1. Matthew J Poes Research Data Specialist Center for Prevention Research and Development University of Illinois 510 Devonshire Dr. Champaign, IL 61820 Phone: 217-265-4576 email: [hidden email] From: SPSSX(r) Discussion [[hidden email]] On Behalf Of John F Hall For information to other listers for answers on research questions. There is now a new file with variables from both time 1 (t1.1 to t1.22) and time 2 (t2.1 to t2.22). I also generated an ID for each case (not really necessary, but just habit for me: compute id = $casenum, then dragged it to the top of the file) changed the original varnames (see table below for an indication of which variable is which: sorry but I don’t have time to generate shorter labels from the original enormously long ones. Bring back the days of max 20 characters!) merged the files for time 1 and 2 and created new variables diff1 to diff22 combining the responses. The 170 cases are the same at both times and are in the same order in both original files, therefore safe to merge. Roz now needs advice and assistance on appropriate statistical analyses and tests. To be honest, I’m not sure we can go much beyond descriptive analysis identifying which variables exhibit which patterns in new variables diff1 to diff22, particularly when they have changed between time points. diff1 to diff22 coded as: Value Time1 Time2 11 = Yes Yes 12 = Yes No 21 = No Yes 22 = No No John F Hall (Mr) Email: [hidden email] Website: www.surveyresearch.weebly.com From: Mohamed Danial [hidden email] Hi Mr Hall Thanks so much for helping me do this. I think it will take ages for me to do what you did to the file. Appreciate your help. I hope there is someone who could help me esp a statistician. I have posted my response to all the questions asked. Hopefully, the response will give a clear picture of what I am trying to do and what I am looking out. Thank you. Roz From: John F Hall <[hidden email]> Managed to add a serial number and create a combined data set roz1.sav (attached). 170 cases, 22 variables with Yes/No at time 1 (t1.1 to t1.22) repeated at time 2 (t2.1 to t2.22), then combined to yield 11 = y/y 12 = y/n 21 = n/y and 22 = n/n in var diff1 to diff22. [Quick check extract below] t1.1 t2.1 diff1 1 2 12 1 1 11 2 1 21 1 2 12 2 2 22 2 2 22 2 1 21 2 1 21 2 2 22 2 2 22
As David says, “What are you attempting to achieve with this data?“ Who are your respondents and why were they asked these questions? You’ll need a statistician to provide appropriate tests to see if there are any significant differences between answers at time 1 and at time 2. Over to you. John F Hall (Mr) Email: [hidden email] Website: http://surveyresearch.weebly.com/ From: John F Hall [[hidden email]] Roz I’ll have a look at these, but I can basically only do the mechanics of SPSS and some thinking about your research question(s). I’m not a statistician, but others on the list are (and still answering in the middle of August!) so I’ve copied my reply to the list (without the files, since it won’t accept attachments or embedded links). Hopefully one or two listers will address themselves to the inferential statistical questions. You have the same 170 cases and the same 22 variables in each of two files representing Time 1 and Time 2. I assume the cases are in the same order in each file. Your variable names are very long (typical in some beginners’ work) and can be replaced by something much shorter, since the same info is already included in the variable labels. The data needs tweaking to change the variable names for time 2 and then the files need to be merged so that all the data are in the same file. It’s also advisable to create a new variable, SERIAL, for each case. I suggest changing the variable names to something reflecting the sequence of items at each time (eg pf1.1 to pf1.22 and pf2.1 to pf2.22) and adding a serial number to each case. This can be done simply. I’ll get back to you after lunch with a new version of your file and the syntax which creates it. John Email: [hidden email] Website: http://surveyresearch.weebly.com/ John Hall From: Mohamed Danial [[hidden email]] Hi Mr Hall Thanks so much for helping. Really need help desperately. I have attached the two SPSS files. PF1 is the first reflection from students transfered into quantative numbers and PF2 is the second one. I guess i need to explain what I have done. What I did is that students wrote a reflection based on a peer feedback activity. I read these reflections and counted the number of times they said e.g. they have learnt, they received assistance etc. Then I have another feedback activity and the students wrote their reflections and I counted the number of times the students said they have learnt etc etc. My Sup believes that numbers from qualitative data adds value and I agree with him. We cant perform the Mcnemar Test because the number of Yes/Yes, No/No and Yes/No and No/Yes are way too small. So I came across multiple responses from the book Surveys in Social Research by David De Vaus which said multiple responses. So I tried, but I have no idea what the results show and my Sup has no idea either. YEs these figures are for peer feedback 1 and 2 and the same population at two different times. The first peer feedback was taken on Week 3 and the second on Week 5. My concern is that if I just substract the percentage, is it valid? For e.g if i use the phase 'significant increase', my Sup just shot me down because he said I cant say that because it is not tested. What test am i suppose to do? I am getting confused, very confused. Appreciate your help. I will look at the tutorials, I have been looking at a number of multiple reposnses youtube video, yeah understood the concept but how do i interpret them? Thanks thanks for the help Roz From: John F Hall <[hidden email]>
|
Re: New to SPSS (Multiple responses)
|
|
In reply to this post by John F Hall
Just checked. Changing SPSS settings works for mult resp, so these may be helpful: mult resp groups t1 'Yes replies at time 1' (t1.1 to t1.22 (1)) t2 'Yes replies at time2' (t2.1 to t2.22 (1)) /freq t1 t2.
John F Hall (Mr) Email: [hidden email] Website: www.surveyresearch.weebly.com From: John F Hall [mailto:[hidden email]] Matthew Thanks for this reply, but it’s Roz’s project. I just did a “rescue job” using some simple mechanics on the SPSS files in response to a plea for help via the list. The values in diff1 to diff22 are a combination of values (1 = Yes, 2 = No) on the original 22 variables at time1 (t1.1 to t1.22) and time2 (t2.1 to t22). I’m no statistician and so can’t offer further advice. There are no missing values, so a simple recode of 2 = 0 for t1.1 to t2.22 would work for your suggested tests: the combinations using: do repeat x = t1.1 to t1.22 /y = t2.1 to t2.22 /z = diff1 to diff22. compute z = x*10+y. end repeat. . . would then yield values 0,1,10 and 11. I can use mult resp to produce dichotomous tables, but I need to generate them without Roz’s original labels. Not sure if changing the settings works with mult resp. Would it help if I sent you the *.sav file I created (off-list)? John F Hall (Mr) Email: [hidden email] Website: www.surveyresearch.weebly.com John From: Poes, Matthew Joseph [[hidden email]] Hi John and Roz, I’d be willing to help if I could better understand what it is you are trying to achieve, and what the issues are. People responded yes or no to a set of questions, and as a result you have a set of yes/no combinations? You want to see if these combinations changed in some meaningful way from pre-test to post-test, correct? The multiple response piece of SPSS is for tables, so obviously that’s not an option. What needs to happen would be variables which reflect the categories of possible combinations. What is the diff score for? In general I don’t use or like diff scores in statistics. It’s an under-powered approach that makes improper assumptions for change, there are better approaches. In this case, I’m just not sure why you wouldn’t use a McNemar test for 2 related samples, and use an exact statistic to handle the power/assumption issue associated with the small sample within groups. Another possibility would be to convert the yes/no responses to 1’s and 0’s, and use them to create ratios. The ratio’s could reflect meaningful categories of yes responses (assuming that is the desired outcome of interest, more yes responses). Then it can be treated as a linear outcome, and you could test if the ratio changed significantly time 1 to time 2. In that case, what I would do is just treat them as a paired samples t-test. Compare the mean % yes responses in T1 to that of T2. If there were demographics and such, just switch to a 2 step regression, and use the T1 % yes responses as a covariate in the first step, T2 as the DV, and then the demographics or other predictors as additional covariates/predictors added in T2. By having the T1 percent yes responses at the first step, you essentially set the intercept to that of the time 1 score, and test if the t2 is significantly greater than t1. So by step 2 you are predicting the variation in T2 scores, accounting for where the people started at T1. Matthew J Poes Research Data Specialist Center for Prevention Research and Development University of Illinois 510 Devonshire Dr. Champaign, IL 61820 Phone: 217-265-4576 email: [hidden email] From: SPSSX(r) Discussion [[hidden email]] On Behalf Of John F Hall For information to other listers for answers on research questions. There is now a new file with variables from both time 1 (t1.1 to t1.22) and time 2 (t2.1 to t2.22). I also generated an ID for each case (not really necessary, but just habit for me: compute id = $casenum, then dragged it to the top of the file) changed the original varnames (see table below for an indication of which variable is which: sorry but I don’t have time to generate shorter labels from the original enormously long ones. Bring back the days of max 20 characters!) merged the files for time 1 and 2 and created new variables diff1 to diff22 combining the responses. The 170 cases are the same at both times and are in the same order in both original files, therefore safe to merge. Roz now needs advice and assistance on appropriate statistical analyses and tests. To be honest, I’m not sure we can go much beyond descriptive analysis identifying which variables exhibit which patterns in new variables diff1 to diff22, particularly when they have changed between time points. diff1 to diff22 coded as: Value Time1 Time2 11 = Yes Yes 12 = Yes No 21 = No Yes 22 = No No John F Hall (Mr) Email: [hidden email] Website: www.surveyresearch.weebly.com From: Mohamed Danial [hidden email] Hi Mr Hall Thanks so much for helping me do this. I think it will take ages for me to do what you did to the file. Appreciate your help. I hope there is someone who could help me esp a statistician. I have posted my response to all the questions asked. Hopefully, the response will give a clear picture of what I am trying to do and what I am looking out. Thank you. Roz From: John F Hall <[hidden email]> Managed to add a serial number and create a combined data set roz1.sav (attached). 170 cases, 22 variables with Yes/No at time 1 (t1.1 to t1.22) repeated at time 2 (t2.1 to t2.22), then combined to yield 11 = y/y 12 = y/n 21 = n/y and 22 = n/n in var diff1 to diff22. [Quick check extract below] t1.1 t2.1 diff1 1 2 12 1 1 11 2 1 21 1 2 12 2 2 22 2 2 22 2 1 21 2 1 21 2 2 22 2 2 22
As David says, “What are you attempting to achieve with this data?“ Who are your respondents and why were they asked these questions? You’ll need a statistician to provide appropriate tests to see if there are any significant differences between answers at time 1 and at time 2. Over to you. John F Hall (Mr) Email: [hidden email] Website: http://surveyresearch.weebly.com/ From: John F Hall [[hidden email]] Roz I’ll have a look at these, but I can basically only do the mechanics of SPSS and some thinking about your research question(s). I’m not a statistician, but others on the list are (and still answering in the middle of August!) so I’ve copied my reply to the list (without the files, since it won’t accept attachments or embedded links). Hopefully one or two listers will address themselves to the inferential statistical questions. You have the same 170 cases and the same 22 variables in each of two files representing Time 1 and Time 2. I assume the cases are in the same order in each file. Your variable names are very long (typical in some beginners’ work) and can be replaced by something much shorter, since the same info is already included in the variable labels. The data needs tweaking to change the variable names for time 2 and then the files need to be merged so that all the data are in the same file. It’s also advisable to create a new variable, SERIAL, for each case. I suggest changing the variable names to something reflecting the sequence of items at each time (eg pf1.1 to pf1.22 and pf2.1 to pf2.22) and adding a serial number to each case. This can be done simply. I’ll get back to you after lunch with a new version of your file and the syntax which creates it. John Email: [hidden email] Website: http://surveyresearch.weebly.com/ John Hall From: Mohamed Danial [[hidden email]] Hi Mr Hall Thanks so much for helping. Really need help desperately. I have attached the two SPSS files. PF1 is the first reflection from students transfered into quantative numbers and PF2 is the second one. I guess i need to explain what I have done. What I did is that students wrote a reflection based on a peer feedback activity. I read these reflections and counted the number of times they said e.g. they have learnt, they received assistance etc. Then I have another feedback activity and the students wrote their reflections and I counted the number of times the students said they have learnt etc etc. My Sup believes that numbers from qualitative data adds value and I agree with him. We cant perform the Mcnemar Test because the number of Yes/Yes, No/No and Yes/No and No/Yes are way too small. So I came across multiple responses from the book Surveys in Social Research by David De Vaus which said multiple responses. So I tried, but I have no idea what the results show and my Sup has no idea either. YEs these figures are for peer feedback 1 and 2 and the same population at two different times. The first peer feedback was taken on Week 3 and the second on Week 5. My concern is that if I just substract the percentage, is it valid? For e.g if i use the phase 'significant increase', my Sup just shot me down because he said I cant say that because it is not tested. What test am i suppose to do? I am getting confused, very confused. Appreciate your help. I will look at the tutorials, I have been looking at a number of multiple reposnses youtube video, yeah understood the concept but how do i interpret them? Thanks thanks for the help Roz From: John F Hall <[hidden email]>
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Re: FW: New to SPSS (Multiple responses)
Administrator
|
In reply to this post by John F Hall
I understand the concept of providing help, but this seems to be verging on doing somebodies job for them.
---
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?" |
Re: FW: New to SPSS (Multiple responses)
|
|
David
I don't know if you've ever been a university teacher, but I have (for 40 years or so) and I've never done students' work for them. I've shown them what to do if they were capable of following the logic, but there are dozens of students who are led (or forced) by incompetent supervisors to believe that quantitative analysis is essential to their research, even when it's completely irrelevant. In this case an interesting piece of research was at a standstill ( with a deadline to meet) and would have been stillborn without some intervention (all too frequently the case these days and even 40 years ago). It's up to Roz to make sense of the data, but as a complete beginner in SPSS I feel he warranted a bit of encouragement and help, short of doing the analysis and reporting for him. No offence, but get off your high horse, mate. I still rate you highly for all your technical wizardry, but you're a long way short on social theory and policy input. John Email: [hidden email] Website: www.surveyresearch.weebly.com -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of David Marso Sent: 15 August 2012 19:54 To: [hidden email] Subject: Re: FW: New to SPSS (Multiple responses) I understand the concept of providing help, but this seems to be verging on doing somebodies job for them. --- John F Hall wrote > > For information to other listers for answers on research questions. > There is now a new file with variables from both time 1 (t1.1 to > t1.22) and time 2 (t2.1 to t2.22). I also generated an ID for each > case (not really necessary, but just habit for me: compute id = > $casenum, then dragged it to the top of the file) changed the original > varnames (see table below for an indication of which variable is > which: sorry but I don’t have time to generate shorter labels from the > original enormously long ones. Bring back the days of max 20 > characters!) merged the files for time 1 and 2 and created new > variables diff1 to diff22 combining the responses. The 170 cases are > the same at both times and are in the same order in both original files, therefore safe to merge. > > > > Roz now needs advice and assistance on appropriate statistical > analyses and tests. To be honest, I’m not sure we can go much beyond > descriptive analysis identifying which variables exhibit which > patterns in new variables diff1 to diff22, particularly when they have > changed between time points. > > > > diff1 to diff22 coded as: > > > > Value Time1 Time2 > > > > 11 = Yes Yes > > 12 = Yes No > > 21 = No Yes > > 22 = No No > > > > > > John F Hall (Mr) > > > > Email: johnfhall@ > > Website: www.surveyresearch.weebly.com > <http://surveyresearch.weebly.com/> > > > > > > > > > > From: Mohamed Danial [mailto:danial_rozi@] > Sent: 15 August 2012 02:36 > To: John F Hall > Cc: 'Bruce Weaver'; 'David Marso' > Subject: Re: New to SPSS (Multiple responses) > > > > Hi Mr Hall > > Thanks so much for helping me do this. I think it will take ages for > me to do what you did to the file. Appreciate your help. I hope there > is someone who could help me esp a statistician. I have posted my > response to all the questions asked. Hopefully, the response will give > a clear picture of what I am trying to do and what I am looking out. Thank you. > > > > Roz > > > > From: John F Hall <johnfhall@> > To: 'Mohamed Danial' <danial_rozi@> > Cc: 'Bruce Weaver' <bruce.weaver@>; 'John F Hall' > <johnfhall@>; 'David Marso' <david.marso@> > Sent: Wednesday, 15 August 2012 1:35 AM > Subject: RE: New to SPSS (Multiple responses) > > Managed to add a serial number and create a combined data set roz1.sav > (attached). 170 cases, 22 variables with Yes/No at time 1 (t1.1 to > t1.22) repeated at time 2 (t2.1 to t2.22), then combined to yield 11 = > y/y 12 = y/n 21 = n/y and 22 = n/n in var diff1 to diff22. > > > > [Quick check extract below] > > > > t1.1 t2.1 diff1 > > > > 1 2 12 > > 1 1 11 > > 2 1 21 > > 1 2 12 > > 2 2 22 > > 2 2 22 > > 2 1 21 > > 2 1 21 > > 2 2 22 > > 2 2 22 > > > > > > > Variable Names > > > ActiontakenafterPFactivityevidenceoffeedforwarde.g.studentswilll > > > Learntawarewhattoavoidwhenwritingessaysocouldimprovee.g.dumpingc > > > RubricsSWAguidestudentshowtomarkmadeclearthecriteriaforsuccessra > > > Identifypeersmistakesgoodpointse.g.didnotexplainelaboratelackofs > > > AssistancebyteacherduringPeerFeedback > > > Studentknowshowtogivefeedbackmarkclearlygradecorrectlyconfused > > > OpportunitytoevaluateownMYlearningstrengthsandweaknesseslearning > > > PFhasbenefittedstudentsinlearninge.g.productivesharpenskillmoref > > > Groupmembersdidwellsatisfiedwithgroupmemberseasyenjoytoworkwithg > > > Grouphaseffectivediscussionsenthusiasticpeershelpfulnotparticipa > > > Peerscouldimprovefromfeedbackgiven > > > Studentsbelievetheygavereceivedconstructivepeerfeedback > > > Studenthastakensometimetogivefeedbackandcompletemarkinglessontoo > > > PFgoodmethodtolearnascouldseevarietyofessaysandlearntviewsthestr > > > Learnthowtobeamarkerandhowtoawardmarkscleareronmarkingwhatteache > > > PFoutcomesofPFdiscussionwillhelpstudentstoscorewellgetbettergrad > > > Peershelpedtoclarifydoubtsconfusionandexplainedwhatwaswrittenmea > > > PFwasenjoyableinterestingfunengaginginformativeremarkablenotbori > > > Markedbetterintermsofaccuracyinmarkingfeedbackgivenandreceived > > > Markwithhonestynotbias > > > WanttodoPFagainmorePFbecausehavelearntSWA > > > Teachertomarkbecausetheygavemoreaccuratefeedbackandmarks > > > Currently defined variables > > > > > > As David says, “What are you attempting to achieve with this data?“ > > > > Who are your respondents and why were they asked these questions? > You’ll need a statistician to provide appropriate tests to see if > there are any significant differences between answers at time 1 and at time 2. > > > > Over to you. > > > > > > John F Hall (Mr) > > > > Email: johnfhall@ > > Website: http://surveyresearch.weebly.com/ > > > > > > > > > > > > From: John F Hall [mailto:johnfhall@] > Sent: 14 August 2012 13:32 > To: 'Mohamed Danial' > Cc: 'SPSSX-L@.UGA'; 'Bruce Weaver' > Subject: RE: New to SPSS (Multiple responses) > > > > Roz > > > > I’ll have a look at these, but I can basically only do the mechanics > of SPSS and some thinking about your research question(s). I’m not a > statistician, but others on the list are (and still answering in the > middle of August!) so I’ve copied my reply to the list (without the > files, since it won’t accept attachments or embedded links). > Hopefully one or two listers will address themselves to the > inferential statistical questions. > > > > You have the same 170 cases and the same 22 variables in each of two > files representing Time 1 and Time 2. I assume the cases are in the > same order in each file. Your variable names are very long (typical > in some beginners’ work) and can be replaced by something much > shorter, since the same info is already included in the variable > labels. The data needs tweaking to change the variable names for time > 2 and then the files need to be merged so that all the data are in the > same file. It’s also advisable to create a new variable, SERIAL, for > each case. I suggest changing the variable names to something > reflecting the sequence of items at each time (eg pf1.1 to pf1.22 and > pf2.1 to pf2.22) and adding a serial number to each case. This can be > done simply. I’ll get back to you after lunch with a new version of your file and the syntax which creates it. > > > > John > > > > Email: johnfhall@ > > Website: http://surveyresearch.weebly.com/ > > > > > > > > John Hall > > > > From: Mohamed Danial [mailto:danial_rozi@] > Sent: 14 August 2012 10:14 > To: John F Hall > Subject: Re: New to SPSS (Multiple responses) > > > > Hi Mr Hall > > Thanks so much for helping. Really need help desperately. I have > attached the two SPSS files. PF1 is the first reflection from students > transfered into quantative numbers and PF2 is the second one. > > > > I guess i need to explain what I have done. What I did is that > students wrote a reflection based on a peer feedback activity. I read > these reflections and counted the number of times they said e.g. they > have learnt, they received assistance etc. Then I have another > feedback activity and the students wrote their reflections and I > counted the number of times the students said they have learnt etc etc. > > > > My Sup believes that numbers from qualitative data adds value and I > agree with him. We cant perform the Mcnemar Test because the number of > Yes/Yes, No/No and Yes/No and No/Yes are way too small. So I came > across multiple responses from the book Surveys in Social Research by > David De Vaus which said multiple responses. So I tried, but I have no > idea what the results show and my Sup has no idea either. > > > > YEs these figures are for peer feedback 1 and 2 and the same > population at two different times. The first peer feedback was taken > on Week 3 and the second on Week 5. My concern is that if I just > substract the percentage, is it valid? For e.g if i use the phase > 'significant increase', my Sup just shot me down because he said I > cant say that because it is not tested. What test am i suppose to do? > I am getting confused, very confused. > > > > Appreciate your help. I will look at the tutorials, I have been > looking at a number of multiple reposnses youtube video, yeah > understood the concept but how do i interpret them? > > > > Thanks thanks for the help > > Roz > > > > From: John F Hall <johnfhall@> > To: 'rozdan' <danial_rozi@>; SPSSX-L@.UGA > Sent: Tuesday, 14 August 2012 5:21 PM > Subject: RE: New to SPSS (Multiple responses) > > > Are these before and after figures on the same variables from the same > population at two different times? How many actual cases do you have? > It may be that all you need is some sort of index for each row > subtracting % for 2nd PF from 1st PF and then trying to make some sort > of theoretical or policy sense from the results. > > If you send me [off-list] the instrument/questionnaire used and your > SPSS *.sav file, I'll have a look and see if I can help. Meanwhile > check out the SPSS tutorials on my website > (http://surveyresearch.weebly.com/summary-guide-to-spss-tutorials.html ). > I'm not convinced that this is a case for multiple response as such, > but check out > http://surveyresearch.weebly.com/33-multiple-response-mult-response.html . > > > John F Hall (Mr) > > Email: johnfhall@ > Website: http://www.surveyresearch.weebly.com/ > > > > > > > -----Original Message----- > From: SPSSX(r) Discussion [mailto:SPSSX-L@.UGA] On Behalf Of rozdan > Sent: 14 August 2012 08:35 > To: SPSSX-L@.UGA > Subject: New to SPSS (Multiple responses) > > Hi > Desperately need help to interpret this data. Very new to SPSS and > needed to do a multiple response from a qualitative data to > quantitative. Fed the numbers in SPSS and had this. The table shows > results from 2 peer feedback activities. The columns (not shaded) is > the first responses during peer feedback and the shaded columns are > the second responses during the second peer feedback activity. > > My Sup said that the results cannot be used because it didnt show > much. I am sure it can and I feel it is valuable because I am sure it > shows something. > Can anyone help how I should interpret this? > > > http://spssx-discussion.1045642.n5.nabble.com/file/n5714676/multiple_r > espons > es_spss.jpg > > > > > -- > View this message in context: > http://spssx-discussion.1045642.n5.nabble.com/New-to-SPSS-Multiple-res > ponses > -tp5714676.html > Sent from the SPSSX Discussion mailing list archive at Nabble.com. > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA (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 > -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/FW-New-to-SPSS-Multiple-responses-tp5714715p5714724.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 Mr Hall,
I appreciate all the help. This means a lot to me because reading the advises have helped me understand better. These valuable advises are something that I didnt get from my Supervisor. I have been struggling to figure all these for 2 months and without any direction, I think I will need to figure out in the dark for another 2 months or more. I never see this and will never see this as getting someone to do my work. What you did and a few others to help me, is indeed an encouragement for me to learn more and giving me good direction as to where I need to go. I do feel encouraged that I am not alone in my quest to gain knowledge. I am learning SPSS and I am happy to tell you that I know how to combine the two data as what you did for me. Appreciate if others do not see Mr Hall (and a few others who have been relentlessly helping me) as doing my work. I do feel discouraged for 2 months figuring this out and this forum was the light at the end of the tunnel for me because finally I can see something. Sorry. |
| Free forum by Nabble | Edit this page |