Spss vs Lisrel
Locked
 
|
Dear all
I would ask for a problem that I encountered in my analyses. I used SPSS to test the following longitudinal hypothesis: Organizational Identification T1 --> (increase) Social Support --> (decrease) Burnout T2 I controlled for Burnout T1, while Social Support is the standardized score residual obtained by regressing Social Support T2 on Social Support T1. I performed this analysis and I found a significant effect of Social Support on Burnout T2. I performed these analyses using both SPSS (hierarchical multiple regression analysis) and PROCESS macro by Hayes (2012). Hovewer, I tried to confirm my hypothesis also using SEM (Lisrel; full model), but when I use SEM the path from Social Support to Burnout T2 was strongly not significant. How is this possible? What could be the explanation? Thank you very much, Lorenzo ===================== 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 |
|
It sounds as though you contolled for T1 burnout twice. Is this correct or have I misread your problem?
Did you mean that you controolled for T1 burnout "by" regressing T2 on T1? Paul R. Swank, Ph.D., Professor Health Promotions and Behavioral Sciences School of Public Health University of Texas Health Science Center Houston ________________________________________ From: SPSSX(r) Discussion [[hidden email]] On Behalf Of [hidden email] [[hidden email]] Sent: Monday, April 08, 2013 10:14 AM To: [hidden email] Subject: Spss vs Lisrel Dear all I would ask for a problem that I encountered in my analyses. I used SPSS to test the following longitudinal hypothesis: Organizational Identification T1 --> (increase) Social Support --> (decrease) Burnout T2 I controlled for Burnout T1, while Social Support is the standardized score residual obtained by regressing Social Support T2 on Social Support T1. I performed this analysis and I found a significant effect of Social Support on Burnout T2. I performed these analyses using both SPSS (hierarchical multiple regression analysis) and PROCESS macro by Hayes (2012). Hovewer, I tried to confirm my hypothesis also using SEM (Lisrel; full model), but when I use SEM the path from Social Support to Burnout T2 was strongly not significant. How is this possible? What could be the explanation? Thank you very much, Lorenzo ===================== 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 |
|
|
In reply to this post by lorenzo.avanzi
Instead of manipulating the MACRO directly, why don't you install the custom dialog? This will ensure that you are not making any obvious errors. Moreover, he has made many advances over the years:
By the way, you can also assess for simple and multiple mediation using bootstrap methods offered in AMOS.
Best, Ryan On Mon, Apr 8, 2013 at 11:14 AM, [hidden email] <[hidden email]> wrote: Dear all |
Administrator
|
In reply to this post by lorenzo.avanzi
How do we know that you set up the LISREL code to be parallel to the Regressions and PROCESS analyses?
Verbal descriptions of what did and did not work are not typically useful in isolating discrepancies. Maybe you should attach your outputs? http://spssx-discussion.1045642.n5.nabble.com/template/NamlServlet.jtp?macro=reply&node=5719335 Nabble allows attachments (see More button). ---
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?" |
|
|
In reply to this post by lorenzo.avanzi
Dear All,
I attached the Lisrel Output Note: B_LAV2 = Burnout T2 RES_SUPP = standardized residual scores of Social Support (obtained by regressing T2 scores of the Social Support at T2 on the corresponding Social Support T1 scores) IDORG = Organizational Identification B_LAV1 = Burntout T1 TSE = Teacher Self-Efficacy as endogenous variable. SPSS_vs_Lisrel.txt |
|
|
In reply to this post by Swank, Paul R
No, I have not checked for Burnout T1 twice.
I used a procedure proposed by Smith and Beaton (2008) to test a mediation with only two times. Following this procedure T1-T2 changes in Social Support (which represent in my hypothesis the "mediator") are measured as the residual scores obtained by regressing T2 scores of the Social Support on the corresponding T1 scores. The mediation model that I postulated is: Organizaitonal Identification measured at Time 1 --> Social Suppor (residual scores T1-T2) --> Burnout measured at Time 2; and I controlled for Burnout measured at Time 1. |
|
Administrator
|
In reply to this post by lorenzo.avanzi
You are missing some vital info! The SPSS Runs and Process.
You really need to highlight the specific issue as well. I am NOT going to swim through all that LISREL output and try to correlate it to your issue! Assume that I have about 5 minutes to contribute to your plight. Streamline accordingly.
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?" |
|
|
Dear,
I attached a new file. At the beginning there is the output of Lisrel, while at the end there is the output of PROCESS The lines to check are indicated by asterisks: ******************************************************************************** In Lisrel Beta from RES_SUPP to B_LAV2 is not significant (-0.01), while in Process is significant: BETA B_LAV2 RES_SUPP IDORG B_LAV1 -------- -------- -------- -------- B_LAV2 - - -0.01 0.06 0.82 (0.04) (0.05) (0.11) -0.23 1.16 7.59 RES_SUPP - - - - 0.28 - - (0.12) 2.37 IDORG - - - - - - - - B_LAV1 - - - - - - - - SPSS_vs_Lisrel.txt |
|
Administrator
|
You need to indicate why you think the LISREL and the SPSS should have anything in common considering you are using 27 vars in LISREL and only 5 or 6 in Process.
140 cases is REALLY TOO SMALL FOR YOUR LISREL MODEL!!! No further comment! -
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?" |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |