multilevel regression

[drfg2008]

SPSS 17 Win Vista / XP

For multilevel data (similar to that in my syntax example). Two questions:


1. Does anyone know if it is possible to draw more than one regression line (at the same time) in a scatter plot with SPSS?

2. Does anyone know an online resource where the SPSS solution to multilevel regression is explained for students not from strong mathematical background (if possible with screenshots –not with syntax- and relative simple explanations).

Thanks

Data example:

input program.
loop a =1 to 600 by 1.
end case.
end loop.
end file.
end input program.
exe.

COMPUTE group2=RV.UNIFORM(0,6).
RECODE group2 (Lowest thru 1=1) (Lowest thru 2=2) (Lowest thru 3=3) (Lowest thru 4=4) (Lowest thru 5=5) (Lowest thru 6=6) INTO group.
EXECUTE .
DELETE VARIABLES a group2.

IF  (group = 1) uv=RV.UNIFORM(0,20).
IF  (group = 2) uv=RV.UNIFORM(10,30).
IF  (group = 3) uv=RV.UNIFORM(20,40).
IF  (group = 4) uv=RV.UNIFORM(30,50).
IF  (group = 5) uv=RV.UNIFORM(40,60).
IF  (group = 6) uv=RV.UNIFORM(50,70).


IF  (group = 6) av=uv * 20 + RV.NORMAL(0,50).
IF  (group = 5) av=uv * 20 + RV.NORMAL(0,50) + 300.
IF  (group = 4) av=uv * 20 + RV.NORMAL(0,50) + 600.
IF  (group = 3) av=uv * 20 + RV.NORMAL(0,50) + 900.
IF  (group = 2) av=uv * 20 + RV.NORMAL(0,50) + 1200.
IF  (group = 1) av=uv * 20 + RV.NORMAL(0,50) + 1500.

EXECUTE.

IF  (group = 1) g1 = 1.
IF  (group = 2) g2 = 1.
IF  (group = 3) g3 = 1.
IF  (group = 4) g4 = 1.
IF  (group = 5) g5 = 1.
IF  (group = 6) g6 = 1.


EXECUTE .

RECODE g1 to g6   (MISSING=0).
EXECUTE.

var lab group 'Group'.
var lab uv 'independent variable: predictor'.
var lab av 'dependent variable: dv'.
var lab g1 'group membership: 1 '.
var lab g2 'group membership: 2'.
var lab g3 'group membership: 3'.
var lab g4 'group membership: 4'.
var lab g5 'group membership: 5'.
var lab g6 'group membership: 6'.


GRAPH
  /SCATTERPLOT(BIVAR)=uv WITH av
  /MISSING=LISTWISE.

[ViAnn Beadle]

Yes, it is possible to draw more than one regression line but how the process is determined by whether the multiple lines are subgroups or separate variables.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of drfg2008
Sent: Saturday, May 07, 2011 3:38 AM
To: [hidden email]
Subject: multilevel regression

SPSS 17 Win Vista / XP

For multilevel data (similar to that in my syntax example). Two questions:


1. Does anyone know if it is possible to draw more than one regression line (at the same time) in a scatter plot with SPSS?
<bigsnip/>

=====================
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

[drfg2008]

Thanks, now I've found it


IGRAPH
  /VIEWNAME='Streudiagramm'
  /X1=VAR(uv) TYPE=SCALE
  /Y=VAR(av) TYPE=SCALE
  /COLOR=VAR(group) TYPE=CATEGORICAL
  /COORDINATE=VERTICAL
  /FITLINE METHOD=REGRESSION LINEAR LINE=MEFFECT SPIKE=OFF
  /YLENGTH=5.2
  /X1LENGTH=6.5
  /CHARTLOOK='NONE'
  /CATORDER VAR(group) (ASCENDING VALUES OMITEMPTY)
  /CATORDER VAR(group) (ASCENDING VALUES OMITEMPTY)
  /SCATTER COINCIDENT=NONE.


However, what about the multilevel regression 'tutorial' for students without knowledge of syntax.

Has anyone an idea, maybe an URL?

[Christopher Stride]

Hi
Can anyone point me towards either some SPSS code (or some R code I can
embed in SPSS), and pref an example or two of using it, for calculating
the Variance Partition Coefficient (VPC) using the simulation method
outlined in Goldstein et al (2002)?
cheers
Chris

=====================
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