2-1-1 & error message PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Eunice Kim posted on Thursday, September 17, 2015 - 4:18 pm
Hello Linda, I have a 2-1-1 model with a level-2 moderator for the first-stage mediation. x (ZOSM) w (RZOSSD), and xw (RZOSMSD) were measured at level-2. m (TrustO) and y (OCB_O) were measured at level-1. I was able to run the analysis with warning message. The warning message says "*** WARNING in MODEL command
In the MODEL command, the following variable is a y-variable on the BETWEEN
level and an x-variable on the WITHIN level. This variable will be treated
as a y-variable on both levels: TRUSTO".

What does it mean? Do I need to do something to fix the issue?

USEVARIABLES ARE Team ZOSM RZOSSD RZOSMSD TrustO OCB_O ;
BETWEEN IS ZOSM RZOSSD RZOSMSD;
CLUSTER IS Team;
MISSING=ALL(999);

ANALYSIS: TYPE IS TWOLEVEL RANDOM;
MODEL:
%WITHIN%
TrustO OCB_O;
OCB_O ON TrustO;

%BETWEEN%
ZOSM RZOSSD RZOSMSD TrustO OCB_O ;
TrustO ON
ZOSM (a1)
RZOSSD
RZOSMSD (a2);
OCB_O ON
ZOSM
RZOSSD
RZOSMSD
TrustO(b);
MODEL CONSTRAINT:
NEW(ind1 wmodval);
wmodval=-1;
ind1=(a1+a2*wmodval)*b;
OUTPUT: TECH1 TECH8 CINTERVAL;
 Bengt O. Muthen posted on Thursday, September 17, 2015 - 6:07 pm
Nothing needs to be done. It's just internal Mplus housekeeping.
 Eunice Kim posted on Friday, September 18, 2015 - 3:37 pm
Hi Bengt,

Thank you so much for your prompt response! Very helpful!:-)

Thanks,
 Andrew Li posted on Wednesday, August 02, 2017 - 4:49 pm
Hello Dr. Muthen,

I have a 2-1-1 model (x-m-y) model with the first stage of the model being moderated by a level-2 moderator W. Is this the right code? Thank you.

missing = all(-999);
usevariables = a1 x m y w inter;
cluster = a1;
between = x w inter;
define:
standardize x w;
inter = x * w;
analysis: type = twolevel random;
model:
%within%
m y;
y on m;
%between%
x m y w inter;
m on
x (p1)
w
inter (p2);
y on
x
w
inter
m (b);
model constraint:
new (ind1 ind2 wmo1 wmo2);
wmo1 = 1; !1 SD of the moderator W
wmo2 = -1; !-1 SD of the moderator W
ind1 = (p1+p2*wmo1)*b;
ind2 = (p1+p2*wmo2)*b;
output: tech1 tech8 cinterval;
 Bengt O. Muthen posted on Thursday, August 03, 2017 - 5:23 pm
Looks correct.
 CHRIS OH posted on Friday, September 01, 2017 - 3:58 am
Hello Dr. Muthen,

I am trying to do 2-1-1 multilevel mediation and wonder which syntax is correct.

I used two individual level INTRI(m), COORD(y) variables.
The difference between the two is that the first used PAM(x), a mean level of individuals' positive affect in the same team, and The second used PA(x), an individual's response for positive affect. It means that individuals in the same team share the same PAM.

1:
USEVARIABLES = TEAMN PAM INTRI COORD;
BETWEEN = PAM;
CLUSTER = TEAMN;
MISSING = ALL (999);
ANALYSIS: TYPE IS TWOLEVEL RANDOM;
MODEL:
%WITHIN%
INTRI COORD;
COORD ON INTRI;
%BETWEEN%
PAM INTRI COORD;
INTRI ON PAM(A);
COORD ON INTRI(B);
COORD ON PAM;
MODEL CONSTRAINT:
NEW(INDB);
INDB=A*B;

2:
USEVARIABLES = TEAMN PA INTRI COORD;
CLUSTER = TEAMN;
MISSING = ALL (999);
ANALYSIS: TYPE IS TWOLEVEL RANDOM;
MODEL:
%WITHIN%
PA INTRI COORD;
COORD ON INTRI;
COORD ON PA;
INTRI ON PA;
%BETWEEN%
PA INTRI COORD;
INTRI ON PA(A);
COORD ON INTRI(B);
COORD ON PA;
MODEL CONSTRAINT:
NEW(INDB);
INDB=A*B;

Thank you in advance.
 Bengt O. Muthen posted on Friday, September 01, 2017 - 1:04 pm
Both are ok, but the second is better because it uses a latent variable decomposition of PA. See

Lüdtke, O., Marsh, H.W., Robitzsch, A., Trautwein, U., Asparouhov, T., & Muthén, B. (2008). The multilevel latent covariate model: A new, more reliable approach to group-level effects in contextual studies. Psychological Methods, 13, 203-229.
 Andrew Li posted on Tuesday, September 12, 2017 - 3:15 pm
Hello Dr. Muthen,

I have a 2-1-1 model (x-m-y) model with the first stage of the model being moderated by a level-2 moderator W. However, m has a random rather than a fixed slope. Also, I have a level-1 control variable (C1) and a level-2 control variable (C2). C1 has a fixed slope. Is this the right code? Thank you.

missing = all(-999);
usevariables = a1 x m y w c1 c2 inter;
cluster = a1;
between = x w c2 inter;
Within = C1;
define:
standardize x w;
inter = x * w;
analysis: type = twolevel random;
model:
%within%
m y C1;
y on C1;
sb | y on m;
%between%
x m y w C2 inter;
m on
x (a)
w
inter (b);
y on m (bb)
c2
x
w
inter;
sb WITH x m y w C2 inter;
[sb] (bw)
model constraint:
new (k ind1 ind2 wmo1 wmo2);
K = bb + bw;
wmo1 = 1; !1 SD of the moderator W
wmo2 = -1; !-1 SD of the moderator W
ind1 = (a+b*wmo1)*k;
ind2 = (a+b*wmo2)*k;
output: tech1 tech8 cinterval;
 Bengt O. Muthen posted on Tuesday, September 12, 2017 - 6:26 pm
This looks ok.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: