Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 33 views

    Defining Model 1 (Null Model) With PASW Menu Commands: Models: Specified Subjects and Repeated

    Uploaded by

    fordlee
    The document describes how to define 5 multilevel models using PASW/SPSS commands. It involves specifying variables as fixed effects, random effects, and interactions between levels. Model 1 is a null model with the intercept varying randomly. Model 2 adds a level-1 predictor. Model 3 adds cross-level predictors. Model 4 specifies a random slope. Model 5 adds cross-level interactions. The steps involve selecting variables and transferring them between dialogue boxes to define the model structure.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Defining Model 1 (Null Model) With PASW Menu Commands: Models: Specified Subjects and Repeated

    Uploaded by

    fordlee
    33 views18 pages
    The document describes how to define 5 multilevel models using PASW/SPSS commands. It involves specifying variables as fixed effects, random effects, and interactions between levels. Model 1 is a null model with the intercept varying randomly. Model 2 adds a level-1 predictor. Model 3 adds cross-level predictors. Model 4 specifies a random slope. Model 5 adds cross-level interactions. The steps involve selecting variables and transferring them between dialogue boxes to define the model structure.

    Original Description:

    OK

    Original Title

    models1-5instructions

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document describes how to define 5 multilevel models using PASW/SPSS commands. It involves specifying variables as fixed effects, random effects, and interactions between levels. Model 1 is a null model with the intercept varying randomly. Model 2 adds a level-1 predictor. Model 3 adds cross-level predictors. Model 4 specifies a random slope. Model 5 adds cross-level interactions. The steps involve selecting variables and transferring them between dialogue boxes to define the model structure.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    33 views18 pages

    Defining Model 1 (Null Model) With PASW Menu Commands: Models: Specified Subjects and Repeated

    Uploaded by

    fordlee
    The document describes how to define 5 multilevel models using PASW/SPSS commands. It involves specifying variables as fixed effects, random effects, and interactions between levels. Model 1 is a null model with the intercept varying randomly. Model 2 adds a level-1 predictor. Model 3 adds cross-level predictors. Model 4 specifies a random slope. Model 5 adds cross-level interactions. The steps involve selecting variables and transferring them between dialogue boxes to define the model structure.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 18

    Ronald H. Heck, Scott L.

    Thomas, Lynn Tabata Fall 2009


    Defining Model 1 (Null Model) with PASW Menu Commands

    Note: The following


    models are constructed
    using the mathdata.sav
    file.

    1 .Go to the PASW


    toolbar and select
    ANALYZE, MIXED
    MODELS, LINEAR.
    This command opens
    the Linear Mixed
    Models: Specified
    Subjects and Repeated
    dialogue box.

    2. From the Linear Mixed Models:


    Specify the Subjects and Repeated
    dialogue box, click to select the
    schcode variable from the left
    column. Then click the right arrow
    button to transfer schcode into the
    Subjects box. This is the grouping
    variable. We could add a second
    grouping variable if we were
    defining a three-level model.
    Click the CONTINUE button to
    display the Linear Mixed Models
    dialogue box.

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    3. The Linear Mixed Models dialogue box is used to define the dependent variable as well as any
    factors (categorical variables) or covariates (interval variables) in the analysis.
    a. Click to select the math variable from the left column. Then click the right arrow button to
    transfer math into the Dependent Variable box.
    b. Since there are no
    fixed effects in the null
    model, click the
    RANDOM button to
    access the Linear
    Mixed Models:
    Random Effects
    dialogue box.

    4. The RANDOM command is used to specify the random effects; that is, which variables are to
    be treated as randomly varying across groups. In this case, only the intercept varies randomly.
    a. Change the default
    covariance type to
    SCALED IDENTITY
    by clicking on the
    pull-down menu to
    make the selection.
    b. Click the Include
    intercept option. This
    identifies it as a
    random effect.
    c. Click to select the
    Subject Groupings
    variable schcode. Then
    click the right arrow
    button to transfer the
    variable into the
    Combinations box.
    This identifies the

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009


    intercept as varying
    across schools.
    Click the CONTINUE
    button to return
    5. In the Linear Mixed
    Models dialogue box, click
    the ESTIMATION button
    to access the Linear Mixed
    Models: Estimation
    dialogue box.
    We will use the default
    restricted maximum
    likelihood (REML) setting.
    Click the CONTINUE
    button to return to the
    Linear Mixed Models
    dialogue box.

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009


    6. In the Linear Mixed
    Models dialogue box,
    click the STATISTICS
    button to access the
    Linear Mixed Models:
    Statistics dialogue box.
    Click and select the
    following three statistics:
    Parameter estimates, tests
    for covariance
    parameters, and
    covariances of random
    effects.
    Then click the
    CONTINUE button to
    return to the Linear
    Mixed Models dialogue
    box.
    7. Finally, in the Linear
    Mixed Models dialogue
    box, click the OK button
    to run the model.

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    Defining Model 2 with PASW Menu Commands


    (PASW settings will
    default to those used for
    the preceding Null
    Model.)
    1. Go to the PASW toolbar
    and select ANALYZE,
    MIXED MODELS,
    LINEAR.
    This command opens the
    Linear Mixed Models:
    Specified Subjects and
    Repeated dialogue box.

    2. The Linear Mixed Models: Specify


    Subjects and Repeated dialogue box,
    displays schcode in the Subjects box.
    Click the CONTINUE button to
    display the Linear Mixed Models
    dialogue box.

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    3. The Linear Mixed Models


    dialogue box displays math
    in the Dependent Variable
    box.
    a. Locate and click the
    variables lowses and
    lowses_mean variables from
    the left column. Then click
    the right arrow button to
    transfer the variables into
    the Covariate(s) box.
    b. Click the FIXED button
    to access the Linear Mixed
    Models: Fixed Effects
    dialogue box.
    4a. Within the Linear Mixed Models: Fixed Effects dialogue box, click the pull-down menu to
    change the factorial setting to Main Effects.
    b. Click to select
    lowses from the
    Factors and
    Covariates box
    then click the
    ADD button to
    move the variable
    into the Model
    box. Note that we
    will only place
    this variable in the
    model at this time
    since it is a level
    one predictor in
    Model 2.
    Click the
    CONTINUE
    button to return to
    the Linear Mixed
    Models dialogue
    box. In the Linear
    Mixed Models
    dialogue box,
    click the RANDOM button to access the Linear Mixed Models: Random Effects dialogue box

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    5. In the Linear Mixed


    Models dialogue box,
    the estimation and
    statistics settings
    remain the same as
    Model 1 so skip over
    these buttons and click
    OK to run the model.

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    Defining Model 3 with PASW Menu Commands

    (PASW settings will default


    to those used for Model 2.)
    1. Go to the PASW toolbar
    and select ANALYZE,
    MIXED MODELS,
    LINEAR.
    This command opens the
    Linear Mixed Models:
    Specified Subjects and
    Repeated dialogue box.

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009


    2. The Linear Mixed Models:
    Specify Subjects and Repeated
    dialogue box, displays
    schcode in the Subjects box.
    Click the CONTINUE button
    to display the Linear Mixed
    Models dialogue box.

    3. The Linear Mixed Models


    dialogue box displays math
    in the Dependent Variable
    box.
    a. Click the
    teacheffect_mean variable
    from the left column. Then
    click the right arrow button
    to transfer the variable into
    the Covariate(s) box.
    b. Click the FIXED button
    to access the Linear Mixed
    Models: Fixed Effects
    dialogue box.

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009


    4. Clear the
    lowses
    variable from
    the Model by
    clicking the
    Remove
    button. We do
    this to
    reorganize the
    information in
    the output
    table.

    5a. Click the


    teacheffect_mean
    then click the
    ADD button to
    move the
    variable into the
    Model box.
    b. Now click the
    lowses_mean
    variable then the
    ADD button to
    move the
    variable into the
    Model box.
    c. Finally, click
    the last variable
    lowses then the
    the ADD button.
    Click the CONTINUE button to return to the Linear Mixed Models dialogue box.

    10

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009


    6. The random effects,
    estimation and statistics
    settings remain the
    same so skip over these
    buttons and click the
    OK button to run the
    model.
    If we did wish to define
    a random slope, we
    could open the
    RANDOM dialogue
    box and specify lowses
    as randomly varying
    before running the
    model.

    11

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    Defining Model 4 with PASW/SPSS Commands

    (PASW settings will


    default to those used for
    Model 3.)
    1 .Go to the PASW
    toolbar and select
    ANALYZE, MIXED
    MODELS, LINEAR.
    This command opens the
    Linear Mixed Models:
    Specified Subjects and
    Repeated dialogue box.

    12

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009


    2. The Linear Mixed Models:
    Specify Subjects and Repeated
    dialogue box, displays schcode in
    the Subjects box.
    Click the CONTINUE button to
    display the Linear Mixed Models
    dialogue box.

    3. The Linear Mixed


    Models dialogue box
    shows the default
    setting from Model 3.
    We will continue using
    the same variables and
    Fixed Effects settings
    so click the RANDOM
    button to access the
    Random Effects screen.

    4. The Random Effect 1 of 1 screen displays Variance Components as the covariance matrix.

    13

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009


    a. Change the
    covariance type
    to Diagonal by
    clicking the pulldown menu to
    make the
    selection.
    b. Click the pulldown menu to
    change the
    factorial setting
    to Main Effects.
    c. Click to select
    the lowses
    variable from the
    Factors and
    Covariates box
    then click the
    ADD button to
    move the
    variable into the
    Model box.
    Click the CONTINUE button to return to the Linear Mixed Models dialogue box.
    5. The estimation and
    statistics settings
    remain the same so
    skip over these buttons
    and click OK to run the
    model.

    14

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    Defining Model 5 with PASW/SPSS Commands

    (PASW settings will


    default to those used for
    the preceding Model 4.)
    1. Go to the PASW
    toolbar and select
    ANALYZE, MIXED
    MODELS, LINEAR.
    This command opens the
    Linear Mixed Models:
    Specified Subjects and
    Repeated dialogue box.

    2. The Linear Mixed Models:


    Specify Subjects and Repeated
    dialogue box, displays schcode in
    the Subjects box.
    Click the CONTINUE button to
    display the Linear Mixed Models
    dialogue box.

    15

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    16

    3. The Linear Mixed


    Models dialogue box
    displays the settings
    used for Model 4.
    We will continue using
    the same variables.
    Click the FIXED
    button to access the
    Linear Mixed Models:
    Fixed Effects dialogue
    box.

    4. Two cross-interactions will be created from the Linear Mixed Models: Fixed Effects dialogue
    box: teacheffect_mean*lowses and lowses_mean*lowses.
    a. Begin by
    clicking to select
    the Build nested
    terms option.
    b. Click to select
    the variable
    teacheffect_mean
    from the Factors
    and Covariates
    box.
    c. Then click the
    down arrow button
    which will insert
    teacheffect_mean
    into the Build
    Term construction
    section.
    d. Next, click the BY* button which will insert the command symbol (*) to the term so it appears
    as: teacheffect_mean*

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    17

    e. Now click to
    select lowses from
    the Factors and
    Covariates box.
    f. Then click the
    down arrow button
    which will insert
    lowses after the *
    (BY) term in the
    Build Term
    section. This
    completes the
    interaction
    command:
    teacheffect_mean*
    lowses.
    g. Click the ADD button to move the interaction teacheffect_mean*lowses into the Model box.

    5. The second cross-level interaction to create is: lowses_mean*lowses


    a. Click to select
    the variable
    lowses_mean from
    the Factors and
    Covariates box.
    b. Then click the
    down arrow button
    which will insert
    lowses_mean into
    the Build Term
    construction area.
    c. Click the BY*
    button to insert the
    computation
    command symbol
    (*) into the term so
    it appears as: lowses_mean*

    Ronald H. Heck, Scott L. Thomas, Lynn Tabata Fall 2009

    18

    d. Next, click to select


    lowses from the
    Factors and
    Covariates box.
    e. Then click the down
    arrow button to insert
    lowses into the Build
    Term construction
    area. This completes
    the interaction:
    lowses_mean*lowsess
    f. Click the ADD
    button to move the
    interaction
    lowses_mean*lowses.
    Click the CONTINUE button to return to the Linear Mixed Models dialogue box. Next click the
    RANDOM button to access Linear Mixed Models: Random Effects.

    6. The estimation and


    statistics settings
    remain the same as
    those used for Model 4
    so skip over these
    buttons and click OK to
    run the model.

    You might also like