Millisecond Forums

IAT SPSS Syntax: Replacing Labels to Calculate D_biep

https://forums.millisecond.com/Topic21808.aspx

By thaphthia - 6/9/2017

Hi everyone,

I am looking to score my IAT using Greenwald's 'D_biep.Generic Inquisit IAT.SPSS form.29Dec07' scoring Syntax. I am having a bit of trouble figuring out how to replace the Syntax labels appropriately, and I am wondering if someone who has done this before would be able to help me out!

In the Syntax, it is stated that: *D_biep will be scored so that higher numbers indicate stronger association of <<attribute_left_label>> with <<TARGET_2_LABEL>>, unless the scoring is reversed.

For the purposes of my analyses, and if at all possible, I would like to score my data such that I get a D_biep that will indicate the strength association between 'fat' and 'SELF' (i.e., such that higher scores indicated greater heavyweight self-identity), and another scoring of my data such that I get a D_biep that will indicate the strength of association between 'thin' and SELF (i.e., such that higher scores indicate greater lightweight self-identity).

 https://www.millisecond.com/forums/Uploads/Images/70d558a9-f0ed-49d1-8c13-ae43.png
  
Referring to the Syntax below, I understand that SELF/OTHER are my 'TARGETS'. However, I would like some help figuring out which is my 'TARGET_1_LABEL'  and which is my 'TARGET_2_LABEL', given that I would like my analyses to focus on 'SELF' (as opposed to 'OTHER').

I also understand that fat/thin are my 'attributes'.  I am wondering if there would be a way to manipulate the Syntax such that one analysis focuses on the association between SELF with 'thin', whereas another analysis focuses on the association between SELF with 'fat'.

Basically, I would like some help with appropriately replacing the labels in the following Syntax to fit with my data and goals for interpreting D_biep.

* replace the labels below with ones meaningful for the targets and attributes used.
VALUE LABELS
   pairing
    0 'single task practice'
    1 '«attribute_left_label»+«TARGET_2_LABEL» and «attribute_right_label»+«TARGET_1_LABEL»'
    2 '«attribute_left_label»+«TARGET_1_LABEL» and «attribute_right_label»+«TARGET_2_LABEL»'
 / order
    1 '«attribute_left_label»+«TARGET_2_LABEL» and «attribute_right_label»+«TARGET_1_LABEL» first'
    2 '«attribute_left_label»+«TARGET_1_LABEL» and «attribute_right_label»+«TARGET_2_LABEL» first' .

I am also unsure about whether the following reverse scoring Syntax will be necessary?

  *************************************************************
** TO REVERSE THE SCORING OF THE IAT, REMOVE THE ASTERISKS **
** FROM THE NEXT FOUR COMMAND LINES.                       **
************************************************************* .

* COMPUTE D_biep = 0 - D_biep .
* COMPUTE D_biep_a = 0 - D_biep_a .
* COMPUTE D_biep_b = 0 - D_biep_b .
* VARIABLE LABELS
    D_biep   '«attribute_left_label»+«TARGET_1_LABEL» association all blocks'
  / D_biep_a '«attribute_left_label»+«TARGET_1_LABEL» association 1st blocka'
  / D_biep_b '«attribute_left_label»+«TARGET_1_LABEL» association 2nd blocka' .

FORMAT D_biep D_BIEP_a D_BIEP_b D_BIEP (F6.3) SUBJECT (F7.0)
    Na Nb (F4.0) M1a M2a M1b M2b (F7.1)
    ERR1a ERR2a ERR1b ERR2b ERR_1 ERR_2 (F5.1) .

SAVE OUTFILE = 'D_BIEP_temp.SAV' .
DESCRIPTIVES ALL .


Any assistance would be greatly appreciated! 
By Dave - 6/9/2017

thaphthia - Friday, June 9, 2017
Hi everyone,

I am looking to score my IAT using Greenwald's 'D_biep.Generic Inquisit IAT.SPSS form.29Dec07' scoring Syntax. I am having a bit of trouble figuring out how to replace the Syntax labels appropriately, and I am wondering if someone who has done this before would be able to help me out!

In the Syntax, it is stated that: *D_biep will be scored so that higher numbers indicate stronger association of <<attribute_left_label>> with <<TARGET_2_LABEL>>, unless the scoring is reversed.

For the purposes of my analyses, and if at all possible, I would like to score my data such that I get a D_biep that will indicate the strength association between 'fat' and 'SELF' (i.e., such that higher scores indicated greater heavyweight self-identity), and another scoring of my data such that I get a D_biep that will indicate the strength of association between 'thin' and SELF (i.e., such that higher scores indicate greater lightweight self-identity).

 https://www.millisecond.com/forums/Uploads/Images/70d558a9-f0ed-49d1-8c13-ae43.png
  
Referring to the Syntax below, I understand that SELF/OTHER are my 'TARGETS'. However, I would like some help figuring out which is my 'TARGET_1_LABEL'  and which is my 'TARGET_2_LABEL', given that I would like my analyses to focus on 'SELF' (as opposed to 'OTHER').

I also understand that fat/thin are my 'attributes'.  I am wondering if there would be a way to manipulate the Syntax such that one analysis focuses on the association between SELF with 'thin', whereas another analysis focuses on the association between SELF with 'fat'.

Basically, I would like some help with appropriately replacing the labels in the following Syntax to fit with my data and goals for interpreting D_biep.

* replace the labels below with ones meaningful for the targets and attributes used.
VALUE LABELS
   pairing
    0 'single task practice'
    1 '«attribute_left_label»+«TARGET_2_LABEL» and «attribute_right_label»+«TARGET_1_LABEL»'
    2 '«attribute_left_label»+«TARGET_1_LABEL» and «attribute_right_label»+«TARGET_2_LABEL»'
 / order
    1 '«attribute_left_label»+«TARGET_2_LABEL» and «attribute_right_label»+«TARGET_1_LABEL» first'
    2 '«attribute_left_label»+«TARGET_1_LABEL» and «attribute_right_label»+«TARGET_2_LABEL» first' .

I am also unsure about whether the following reverse scoring Syntax will be necessary?

  *************************************************************
** TO REVERSE THE SCORING OF THE IAT, REMOVE THE ASTERISKS **
** FROM THE NEXT FOUR COMMAND LINES.                       **
************************************************************* .

* COMPUTE D_biep = 0 - D_biep .
* COMPUTE D_biep_a = 0 - D_biep_a .
* COMPUTE D_biep_b = 0 - D_biep_b .
* VARIABLE LABELS
    D_biep   '«attribute_left_label»+«TARGET_1_LABEL» association all blocks'
  / D_biep_a '«attribute_left_label»+«TARGET_1_LABEL» association 1st blocka'
  / D_biep_b '«attribute_left_label»+«TARGET_1_LABEL» association 2nd blocka' .

FORMAT D_biep D_BIEP_a D_BIEP_b D_BIEP (F6.3) SUBJECT (F7.0)
    Na Nb (F4.0) M1a M2a M1b M2b (F7.1)
    ERR1a ERR2a ERR1b ERR2b ERR_1 ERR_2 (F5.1) .

SAVE OUTFILE = 'D_BIEP_temp.SAV' .
DESCRIPTIVES ALL .


Any assistance would be greatly appreciated! 

> I would like some help figuring out which is my 'TARGET_1_LABEL'  and which is my 'TARGET_2_LABEL'

You'll have to consult the script you used to collect the data. Also see https://www.millisecond.com/forums/Topic3444.aspx

> I also understand that fat/thin are my 'attributes'.  I am wondering if there would be a way to manipulate the Syntax such that one analysis focuses on the association
> between SELF with 'thin', whereas another analysis focuses on the association between SELF with 'fat'.

But these come down to the same thing, no? The only difference would be the sign of the D-score, it's absolute value would be the same.