* Encoding: UTF-8.
***Basic data cleaning.

Sort Cases by Progress(A).

Frequencies Variables=SIS.

*Eliminate case that did not agree via SIS.

Select If SIS ne 2.
Execute.

*Eliminate cases that failed the truthfulness check or that did not reach that check (quit before).

Frequencies Variables=Latveria.

Select If Latveria = 1.
Execute.

*Eliminate further cases based on the telephone check or that did not reach that check (quit before).

Frequencies Variables=Telephone.

Select If Telephone = 2.
Execute.

*Eliminate any cases that failed the color attention check or that did not reach that check (after frequencies, no issues identified).

Frequencies Variables = Color_1 to Color_5.

*Eliminate anyone who failed the second attention check or who did not reach that check.

Frequencies Variables = Trust1a_3.

Select If Trust1a_3=3.
Execute.

*Because this is a pilot test, eliminate anyone who did not progress past the image review.

Frequencies Variables=Progress.

Select If Progress>81.
Execute.

***Recode Trust in Science scale per author instructions.

Recode Trust1a_1 Trust1a_2 Trust1a_4 Trust1a_5 Trust1a_7 Trust1a_9 Trust1b_3 Trust1b_7 Trust1b_8 Trust1b_9 Trust1b_10 Trust1b_11 (5=1) (4=2) (2=4) (1=5).
Execute.

Recode Trust2a_1 Trust2a_2 Trust2a_3 Trust2a_4 Trust2a_6 Trust2a_8 Trust2b_3 Trust2b_7 Trust2b_8 Trust2b_9 Trust2b_10 Trust2b_11 (5=1) (4=2) (2=4) (1=5).
Execute.

***Check reliability of trust in science scale for total sample.

RELIABILITY
  /VARIABLES=Trust1a_1 Trust1a_2 Trust1a_4 Trust1a_5 Trust1a_6 Trust1a_7 Trust1a_8 Trust1a_9 
    Trust1a_10 Trust1a_11 Trust1b_1 Trust1b_2 Trust1b_3 Trust1b_4 Trust1b_5 Trust1b_6 Trust1b_7 
    Trust1b_8 Trust1b_9 Trust1b_10 Trust1b_11
  /SCALE('Trust in Science') ALL
  /MODEL=ALPHA
  /STATISTICS=SCALE CORR
  /SUMMARY=TOTAL.

***Create scale.

Compute TrustPre=(Trust1a_1+Trust1a_2+(SUM(Trust1a_4 to Trust1b_11)))/21.
Execute.

Descriptives Variables=TrustPre
  /Statistics=Mean Stddev Min Max.

*Assign arms.

Compute Arm = 0.
Execute.

If (Timing1_First_Click ge 0) Arm = 1.
If (Timing2_First_Click ge 0) Arm = 2.
If (Timing3_First_Click ge 0) Arm = 3.
If (Timing4_First_Click ge 0) Arm = 4.
If (Timing5_First_Click ge 0) Arm = 5.

Frequencies Variables Arm.

Sort Cases by Arm.
Split File Separate by Arm.

Descriptives Variables=TrustPre
  /Statistics=Mean Stddev Min Max.

***Check reliability of posttest trust scale.

Split File Off.

RELIABILITY
  /VARIABLES=Trust2a_1 to Trust2b_11 
  /SCALE('Trust in Science2') ALL
  /MODEL=ALPHA
  /STATISTICS=SCALE CORR
  /SUMMARY=TOTAL.

***Create scale.

Compute TrustPost=(SUM(Trust2a_1 to Trust2b_11))/21.
Execute.

Descriptives Variables=TrustPost
  /Statistics=Mean Stddev Min Max.

***Review Changes in Trust over TIme.

Split File Separate by Arm.

Descriptives Variables=TrustPost
 /Statistics=Mean Stddev Min Max.

T-Test Pairs=TrustPre with TrustPost (Paired)
  /Criteria=CI(.9500)
  /Missing=Analysis.

***Review narrative believability index

Split File Off.

Recode NBI1_4 NBI1_6 NBI1_8 NBI1_10 (7=1) (6=2) (5=3) (3=5) (2=6) (1=7).
Execute.

RELIABILITY
  /VARIABLES=NBI1_1 to NBI2_2 
  /SCALE('Believability') ALL
  /MODEL=ALPHA
  /STATISTICS=SCALE CORR
  /SUMMARY=TOTAL.

Compute Believe=(SUM(NBI1_1 to NBI2_2)/12).
Execute.

Descriptives Variables=Believe
  /Statistics=Mean Stddev Min Max.

ONEWAY Believe BY Arm
  /MISSING ANALYSIS
  /POSTHOC=TUKEY GH ALPHA(0.05).

Split File Separate by Arm.

Descriptives Variables=Believe
  /Statistics=Mean Stddev Min Max.

***Print out qualitative data by arm.

Frequencies Variables=Description.

***Compute subscales for reviewers.

Split File Off.

Compute Plausibility=((NBI1_1+NBI1_5+NBI1_9)/3).
Compute Completeness=((NBI1_2+NBI1_6+NBI1_10)/3).
Compute Consistency=((NBI1_3+NBI1_7+NBI2_1)/3).
Compute Coverage=((NBI1_4+NBI1_8+NBI2_2)/3).
Execute.

Descriptives Variables=Plausibility Completeness Consistency Coverage
  /Statistics=Mean Stddev Min Max.