Online Supplement A: Data Extraction Template

Online Supplement A: Data Extraction Template

AUTHORS (year, study number)

TYPE: In what type of publication is the study reported?

CONFERENCE ABSTRACT
JOURNAL ARTICLE
DOCTORAL THESIS
OTHER (details below)

FRAMING_DESIGN: All of the studies compare gain (emphasize the benefits of complying with the recommended behaviour) and loss frames (emphasize the costs of not complying with the recommended behaviour). Was this manipulation between or within-Ps?

BETWEEN-Ps
WITHIN-Ps

WITHIN_COUNTERBALANCING: If WITHIN-Ps was the order of the framing manipulations counterbalanced or randomised?

YES
NO

UNIT_OF_ALLOCATION: If BETWEEN-Ps at what unit have participants been assigned to the framing conditions?

PARTICIPANT
GROUP
NO INFORMATION PROVIDED

RANDOMISATION_METHOD_DESCRIBED: If BETWEEN-Ps was the method of randomisation explicitly/ clearly described?

YES
NO

If YES give details:

FRAME_DESCRIPTIONS: Were the framing interventions explicitly/clearly described?

YES
NO

Provide brief details of the gain/loss frames below:

GAIN_FRAME
LOSS_FRAME

CONTROL_FRAME: Were the gain and loss frames compared against a control or neutral framing condition?

YES
NO

If YES give details:

INTERVENTION_MEDIA: What type of media was used to deliver the framing manipulations?

WRITTEN
VIDEO
SLIDE SHOW/ POWERPOINT
AUDIO
OTHER (give details below)

INTERVENTION_DELIVERY: Did the authors take any steps to ensure, or conduct any analyses to check, that participants received and attended to the framing intervention?

YES
NO

If YES give details:

POPULATION: Describe the type of population that were recruited for this study – with details of how they were recruited.

ELIGIBILITY_CRITERIA: Were the eligibility criteria specified?

YES
NO

If YES give details:

DEMOGRAPHICS: Enter the following details (if provided) about the participants below

MEAN_AGE
MIN_AGE
MAX_AGE
PERCENT_M
PERCENT_F

POWER_CALCULATION_SAMPLE_SIZE: Was the sample size for each condition determined by means of a power calculation?

YES
NO

If YES give details:

SAMPLE_SIZE: Enter in the table the time point(s) at which participants were followed up post-intervention from 0 (immediate) to X months. Also enter (if data provided) the sample sizes in total and/or in each framing condition.

BASELINE / FOLLOW_UP1 / FOLLOW_UP2 / FOLLOW_UP3
TIME / ----
N_TOTAL
N_GAIN
N_LOSS
N_CONTROL

FOLLOWUP_TIMING: Was the timing of the follow-up measurements comparable in all conditions?

YES
NO

DROPOUT_REASONS: If there was a loss of participants between baseline and follow-up(s) were reasons provided?

YES
NO
N/A (NO DROPOUTS)

If YES give details:

DROPOUT_ANALYSIS: If there was a loss of participants between baseline and follow-up(s) were the characteristics of the dropouts compared with the completers of the study?

YES
NO
N/A (NO DROPOUTS)

If YES give details:

BEHAVIOR(S): List the specific behaviour or behaviours that the framing intervention aimed to address

OUTCOME_MEASURE(S): What type(s) of outcome measure(s) were used test the effects of the framing intervention? Enter details for all measures used.

X if used / Number of items / Cronbach’s alpha
ATTITUDES
INTENTIONS
COMPOSITE ATTITUDE-INTENTION MEASURE
BEHAVIOR
PERCEIVED PERSUASIVENESS
OTHER (enter details below)

OUTCOME_METHOD: Which of the following were used to collect the outcome measure(s)? X all that apply

SELF-COMPLETION QUESTIONAIRE
TELEPHONE INTERVIEW
FACE-TO-FACE INTERVIEW
OTHER (give details below)

BLIND_ASSESSMENT: If TELEPHONE INTERVIEW or FACE-TO-FACE INTERVIEW was the interviewer blinded to which framing intervention the participant received?

YES
NO
NO INFORMATION

GROUP_DIFFS_AT_BASELINE: Were any analyses conducted to compare or control for pre-intervention differences between framing conditions?

YES
NO

If YES give details:

INTENTION_TO_TREAT: If there was a loss of participants between baseline and follow-up(s) did the analysis include an intention to treat analysis (i.e., were all participants randomised to conditions – including those who were lost at followup - included in the analysis of the effectiveness of the framing interventions)?

YES
NO
N/A (NO DROPOUTS)

FRAMING_MAIN_EFFECT

Include results of analysis of the main effect of the framing manipulation. Describe tests used and include any relevant inferential/ descriptive statistics to enable effect size calculations (e.g., F-tests/ t-tests, means/ SDs etc.)

INTERACTION_EFFECTS

Include results of any tests of two-way or three-way interactions (which include framing) that have been reported. Describe tests used and include any relevant inferential/ descriptive statistics to enable effect size calculations (e.g., F-tests/ t-tests, means/ SDs etc.)

Online supplement B: Method used to calculate quality scores

Quality scores were calculated for each study using a set of methodological criteria adapted from previously published systematic reviews conducted by van den Berg et al. (2007) and Edwards et al. (2001). Although assigning equal weight to each criterion is somewhat arbitrary the scores provide a guide about the weight that might be attached to the results of individual papers. The maximum score that can be obtained is 13 (the study is scored 1 on any criteria that are not relevant – e.g., if there was no loss of participants between baseline and follow-up the study would be assigned a score of 1 for Q9). The quality scores ranged from 5.0 to 10.5 (M=7.74, SD=1.14)

Was the sample size recruited sufficient to power the experiment to detect either a medium effect size r≤0.3 (1 point) or small effect size r≤0.1 (2 points)?
(Either) If the experiment was a between-participants design was assignment to framing conditions at the level of participants (rather than groups)? (or) If the experiment was a within-participant design was the order of the framing manipulations counterbalanced or randomized?
Was the method used to generate the random allocation sequence described (0.5) and did it ensure that the intervention allocations were concealed at study enrolment (0.5)?
Were the gain and loss framing conditions explicitly/ clearly described?
Did the authors take any steps to ensure, or conduct any analyses to check, that participants received and attended to the framing intervention?
Was a follow-up measurement performed ≥ 1 week after randomization?
Was the timing of the follow-up measurements comparable in all conditions?
Were there fewer than 10% dropouts between baseline and follow-ups?
If there was a loss of participants between baseline and follow-ups were the characteristics of the dropouts compared with the completers of the study (0.5) and/or was an intention to treat analysis conducted (0.5)?
Was a measure of behavior included (0.5) and was it objective (rather than self-report) (0.5)?
Were outcome assessors blind from knowledge of which intervention a participant received?
Were any analyses conducted to compare or control for pre-intervention differences between framing conditions?