Squiers 2017.

Methods	Study design: RCT Country: USA Recruitment: online advertising and search ads via Facebook, Craigslist, Pandora, and Google, Yahoo! and Bing, plus email recruitment via market research panels Study dates: 2013‐14
Participants	Baseline characteristics (n = 4027) Age 18‐21: 20.8% (N = 839) Female: 70.2% (N = 2825) Less than high school: 5.6% (N = 225) Time to first cigarette 5 min: 27.6% (N = 1110) White: 73.7% (N = 2967) Inclusion criteria: aged 18‐29 years. Reside in the USA. Smoke cigarettes at least 5 days/month. Be interested in quitting cigarette use. Not be involved in a cessation programme. Have an active email account. Be able to receive text messages on their mobile phone. Be the only member of your household participating in this study. Be willing to share contact information with the study team in order to share information about the study on a timely basis. Exclusion criteria: didn't complete baseline survey. The anti‐fraud process included automated duplication checks of phone numbers, email addresses, and IP addresses. If duplicates were detected, the individual was excluded from the study. Failure to provide informed consent. Refused the honesty pledge. Faced technical difficulties (e.g. undelivered text messages). Texted “STOP” at any point before their QD. Opted out of the study entirely by notifying the project team, typically via email
Interventions	Arm 1: periodic cessation assessments and QD reminder messages (total 11 messages) Arm 2: arm 1 messages above, plus motivational preparatory messages for 2 weeks prior to participants’ QD (total 40 messages) Arm 3: all of the messages from Arms 1 and 2 above, plus 6 weeks of follow‐up post‐QD messages (total 127 messages)
Outcomes	Definition of abstinence: self‐reported 7‐day repeat point prevalence abstinence at 32 weeks
Funding source	National Institutes of Health, National Cancer Institute HHSN261201400002B, HHSN26100006, HHSN26100007
Conflicts of interest	None
Notes	We split Arm 1 and compared it with arms 2 and 3 in the comparison of higher‐ versus lower‐frequency text messaging
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information about the sequence generation process
Allocation concealment (selection bias)	Low risk	Not stated but as all done online it is very unlikely to have any bias
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Minimal contact with investigators
Incomplete outcome data (attrition bias) All outcomes	Low risk	There was a 64.4% response rate overall. Arm 1: 846/1313; Arm 2: 933/1400; Arm 3: 824/1314. ITT analysis