Biochemical methods to verify abstinence
Especially with youth and pregnant women, it is important to verify abstinence at follow-up contacts using biochemical methods, like carbon monoxide monitoring and serum and urine cotinine tests. The validity of self-reports is open to serious question.

Confidence intervals
When information from a sample population is used to describe an entire population, confidence intervals indicate the range of confidence expected in making that projection. A confidence interval is indicated by two numbers usually set in parentheses and separated by a comma, for example (44, 50).

If an opinion poll indicates that candidate A is supported by 47% of the people surveyed, the poll takers describe how certain they are of their estimation with the confidence interval. While it's not known exactly how many people in the total population support candidate A, a confidence interval of (44, 50) would indicate a 95% confidence that the true percentage of support in the population is somewhere between 44% and 50%. This is what the footnotes on opinion poll results refer to when they say, "plus or minus 3 percentage points."

The 95% confidence interval is the one used most often (it is linked to the P value of .05 discussed in statistically significant results), but occasionally one sees 90% or 99% confidence intervals.

Consistent with other studies
The results of a study being considered should be similar to the results found in other studies that evaluated the same type of tobacco cessation intervention.

Double-blind study
In this type of study, neither the participants nor the people conducting the program and its evaluation know which individuals participating in the study are in the experiment and which are in the control group.

Drop-out/attrition rate
When more than 30% of the participants in a study drop out, or are not measured for abstinence at follow-up contacts, the results may be biased. The intervention can't be fairly evaluated for effectiveness because it is unknown if participants remain abstinent after the intervention. The data is insufficient.

Evaluation
A study of the study that determines if a program or intervention causes the desired changes in the target population.

Follow-up measurements
To determine if a tobacco cessation intervention is effective in helping tobacco users quit, researchers need to check for abstinence at intervals after the intervention. In general, a larger group of participants would quit using tobacco right after the intervention; however, with time, some participants relapse, and the effect of the intervention decreases. Therefore, the farther away from the intervention the measurements are made, the more we know if the intervention is effective in producing long-term abstinence.