## EE: Nicotine (2005)

**Definitions:**

**Pack-years – **Is determined by multiplying the number of cigarettes per day by the number of years smoking and divided by 20 (i.e., the number of cigarettes in a pack.

**Fisher’s Least Significant Difference** (LSD; also known as Fisher’s protected test). Is an analysis of variance test against the alternative hypothesis that at least one of the means differs from the rest; it is analogous to a two-sample t-test for pair-wise comparisons in an ANOVA,. However, the test uses the variance estimate from the ANOVA rather than from only the two groups being compared (i.e., uses a pooled estimator of the population variance from all the samples used, rather than the pooled sample variance from samples “*i*” and “*j*”). It is identical to the a-priori test, except that it requires a significant omnibus F. At one time, the protected procedure was thought to control the Type 1 error rate (i.e., inappropriate rejection of the null hypothesis) via the omnibus test and would still have good power if you were allowed to proceed to the LSD test (i.e., a significant F). However, in 1993, it is reported that the protected test does NOT control the Type 1 error rate and is not recommended (See Zwick R. Pairwise comparison procedures for one-way analysis of variance designs. In Keren G & Lewis C (*Eds*). *A handbook of data analysis in the behavioral sciences*. Lawrence Erlbaum Associates. 1993. pp 43-71).

**Nicotine Exposure Definitions** (arbitrarily assigned within project evidence analysis; and, when nicotine content/cigarette was unknown, used estimate of 0.9 mg nicotine/cigarette to aid with comparisons).

**High** exposure (i.e., dose): >/= 11.0 mg nicotine/exposure (or ~>/= 12 cigarettes)

**Moderately high** exposure (i.e., dose): >/= 3 mg to < 11.0 mg nicotine/exposure (or ~ 3 – 11 cigarettes)

**Moderate** exposure (i.e., dose): >/= 1 mg to < 3 mg/exposure (or ~1-2 cigarettes)

**Low** exposure (i.e., dose): < 1 mg nicotine/exposure (or ~1/2 cigarette)