Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 May 1.
Published in final edited form as: Drug Alcohol Depend. 2022 Mar 8;234:109397. doi: 10.1016/j.drugalcdep.2022.109397

Construct Validity of the Cigarette Ratings Scale and Associations with Tobacco Use and Product Feature Outcomes

Andrea C Johnson 1,2, Melissa Mercincavage 1,2, Valentina Souprountchouk 1,2, Anupreet K Sidhu 1,2, Andrea C Villanti 2,3, Cristine D Delnevo 2,3, Andrew A Strasser 1,2,4
PMCID: PMC9018530  NIHMSID: NIHMS1788149  PMID: 35287035

Abstract

Background:

Subjective ratings are inconsistently associated with behavioral outcomes such as tobacco use and there is no current standard. The Cigarette Ratings Scale is an ideal measure for further evaluation because it has been widely used in tobacco regulatory science and tobacco industry research.

Purpose:

This study investigated the construct validity of the Cigarette Ratings Scale and associations with tobacco use and product feature outcomes.

Methods:

Using secondary analysis of baseline data from five research trials, we conducted an exploratory factor analysis in one sample and validated the factor solution in a second sample. We then examined the relationship of the averaged subscales with tobacco outcomes and cigarette product features among current adult cigarette smokers (N=752) who smoked ≥5 cigarettes daily for ≥5 years.

Results:

The results supported a three-factor solution: 1. Product harshness evaluation, 2. Smoking satisfaction, and 3. Positive sensory experience. Multivariable general linear models indicated that cigarettes per day was associated with a lower harshness rating b=−0.29 (95% CI: −0.51, −0.07) and higher positive sensory experience b=0.32 (95% CI: 0.08, 0.56). FTND average dependence scores were associated with a more positive sensory experience b=1.08 (95% CI: 0.28, 1.89). CO boost was associated with smoking satisfaction b=0.77 (95% CI: 0.30, 1.26).

Conclusions:

The Cigarette Ratings Scale subscales were primarily associated with behavioral outcomes, biological exposure, and nicotine dependence. This can help addiction efforts to determine how subjective evaluations of tobacco products relate to use behaviors.

Keywords: tobacco, cigarettes, subjective ratings, cigarette ratings scale

1. Introduction

Cigarette smoking is the leading cause of death globally (U.S. Department of Health and Human Services, 2014). Evidence indicates tobacco marketing is associated with individual level outcomes (e.g., positive attitudes toward smoking) that lead to tobacco use and nicotine addiction (Flay, 1999; National Cancer Institute, n.d.). For instance, cigarette companies previously marketed their products with “light” and “low” descriptors that were shown to be misleading, and such explicit descriptors are now prohibited (Government Printing Office, 2009). Product and packaging design features such as color or flavor, size, and brand imagery are implicit approaches to circumvent existing marketing restrictions (U.S. Department of Health and Human Services, 2014). These features have been shown to be associated with perceptions and subjective ratings of a product that enable the illusion of lower risk and better taste (e.g., red packages suggesting full flavor) (National Cancer Institute, n.d.). Yet, study designs and theoretical frameworks generally assume that positive perceptions, expectations, and subjective ratings are universally reinforcing and a prominent driver of cigarette use leading to nicotine addiction.

Clinical trials studying cigarettes often measure participants’ subjective ratings or responses to capture the self-reported experience of using a product. A 2009 review showed there is no current standard in the field of addiction studying tobacco regulatory science, and a variety of subjective rating domains and measures are reported in the literature (Hanson et al., 2009). This commonly includes product evaluation ratings (Hanson et al., 2009) and several scales exist (e.g., Modified Cigarette Evaluation Questionnaire, Product Evaluation Scale) that capture constructs such as satisfaction, psychological reward, aversion, craving reduction, and enjoyment (Cappelleri et al., 2007; Hatsukami et al., 2013). Furthermore, the current evidence of these scales in predicting tobacco use outcomes is mixed and overall there is a dearth of evidence to demonstrate a direct relationship between subjective responses and behavioral outcomes relevant to tobacco control.

Subjective ratings have inconsistently been associated with behavioral outcomes including self-reported (e.g., cigarettes per day) or objective (e.g., carbon monoxide levels) tobacco use. Some work shows that satisfaction (e.g., tastes good) with a reduced nicotine cigarette is associated with greater clinical measures of behavior (e.g., expired carbon monoxide, puff choice) as well (Karelitz & Perkins, 2021; Smith et al., 2019). However, other subjective items (e.g., taste, strength) have been shown to moderate the associations with reduced nicotine cigarette use when accounting for beliefs and age (Cassidy et al., 2019; Mercincavage, Saddleson, et al., 2017). It is difficult to discern the effects with conceptual overlap among items (e.g., taste). Furthermore, existing intervention and randomized controlled trials have primarily used subjective rating scales to evaluate reduced nicotine or reduced exposure products. Fewer recent studies have assessed subjective ratings for one’s usual cigarette brand that individuals are drawn to and use regularly in natural settings (Mercincavage et al., 2016, 2018, 2020; Strasser et al., 2013).

The lack of consistent findings indicate subjective ratings would benefit from further psychometric evaluation. Given the need for common measures across tobacco studies, the PhenX Toolkit is a valued resource (O’Connor et al., 2020; Piper et al., 2020). Yet, there are no consensus measures in the PhenX Toolkit for subjective ratings of tobacco products (O’Connor et al., 2020; Piper et al., 2020). One of the existing subjective rating scales, the Cigarette Ratings Scale, has been widely used in tobacco regulatory science and tobacco industry research. It encapsulates a user’s experience with their usual brand of cigarettes that have shown to be associated with appealing tobacco marketing features (Mercincavage et al., 2020; Mercincavage, Saddleson, et al., 2017; Mercincavage, Wileyto, et al., 2017; Strasser et al., 2005, 2007, 2013). Moreover, this dovetails with historical aspects of the scale items and its origins in industry-funded research (Phillip Morris International, n.d.). Therefore, studying the Cigarette Ratings Scale has the potential to add to the body of research on measurement and product marketing and labeling for addictive products such as cigarettes.

The current study aims to assess the construct validity of the Cigarette Ratings Scale. We assess the reliability and validity of the scale when evaluating adult cigarette smokers’ usual cigarette brand, using secondary analysis of baseline data from five randomized trials by combining datasets and randomly sampling two groups. We conduct an exploratory factor analysis on the first group and confirmatory factor analysis on the second group. We then examine associations with tobacco use and product feature outcomes. We hypothesize that subjective ratings will be associated with tobacco use and product feature outcomes.

2. Methods

2.1. Sampling & Design

The current study is a secondary data analysis of baseline data from five research trials (NCT01898559, NCT01898507, NCT01202942, NCT02301351). The trials studied combustible tobacco products and their marketing, sampled adults ages 21–65 years that were current cigarette smokers defined by at least 5 cigarettes smoked daily for 5 years or more. Individuals were recruited in an around the Philadelphia, Pennsylvania region. All studies included identical baseline procedures at the start of each study to assess ratings of participants’ usual cigarette brand and four studies excluded menthol smokers (Mercincavage et al., 2016, 2018, 2020; Strasser et al., 2013).

2.1.2. Procedures

For each of the randomized trials, participants completed an initial laboratory visit (Day 0) to provide written informed consent, verify eligibility, and smoke two cigarettes with 45 min between each through topography equipment. All sessions were completed between 8:00 AM and 12:00 PM to control for diurnal variations in cigarette smoking. Participants were not instructed to abstain from smoking, but to smoke as usual leading up to their session. The baseline period standardized the time since last cigarette across participants by having smokers smoke a cigarette at session onset. Prior to smoking the first cigarette, participants completed demographic and self-reported tobacco use outcomes, while research staff recorded cigarette product features of the participant’s usual cigarette brand. For each of the two cigarettes, participants provided pre-cigarette carbon monoxide (CO) readings, smoked a cigarette, completed the Cigarette Ratings Scale, and then four minutes later completed post-cigarette CO readings. Participants smoked their usual cigarette brand during the initial visit (Day 0), for the next five days, and then completed a second in-person visit (Day 5). Day 5 repeated Day 0 procedures with one cigarette prior to each study’s unique intervention procedures. This allows for direct comparisons in subjective responses among studies using identical conditions for three recorded cigarette readings (Day 0, Cigarette 1; Day 0, Cigarette 2; Day 5, Cigarette 1) (Mercincavage et al., 2016, 2018, 2020; Strasser et al., 2013).

2.2. Measures

2.2.1. Cigarette Ratings Scale

The Cigarette Ratings Scale measured 14 items: strength, harshness, heat, draw, taste, satisfaction from smoking, burn rate, taste, too mild, harshness of smoking, after taste, staleness, strength of smoke, and smoke smell. Using a visual analogue scale, participants were instructed to place a vertical line along a 0–100 millimeter line with anchor items at either end for each characteristic. A sample of the scale is outlined in Supplemental Table 1. We reverse coded two items (draw and harshness of smoke) so that higher values were affirmative in nature.

2.2.2. Tobacco use outcomes

Tobacco outcomes included both self-reported use and objective measurements. Self-reported tobacco use included cigarettes per day (CPD) and the Fagerström Test for Nicotine Dependence (FTND) scale at Day 0 prior to the Cigarette Ratings Scale (Heatherton et al., 1991). Objective tobacco use included CO boost in parts per million (ppm) using Vitalograph Inc (Mercincavage et al., 2018). We collected up to three CO readings for each cigarette, before smoking their usual cigarette brand and after filling out the Cigarette Ratings Scale. We took the difference between the pre- and post-reading to calculate a CO boost change score between each of the readings to approximate exposure from a single cigarette (Mercincavage et al., 2018). Collecting CO boost with these procedures allows for greater precision in determining cigarette-specific exposure (Burling et al., 1985).

2.2.3. Product features

Research staff collected information on cigarette product features, including cigarette brand name, size, pack type, flavor or color, and menthol or nonmenthol (Centers for Disease Control and Prevention, 2021).

2.2.4. Demographics

Demographic variables included age, sex, race, ethnicity, education, and income (Mercincavage et al., 2018).

2.3. Analyses

We conducted an exploratory factor analysis in one sample and validated the factor solution in the second sample (Alam et al., 2020; Kim et al., 2021). We first merged the five study datasets. In the event individuals participated in more than one study, their first data contribution was retained while any subsequent contribution(s) were removed (n = 50), so that the final dataset contained the first baseline assessment for N = 752 unique participants. We then randomly assigned individuals into two groups. We determined the first group’s sample size based on a suitable 1:5 case to variable ratio (Hair et al., 2009). This allotted 25% of the sample for Group 1 and 75% of the sample for Group 2 to maximize available power for confirmatory factor analysis. We focused on the Day 0, Cigarette 1 outcomes and include other time points for latent robustness checks. We included all individuals despite attrition between Day 0 and Day 5. Yet, we stratified the two groups by attrition (yes/no) for those that completed Day 0, but did not complete Day 5 or other study procedures due to differences found in prior work (Mercincavage, Wileyto, et al., 2017). We then tested differences between each of the two groups among demographics, tobacco outcomes, cigarette product features, and the Cigarette Ratings Scale. We used Chi-Square tests for categorical variables and t-tests for continuous variables. We used Mplus version 7.4 for the exploratory factor analysis and SAS version 9.4 for all other analyses.

2.3.1. Exploratory Factor Analysis

Once the two groups were deemed equivalent, we conducted an exploratory factor analysis (EFA) on Group 1. We used a common factor analysis approach with maximum likelihood estimation and robust standard errors (Hair et al., 2009). This is a standard extraction approach for determining latent dimensions or constructs (Hair et al., 2009). We extracted factors based on cumulative information of eigenvalues greater than 1, a scree plot, and significant oblique rotated loadings using row standardization at p < .05 (Hair et al., 2009). Loading items above 0.45 were retained to conservatively include 12 items from the original scale. Those at or below 0.45 were not considered to have practical significance (Brown, 2015; Hair et al., 2009).

2.3.2. Confirmatory Factor Analysis

After conducting the EFA, we first describe the factor solution in Group 2. We assess the reliability and validity of the extracted factors. For reliability, we ran Cronbach alphas and McDonald omegas for each cigarette evaluation (Hayes & Coutts, 2020). We also reviewed the factors as latent constructs. We reviewed fit indices and modification indices to optimize suitable fit. To determine the robustness of these variables in latent form we tested nested models for conditional independence and metric measurement invariance across all three measurements (Day 0, Cigarette 1; Day 0, Cigarette 2; Day 5, Cigarette 1) (Brown, 2015; Geiser, 2013). We used standard indices to evaluate cumulative model fit for all measurement models such as the Chi Square test of model fit greater than .05, Root Mean Squared Error of Approximation (RMSEA) less than .05, Standardized Root Mean Square Residual (SRMR) less than .08, and the Comparative Fit Index (CFI) above .90 (Brown, 2015; Geiser, 2013). We used full information maximum likelihood estimation and bootstrapping for latent analyses. We report latent-related results in the Supplemental Materials.

For validity, we examined the relationship of the averaged subscales with tobacco outcomes and cigarette product features. We assessed bivariate differences and report ANOVA least square means for categorical variables or Pearson correlations for continuous variables. We also used multivariable general linear models to test the association of the scales with self-reported (CPD, FTND) and objective (CO boost) tobacco outcomes. We included a categorical variable for study as a covariate in all models. To avoid multicollinearity concerns and maintain parsimonious models, this was the only covariate used in modeling provided it was statistically significant in relation to several demographics, tobacco outcomes, and cigarette product features (e.g., age, race, education, income, cigarettes per day, cigarette brand, pack size, type, and flavor, and menthol status, p’s <.05). We conducted all analyses using standard extraction and validation method recommendations (Brown, 2015; Geiser, 2013; Hair et al., 2009).

3. Results

Combining the five clinical trial datasets among those participants completing one study yielded a total sample of N = 752. We stratified by attrition and formed the two subsamples, including Group 1 (n = 188) and Group 2 (n = 564). Table 1 shows no differences between the groups on sample characteristics including demographics, tobacco outcomes, and cigarette product features. The sample characteristics are generally reflective of demographic estimates in the United States (Census Bureau, 2019) including 26% Non-White individuals, though with lower estimates for Hispanics (3%) and females (27%). Table 2 shows the Cigarette Ratings Scale averages for each item on Day 0, Cigarette 1. There were no statistically significant differences between the groups on each of the items. As expected, the study variable differed by attrition, age, race, education, cigarettes per day, cigarette brand, size, type, flavor or color, and menthol status (p’s <.05).

Table 1.

Sample Characteristics

Variable [min-max] Full Sample (N=752) N(%) or M(SD) Group 1 (N=188) N(%) or M(SD) Group 2 (N=564) N(%) or M(SD) p value
Year Enrolled [2007–2019] 2014 (3.3) 2014 (3.3) 2014 (3.4) 0.70
Study 0.47
 Study 1 91 (12.1%) 25 (13.3%) 66 (11.7%)
 Study 2 81 (10.8%) 22 (11.7%) 59 (10.5%)
 Study 3 265 (35.2%) 73 (38.8%) 192 (34.0%)
 Study 4 273 (36.3%) 60 (31.9%) 213 (37.8%)
 Study 5 42 (5.6%) 8 (4.3%) 34 (6.0%)
Age [21–65] 40.5 (11.8) 39.8 (12.0) 40.7 (11.8) 0.36
Race 0.37
 American Indian/Alaska Native 4 (0.5%) 1 (0.5%) 3 (0.5%)
 Asian 14 (1.9%) 2 (1.1%) 12 (2.2%)
 Black/African American 140 (18.9%) 34 (18.4%) 106 (19.0%)
 Native Hawaiian/Pacific Islander 1 (0.1%) 1 (0.5%) 0 (0.0%)
 White 547 (73.7%) 141 (76.2%) 406 (72.9%)
 More than one race 36 (4.9%) 6 (3.2%) 30 (5.4%)
Ethnicity 0.90
 Hispanic 25 (3.3%) 6 (3.2%) 19 (3.4%)
 Non-Hispanic 725 (96.7%) 182 (96.8%) 543 (96.6%)
Sex 0.97
 Female 203 (27.3%) 50 (27.2%) 153 (27.3%)
 Male 541 (71.7%) 134 (72.8%) 407 (72.7%)
Education 0.18
 High school graduate or less 250 (33.2%) 55 (29.3%) 195 (34.6%)
 Some college or beyond 502 (66.8%) 133 (70.7%) 369 (65.4%)
Income 0.24
 <$20,000 340 (45.3%) 92 (48.9%) 315 (56.0%)
 $20,000 and above 411 (54.7%) 96 (51.1%) 248 (44.1%)
Cigarettes per Day [3–60] 17.9 (7.5) 17.6 (7.0) 18.0 (7.6) 0.56
FTND Score [0–10] 5.4 (2.0) 5.6 (1.9) 5.3 (2.0) 0.09
CO Boost [−7–17] 5.2 (3.0) 5.4 (3.3) 5.1 (3.0) 0.26
Cigarette Brand 0.50
 American Spirit 23 (3.1%) 7 (3.7%) 16 (2.8%)
 Camel 72 (9.6%) 12 (6.4%) 60 (10.6%)
 Marlboro 450 (59.8%) 116 (61.7%) 334 (59.2%)
 Maverick 36 (4.8%) 9 (4.8%) 27 (4.8%)
 Newport 67 (8.9%) 19 (10.1%) 48 (8.5%)
 Pall Mall 27 (3.6%) 9 (4.8%) 18 (3.2%)
 Other brand 77 (10.2%) 16 (8.5%) 61 (10.8%)
Cigarette Size 0.42
 Regular 454 (61.4%) 109 (58.9%) 345 (62.3%)
 100s 285 (38.6%) 76 (41.1%) 209 (37.7%)
Cigarette Pack Type 1.00
 Hard pack 736 (98.4%) 184 (98.4%) 552 (98.4%)
 Soft pack 12 (1.6%) 3 (1.6%) 9 (1.6%)
Cigarette Flavor or Color 0.10
 Full flavor or Red Color 503 (67.0%) 138 (73.4%) 365 (64.8%)
 Lights or Gold Color 212 (28.2%) 43 (22.9%) 169 (30.0%)
 Ultra Light, Silver, or Other Color 36 (4.8%) 7 (3.7%) 29 (5.2%)
Menthol 0.77
 Menthol 68 (9.0%) 18 (9.6%) 50 (8.9%)
 Non-Menthol 684 (91.0%) 170 (90.4%) 514 (91.1%)
Open in a new tab

Sporadic missing data <5%

Table 2.

Cigarette Ratings Scale Items

Item Label Full Sample (N=752) M (SD) Group 1 (N=188) M (SD) Group 2 (N=564) M (SD)
1. Strength 59.9 (21.8) 59.6 (22.5) 59.9 (21.6)
2. Harshness 45.3 (22.9) 44.9 (22.6) 45.4 (23.0)
3. Heat 29.8 (22.9) 30.9 (23.2) 29.4 (22.8)
4. Draw 64.7 (27.6) 66.2 (26.8) 64.3 (27.9)
5. Taste 60.4 (23.0) 58.6 (22.3) 61.0 (23.1)
6. Satisfaction from smoking 63.6 (25.5) 63.4 (25.6) 63.6 (25.1)
7. Burn rate 57.7 (26.8) 59.0 (27.0) 57.3 (26.7)
8. Mild taste 45.2 (25.3) 43.0 (25.4) 45.9 (25.2)
9. Too mild 67.6 (25.5) 67.2 (26.7) 67.7 (25.1)
10. Harshness of smoke 37.4 (27.5) 37.1 (26.1) 37.5 (28.0)
11. After taste 48.6 (26.1) 46.8 (26.2) 49.2 (26.1)
12. Staleness 70.2 (27.9) 71.2 (26.7) 69.8 (28.3)
13. Strength of smoke 55.9 (23.4) 55.6 (24.8) 55.9 (23.0)
14. Smoke smell 54.6 (24.8) 58.9 (21.9) 60.3 (21.3)
Open in a new tab

3.1. Exploratory Factor Analysis

Table 3 outlines the rotated factor loadings from the exploratory factor analysis. The results indicate a three-factor solution. Factors were labeled in relation to existing literature (Hanson et al., 2009): 1. Product harshness evaluation (3 items: very harsh, very hot, not mild taste), 2. Smoking satisfaction (6 items: very strong, easy draw, smoking satisfaction, not too mild for me, did not seem stale, smoke seemed very strong), and 3. Positive sensory experience (3 items: good taste, good aftertaste, pleasant smell). From the original scale, harshness of smoke and burn rate were removed as their loadings were less than 0.45. Harshness of smoke loaded onto all three constructs and burn rate loaded onto smoking satisfaction all at minimal levels. Analyses allowed for cross-loadings representing real-world correlations that can occur between latent constructs. For instance, strength cross-loaded onto both product harshness evaluation and smoking satisfaction. Yet, strength loaded most strongly onto smoking satisfaction and was at our 0.45 threshold of lower practical significance for product harshness evaluation, so we categorized it under smoking satisfaction.

Table 3.

Exploratory Factor Analysis

Item Label Original Scale Item Number Direction Product Harshness Evaluation Smoking Satisfaction Positive Sensory Experience
1. Harshness 2 Very harsh 0.76 * 0.02 −0.08
2. Heat 3 Very hot 0.50 * 0.11 −0.06
3. Mild taste 8 Not mild taste 0.55 * 0.00 −0.18
4. Strength 1 Very strong 0.45* 0.52 * 0.01
5. Draw 4 Easy draw −0.34 0.53 * −0.01
6. Satisfaction from smoking 6 Smoking satisfaction 0.02 0.53 * 0.23
7. Too mild 9 It was not too mild for me 0.11 0.77 * −0.13
8. Staleness 12 Somehow it did not seem stale −0.16 0.62 * 0.15
9. Strength of smoke 13 Smoke seemed very strong 0.37 0.46 * 0.12
10. Taste 5 Good taste 0.01 0.38* 0.48 *
11. After taste 11 Left a good aftertaste in my mouth −0.05 0.00 0.79 *
12. Smoke smell 14 Pleasant smell 0.01 −0.06 0.75 *
13. Burn rate 7 Did not burn too fast in too few puffs −0.03 0.42* 0.08
14. Harshness of smoke 10 Smoke seemed too harsh 0.43* −0.31* −0.41*
Open in a new tab
*

p <.05. Bolded estimates indicate retained items for each factor.

3.2. Confirmatory Factor Analysis

Product harshness evaluation had a mean of 40.3 (SD = 18.3) and was modestly correlated with smoking satisfaction (r = 0.11, p = .01) and negatively correlated with positive sensory experience (r = −0.18, p <.001). Smoking satisfaction had a mean of 63.5 (SD = 16.8) and was moderately correlated with positive sensory experience (r = 0.40, p <.001). Positive sensory experience had a mean of 54.9 (SD = 19.6).

3.2.1. Reliability

The Cronbach alphas for Day 0, Cigarette 1 were as follows: product harshness evaluation (α = 0.66), smoking satisfaction (α = 0.75), positive sensory experience (α = 0.70). The Supplemental Materials outlines the latent omegas and factor loadings for each construct. Fit indices were all within standard ranges and omega reliability was slightly higher than the Cronbach alpha estimates. The McDonald omega for Day 0, Cigarette 1 were as follows: product harshness evaluation (ω = 0.69), smoking satisfaction (ω = 0.76), positive sensory experience (ω = 0.71). Easy draw had less impact for smoking satisfaction with regard to its loadings across the time points. Yet, conditional independence and measurement invariance testing indicated that cross-time correlations improved model fit such that there was suitable stability of the construct loadings across all three measurements (Supplemental Table 2).

3.2.2. Validity

Tobacco use and product feature outcomes.

Table 4 shows the rating scale factors and their bivariate associations with demographics, tobacco outcomes, and cigarette product features. Those that indicated their usual brand is harsh were typically younger, Asian or White, with higher education, smoked fewer CPDs using full flavor or red color cigarettes (p’s <.05). Those that indicated satisfaction with their usual brand were typically younger, with higher education, and higher average CO boost (p’s <.05). Finally, those that indicate a positive sensory experience typically smoked more cigarettes per day and had higher nicotine dependence scores (p’s <.05).

Table 4.

Bivariate Associations between Participant Characteristics and Cigarette Ratings Subscales

Product harshness evaluation LSM(SE) or r p value Smoking satisfaction LSM(SE) or r p value Positive sensory experience LSM(SE) or r p value
Age −0.11 0.01 −0.08 0.05 0.06 0.17
Race 0.03 0.53 0.24
 American Indian/Alaska Native 32.1 (10.5) 57.3 (9.6) 73.9 (11.2)
 Asian 50.9 (5.2) 56.7 (4.8) 50.0 (5.6)
 Black/African American 37.0 (1.8) 62.9 (1.6) 55.1 (1.9)
 Native Hawaiian/Pacific Islander 36.1 (3.3) 62.9 (3.0) 59.8 (3.6)
 White 41.2 (0.9) 64.2 (0.8) 54.6 (1.0)
 More than one race
Ethnicity 0.23 0.15 0.28
 Hispanic 35.3 (4.2) 58.1 (3.9) 50.1 (4.5)
 Non-Hispanic 40.4 (0.8) 63.8 (0.7) 55.1 (0.8)
Sex 0.63 0.09 0.80
 Female 39.5 (1.5) 65.5 (1.4) 55.2 (1.6)
 Male 40.4 (0.9) 62.8 (0.8) 54.7 (1.0)
Education 0.01 0.01 0.39
 High school graduate/GED or less 37.3 (1.3) 60.9 (1.4) 55.9 (1.5)
 Some college or more 41.8 (1.0) 65.0 (0.8) 54.4 (1.0)
Income 0.19 0.77 0.71
 <$20,000 39.2 (1.2) 62.3 (1.1) 55.3 (1.3)
 $20,000+ 41.2 (1.0) 63.7 (0.9) 54.7 (1.1)
Cigarettes per Day −0.11 0.01 0.02 0.61 0.09 0.04
FTND Score −0.08 0.06 0.02 0.62 0.10 0.01
CO Average Boost 0.04 0.35 0.13 0.002 0.04 0.31
Cigarette Brand 0.99 0.24 0.49
 American Spirit 42.0 (4.6) 61.0 (4.2) 52.1 (4.9)
 Camel 40.2 (2.4) 64.3 (2.2) 52.9 (2.6)
 Marlboro 40.3 (1.0) 63.9 (0.9) 56.1 (1.1)
 Maverick 41.4 (3.5) 62.4 (3.2) 58.2 (3.8)
 Newport 40.7 (2.7) 67.9 (2.4) 53.4 (2.8)
 Pall Mall 41.0 (4.3) 58.8 (3.9) 49.8 (4.6)
 Other brand 38.6 (2.4) 60.1 (2.1) 52.3 (2.5)
Cigarette Size 0.39 0.12 0.38
 Regular 40.8 (1.0) 62.7 (0.9) 54.3 (1.1)
 100s 39.5 (1.3) 65.0 (1.2) 55.8 (1.4)
Cigarette Pack Type 0.78 0.32 0.26
 Hard pack 40.3 (0.8) 63.5 (0.7) 55.1 (0.8)
 Soft pack 42.0 (6.1) 69.2 (5.6) 47.7 (6.5)
Cigarette Flavor/Color 0.001 0.14 0.59
 Full flavor or Red Color 42.3 (0.9) 63.9 (0.9) 55.5 (1.0)
 Lights or Gold Color 35.7 (1.4) 62.0 (1.3) 53.7 (1.5)
 Ultra Light, Silver, or Other Color 41.7 (3.4) 68.2 (3.1) 54.0 (3.7)
Menthol 0.77 0.47 0.94
 Menthol 39.5 (2.6) 65.2 (2.4) 54.7 (2.8)
 Non-Menthol 40.3 (0.8) 63.4 (0.7) 54.9 (0.9)
Open in a new tab

Multivariable general linear models indicated that every unit of cigarettes per day was associated with a lower harshness rating b = −0.29 (95% CI: −0.51, −0.07) and higher positive sensory experience b = 0.32 (95% CI: 0.08, 0.56). FTND average dependence score was associated with a higher positive sensory experience b = 1.08 (95% CI: 0.28, 1.89). Additionally, CO boost was associated with smoking satisfaction b = 0.77 (95% CI: 0.30, 1.26).

4.0. Discussion

Using data from over 700 adult daily cigarette smokers who participated in five clinical trials with identical baseline procedures, the current study aimed to assess the construct validity of the Cigarette Ratings Scale. Our exploratory factor analysis of the Cigarette Ratings Scale extracted three factors: product harshness evaluation, smoking satisfaction, and positive sensory experience. These factors are complementary to the existing literature and align with existing scale constructs (Cappelleri et al., 2007; Hanson et al., 2009). Yet, the current study extends understanding of these constructs in several ways. First, scales often broadly collect strength, satisfaction, and taste-related ratings, yet they are often vague in nature by asking about such terms alone without greater specification (Cappelleri et al., 2007; Karelitz & Perkins, 2021). This study helps to further operationalize these constructs and distinguish them from other concepts within their nomological network. For example, we conceptualize strength along a spectrum to capture the range of experience, including harshness, heat, and taste intensity. Smoking satisfaction also typically asks about smoking enjoyment or satisfaction only, though it is often unclear what aspects are satisfying or enjoyable. Our assessment of the Cigarette Ratings Scale indicates that satisfaction is complex, drawing upon several aspects that seem to vary “in the moment.” Results showed strength, albeit not a harsh strength, is an aspect of satisfaction along with other momentary assessment features (e.g., easy draw, not stale). Furthermore, our assessment of taste is not solely about good taste or conflated with satisfaction in this context but is operationalized as part of a comprehensive sensory experience including aftertaste and pleasant smell.

Secondly, several subjective rating scales often use items individually and have shown limited effects when assessed independently in previous work (Mercincavage et al., 2016, 2018, 2020; Strasser et al., 2013). We extend this work by creating subscales with several items that we validated as latent constructs shown to be consistent across a five day period with three data collection points. This indicates rigorous testing of the scale, though future work should further explore other temporal associations with the factors in relation to tobacco use and product feature outcomes. Additionally, an important distinction of the current study is that the scale was used to assess current smoker’s ratings of their usual cigarette brand on its own, not compared to other products (Brandon et al., 2019). This serves as a basis for cigarette ratings that participants were drawn to and use regularly in a natural setting and currently use several times a day. Future work could replicate these findings, assess absolute ratings for other addictive substances, and compare them relative to each other as well.

A confirmatory factor analysis of the Cigarette Ratings Scale showed sufficient face validity in that product harshness was modestly associated with smoking satisfaction because strength was an aspect of both. Furthermore, there was sufficient discriminant validity such that a positive sensory experience was associated with being satisfied, but not with the harshness of the product. The subjective rating subscales were associated with unique tobacco use and product feature outcomes as well. Greater daily cigarette consumption was associated with lower harshness ratings and more positive sensory experience. Higher nicotine dependence scores were associated with more positive sensory experience. Greater CO boost was primarily associated with greater smoking satisfaction as well.

Our findings indicate that self-reported smoking history measures such as cigarettes per day and nicotine dependence scores were associated with global product evaluations of harshness and a positive sensory experience. Yet, smoking satisfaction was the only subjective rating subscale associated with an objective measure of cigarette use. This could indicate temporal or contextual differences with the use of this subscale. Though cigarette smokers may continue smoking because they are generally satisfied with a product, the specific smoking satisfaction items measured here (i.e., strength, draw, staleness) captured aspects of the product use experience associated with greater carbon monoxide intake. Notably, as this subscale was not associated with daily cigarette consumption or dependence scores, our findings suggest the potential for intrapersonal and interpersonal factors that may modify one’s satisfaction in the moment (e.g., mood, emotional response, peer presence) that are distinct from global product ratings (Piasecki et al., 2014).

There are limitations to the current study. Namely, the clinical trials providing data for these analyses recruited daily cigarette smokers from a specific US regional area. Moving forward, this scale should be tested for external validity with other populations and products to compare ratings with associated tobacco use and their product features. This could include further assessment of subgroup differences (e.g., age, race, education) that were found. The Cronbach’s alpha for the harshness subscale was lower compared to the other subscales and easy draw had less impact for smoking satisfaction with regard to its loadings across the time points. However, Cronbach alpha values can be sensitive to the number of items, the McDonald Omegas was slightly higher for all factors, and measurement invariance testing indicated that there was suitable stability of the construct loadings across all three measurements. Lastly, associations in the multivariable modeling indicated there were several statistically significant relationships, though some estimates were modest which may limit their practical significance. Further work should study this scale compared to other product evaluation scales and common theoretical concepts (e.g., emotional response, self-efficacy, risk perceptions) as well as other product features (e.g., ventilation) (Pennings et al., 2021) to determine their relative importance.

Our analysis of the construct validity for the Cigarette Ratings Scale provides both theoretical and practical contributions. Psychometric evaluation provided a greater understanding of subjective ratings’ construct validity for product harshness evaluation, smoking satisfaction, and positive sensory experience for current smokers’ usual cigarette brand. Furthermore, subscales were associated with both subjective and objective measures of tobacco use, though minimally with product marketing features. Our findings have direct application to future research, pointing to opportunities to pare down items to the subscales that uniquely relate to established behavioral outcomes that could be further studied to evaluate other current and emerging tobacco products. This extends previous work by identifying a concise approach to studying the associations between subjective ratings and daily cigarette use, biological exposure (carbon monoxide boost), and nicotine dependence. These are important contributions for researchers and practitioners working in addiction efforts. Tobacco regulatory science can also leverage this scale to triangulate product experience and regulatory-relevant outcomes. Use of the Cigarette Ratings Scale may allow for direct comparisons to industry research and support regulatory efforts to reduce cigarette’s disproportional burden on public health.

Supplementary Material

1

Highlights.

  • Subjective ratings inconsistently associated with outcomes and no current standard.

  • Study investigated the construct validity of the Cigarette Ratings Scale.

  • Subscales primarily associated with behavior and nicotine dependence.

Funding

Research reported in this publication was supported by the National Cancer Institute (NCI) of the National Institutes of Health (NIH) and the U.S. Food and Drug Administration (FDA) Center for Tobacco Products under Award Number U54CA229973 and by NCI Award Number K07218366. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the FDA.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflict of Interest

The authors have no conflicts declared.

REFERENCES

  1. Alam MM, Ward KD, Bahelah R, Kalan ME, Asfar T, Eissenberg T, & Maziak W (2020). The Syrian Center for Tobacco Studies-13 (SCTS-13): Psychometric evaluation of a waterpipe-specific nicotine dependence instrument. Drug and Alcohol Dependence, 215, 108192. 10.1016/j.drugalcdep.2020.108192 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Brandon KO, Simmons VN, Meltzer LR, Drobes DJ, Martínez Ú, Sutton SK, Palmer AM, Bullen CR, Harrell PT, & Brandon TH (2019). Vaping characteristics and expectancies are associated with smoking cessation propensity among dual users of combustible and electronic cigarettes. Addiction, 114(5), 896–906. 10.1111/add.14551 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Brown TA (2015). Confirmatory factor analysis for applied research (2nd ed). Guilford Press. [Google Scholar]
  4. Burling TA, Stitzer ML, Bigelow GE, & Mead AM (1985). Smoking topography and carbon monoxide levels in smokers. Addictive Behaviors, 10(3), 319–323. 10.1016/0306-4603(85)90014-0 [DOI] [PubMed] [Google Scholar]
  5. Cappelleri JC, Bushmakin AG, Baker CL, Merikle E, Olufade AO, & Gilbert DG (2007). Confirmatory factor analyses and reliability of the modified cigarette evaluation questionnaire. Addictive Behaviors, 32(5), 912–923. 10.1016/j.addbeh.2006.06.028 [DOI] [PubMed] [Google Scholar]
  6. Cassidy RN, Tidey JW, Cao Q, Colby SM, McClernon FJ, Koopmeiners JS, Hatsukami D, & Donny EC (2019). Age Moderates Smokers’ Subjective Response to Very-Low Nicotine Content Cigarettes: Evidence from a Randomized Controlled Trial. Nicotine & Tobacco Research: Official Journal of the Society for Research on Nicotine and Tobacco, 21(7), 962–969. 10.1093/ntr/nty079 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Census Bureau. (2019). American Community Survey. Demographic and Housing Estimates. https://data.census.gov/cedsci/table?q=United%20States&tid=ACSDP1Y2019.DP05 [Google Scholar]
  8. Centers for Disease Control and Prevention. (2021). Tobacco Brand Preferences. Smoking & Tobacco Use. https://www.cdc.gov/tobacco/data_statistics/fact_sheets/tobacco_industry/brand_preference/index.htm [Google Scholar]
  9. Flay BR (1999). Understanding environmental, situational and intrapersonal risk and protective factors for youth tobacco use: The Theory of Triadic Influence. Nicotine & Tobacco Research: Official Journal of the Society for Research on Nicotine and Tobacco, 1 Suppl 2, S111–114; discussion 569–570. 10.1080/14622299050011911 [DOI] [PubMed] [Google Scholar]
  10. Geiser C (2013). Data analysis with Mplus. Guilford Press. [Google Scholar]
  11. Government Printing Office. (2009). Family Smoking Prevention and Tobacco Control and Federal Retirement Reform. 111th Congress Public Law 111–31. https://www.govinfo.gov/content/pkg/PLAW-111publ31/pdf/PLAW-111publ31.pdf
  12. Hair J, Black W, Babin B, & Anderson R (2009). Multivariate Data Analysis (7th ed.). Pearson. [Google Scholar]
  13. Hanson K, O’Connor R, & Hatsukami D (2009). Measures for assessing subjective effects of potential reduced-exposure products. Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, Cosponsored by the American Society of Preventive Oncology, 18(12), 3209–3224. 10.1158/1055-9965.EPI-09-0971 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hatsukami DK, Zhang Y, O’Connor RJ, & Severson HH (2013). Subjective responses to oral tobacco products: Scale validation. Nicotine & Tobacco Research: Official Journal of the Society for Research on Nicotine and Tobacco, 15(7), 1259–1264. 10.1093/ntr/nts265 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hayes AF, & Coutts JJ (2020). Use Omega Rather than Cronbach’s Alpha for Estimating Reliability. But…. Communication Methods and Measures, 14(1), 1–24. 10.1080/19312458.2020.1718629 [DOI] [Google Scholar]
  16. Heatherton TF, Kozlowski LT, Frecker RC, & Fagerström KO (1991). The Fagerström Test for Nicotine Dependence: A revision of the Fagerström Tolerance Questionnaire. British Journal of Addiction, 86(9), 1119–1127. 10.1111/j.1360-0443.1991.tb01879.x [DOI] [PubMed] [Google Scholar]
  17. Karelitz JL, & Perkins KA (2021). Acute subjective sensory perceptions predict relative reinforcing effects of smoked nicotine. Addictive Behaviors, 117, 106835. 10.1016/j.addbeh.2021.106835 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kim J, Kambari Y, Taggar A, Quilty LC, Selby P, Caravaggio F, Ueno F, Torres E, Song J, Pollock BG, Graff-Guerrero A, & Gerretsen P (2021). A measure of subjective substance use disorder awareness—Substance Use Awareness and Insight Scale (SAS). Drug and Alcohol Dependence, 231, 109129. 10.1016/j.drugalcdep.2021.109129 [DOI] [PubMed] [Google Scholar]
  19. Mercincavage M, Albelda B, Mays D, Souprountchouk V, Giovenco DP, Audrain-McGovern J, & Strasser AA (2020). Shedding “light” on cigarette pack design: Colour differences in product perceptions, use and exposure following the US descriptor ban. Tobacco Control, tobaccocontrol-2020–055886. 10.1136/tobaccocontrol-2020-055886 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Mercincavage M, Lochbuehler K, Wileyto EP, Benowitz NL, Tyndale RF, Lerman C, & Strasser AA (2018). Association of Reduced Nicotine Content Cigarettes With Smoking Behaviors and Biomarkers of Exposure Among Slow and Fast Nicotine Metabolizers: A Nonrandomized Clinical Trial. JAMA Network Open, 1(4), e181346. 10.1001/jamanetworkopen.2018.1346 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Mercincavage M, Saddleson ML, Gup E, Halstead A, Mays D, & Strasser AA (2017). Reduced nicotine content cigarette advertising: How false beliefs and subjective ratings affect smoking behavior. Drug and Alcohol Dependence, 173, 99–106. 10.1016/j.drugalcdep.2016.12.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Mercincavage M, Souprountchouk V, Tang KZ, Dumont RL, Wileyto EP, Carmella SG, Hecht SS, & Strasser AA (2016). A Randomized Controlled Trial of Progressively Reduced Nicotine Content Cigarettes on Smoking Behaviors, Biomarkers of Exposure, and Subjective Ratings. Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, Cosponsored by the American Society of Preventive Oncology, 25(7), 1125–1133. 10.1158/1055-9965.EPI-15-1088 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Mercincavage M, Wileyto EP, Saddleson ML, Lochbuehler K, Donny EC, & Strasser AA (2017). Attrition during a randomized controlled trial of reduced nicotine content cigarettes as a proxy for understanding acceptability of nicotine product standards. Addiction, 112(6), 1095–1103. 10.1111/add.13766 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. National Cancer Institute. (n.d.). The Role of the Media in Promoting and Reducing Tobacco Use. Tobacco Control Monograph No. 19. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute. NIH Pub. No. 07–6242, June 2008
  25. O’Connor R, Watson CH, Swan GE, Nettles DS, Geisler RC, Hendershot TP, & PhenX TRR Agent Working Group. (2020). PhenX: Agent measures for tobacco regulatory research. Tobacco Control, 29(Suppl 1), s20–s26. 10.1136/tobaccocontrol-2019-054976 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Pennings JLA, Ferris Wayne G, Klerx WNM, Pauwels CGGM, & Talhout R (2021). Variations in cigarette brand characteristics: Can consumers tell the difference? Tobacco Control, tobaccocontrol-2021–056856. 10.1136/tobaccocontrol-2021-056856 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Phillip Morris International. (n.d.). Cigarette ratings questionnaire. Bates No. 2022259048 www.pmdocs.com
  28. Piasecki TM, Trela CJ, Hedeker D, & Mermelstein RJ (2014). Smoking antecedents: Separating between- and within-person effects of tobacco dependence in a multiwave ecological momentary assessment investigation of adolescent smoking. Nicotine & Tobacco Research: Official Journal of the Society for Research on Nicotine and Tobacco, 16 Suppl 2, S119–126. 10.1093/ntr/ntt132 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Piper ME, Brown DC, Hendershot TP, Swan GE, & PhenX TRR Host: Social/Cognitive Working Group. (2020). PhenX: Host: Social/Cognitive measures for tobacco regulatory research. Tobacco Control, 29(Suppl 1), s5–s12. 10.1136/tobaccocontrol-2018-054467 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Smith TT, Donny EC, Luo X, Allen AM, Carroll DM, Denlinger-Apte RL, Dermody SS, Koopmeiners JS, McClernon FJ, Pacek LR, Vandrey R, & Hatsukami DK (2019). The Impact of Gradual and Immediate Nicotine Reduction on Subjective Cigarette Ratings. Nicotine & Tobacco Research: Official Journal of the Society for Research on Nicotine and Tobacco, 21(Suppl 1), S73–S80. 10.1093/ntr/ntz158 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Strasser AA, Ashare RL, Kaufman M, Tang KZ, Mesaros AC, & Blair IA (2013). The effect of menthol on cigarette smoking behaviors, biomarkers and subjective responses. Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, Cosponsored by the American Society of Preventive Oncology, 22(3), 382–389. 10.1158/1055-9965.EPI-12-1097 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Strasser AA, Ashare RL, Kozlowski LT, & Pickworth WB (2005). The effect of filter vent blocking and smoking topography on carbon monoxide levels in smokers. Pharmacology, Biochemistry, and Behavior, 82(2), 320–329. 10.1016/j.pbb.2005.09.001 [DOI] [PubMed] [Google Scholar]
  33. Strasser AA, Lerman C, Sanborn PM, Pickworth WB, & Feldman EA (2007). New lower nicotine cigarettes can produce compensatory smoking and increased carbon monoxide exposure. Drug and Alcohol Dependence, 86(2–3), 294–300. 10.1016/j.drugalcdep.2006.06.017 [DOI] [PubMed] [Google Scholar]
  34. Department of Health U.S. and Human Services. (2014). The Health Consequences of Smoking—50 Years of Progress: A Report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1788149-supplement-1.docx (17.4KB, docx)

RESOURCES