Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Oct 31.
Published in final edited form as: Subst Use Misuse. 2023 Oct 31;58(14):1883–1894. doi: 10.1080/10826084.2023.2257312

Modeling Idiographic Longitudinal Relationships between Affect and Cigarette Use: An Ecological Momentary Assessment Study

Joseph A Spillane 1, Peter Soyster 1,*
PMCID: PMC10872632  NIHMSID: NIHMS1930841  PMID: 37735802

Abstract

Despite public knowledge of the adverse health effects of tobacco use, cigarettes remain widely used due to the addictive nature of nicotine. Physiologic adaptation to the presence of nicotine over time leads to unpleasant effects during withdrawal periods. Alongside these physiological effects, tobacco users often report changes in their consumption of tobacco in response to their emotional state. We hypothesized that idiographic, or person-specific level, increases in participants’ negative affect (NA) and positive affect (PA) ratings at a given time point would be associated with higher and lower craving and smoking over the following several hours, respectively. Fifty-two participants completed block randomized ecological momentary assessment surveys on their smartphones 4 times per day for 30 days, reporting from 0–100 their level of seven discrete emotions, stress, current craving, and smoking behavior. We analyzed the relationships between affect and smoking and craving using idiographic generalized linear models. While some participants exhibited the hypothesized relationships, each participant varied in the strength and direction of the relationships between affect and craving/smoking. These outcomes were partially moderated at the group level by anxiety/depression at baseline, but not by level of nicotine dependence or sex. This suggests that the factors driving cigarette use vary significantly between individuals.

Keywords: tobacco, craving, affect, idiographic, ecological momentary assessment

Introduction

For over five decades, the negative health impacts of cigarette use have been a focus of public health. However, cigarettes and other nicotine-containing products remain widely available and used by millions (Hu, 2016; US Dept. of Health and Human Services [USDHHS], 2020). In 2018, 13.7% of the US population, about 34.2 million people, regularly smoked cigarettes despite clear links to higher rates of mortality via cancers, respiratory disorders, and cardiovascular disorders (Hu, 2016; USDHHS, 2020; Wang et al., 2018). In the United States alone, tobacco use is the leading preventable cause of death and results in over $170 billion in healthcare costs and productivity loss (USDHHS, 2020). Despite public awareness of these negative health effects, many people find it difficult to quit smoking.

Nicotine, the primary psychoactive ingredient in tobacco, is highly addictive and a key determinant of the continuation of tobacco use (Benowitz, 2010; De Biasi & Dani, 2011). Physiologically, the acute effects of nicotine are generally reported as pleasant and include a head rush or ‘buzz’, muscle relaxation, and improved focus (Benowitz, 2009). Social factors such as encouragement by peers or family members have also been shown to reinforce tobacco use (Christakis & Fowler, 2008). The physiological changes associated with nicotine ingestion can contribute to an acute increase in positive affect (PA) following use (Watkins, Koob, & Athina, 2000). Use routinely evokes a hedonically positive physiologic and affective state, which can lead to repeated and increased use (Watkins, Koob, & Athina, 2000). Such positive reinforcement often leads a person to want to use more of the substance, but higher frequency of use and increased amount of nicotine ingested creates an allostatic burden on the user’s endemic nicotinergic system leading to physiological insensitivity, or tolerance, to nicotine (George & Koob, 2017). As tolerance grows, one must ingest more nicotine to experience the same effect and will experience stronger withdrawal after cessation (Koob, 2006; McLaughlin, Dani, & DeBiasi, 2015).

Over time, positive reinforcement decreases through opponent processes -- intrinsic anti-reward systems that counteract the brain’s sensitivity to positive stimuli -- producing an aversive withdrawal state which can be alleviated through nicotine administration (George & Koob, 2017; Koob 2015). Without the presence of nicotine, symptoms of withdrawal typically begin within 4 to 24 hours after cessation (McLaughlin, Dani, & DeBiasi, 2015). Tobacco withdrawal symptoms include irritability; frustration or anger; anxiety; difficulty concentrating; increased appetite; restlessness; depressed mood; and insomnia (Diagnostic and Statistical Manual of Mental Disorders, 5th Edition; American Psychiatric Association, 2013). This suggests that those who use tobacco are likely to feel increased negative affect (NA) and decreased positive affect (PA) during periods of withdrawal, which is alleviated by smoking tobacco again (Baker, Piper, McCarthy, Majeskie, & Fiore, 2004; Koob, 2015; Piper et al., 2008). Though negative reinforcement is not often thought to be the impetus of drug use, using nicotine takes away the unpleasant effects of nicotine withdrawal, including the experience of negative affect, such that chronic smoking is thought to be primarily maintained by escape or avoidance of withdrawal state (Baker, et al., 2004; George & Koob, 2017; Koob & Le Moal, 1997; Piper et al., 2008).

Craving, the conscious desire to smoke a cigarette, has been found to contribute to smoking relapse, and tends to be higher in tobacco users who are more dependent on nicotine and those with higher NA (Carter, et al., 2008; Robinson, et al., 2011; Swan, Ward, & Jack, 1996; for a review, see: Wray, Gass, & Tiffany, 2013). However, the nature of the relationship between craving and smoking has been shown to depend on the analyzed timescale, as moment-to-moment increases in craving predicts increased smoking, and reciprocally, smoking predicts decreased craving (Chandra, Scharf, & Shiffman, 2011; Motschman, Germeroth, & Tiffany, 2018).

Idiographic Modeling of Tobacco Use

While the above mentioned nomothetic, or group-level, analyses of smoking behaviors have yielded important information about the average experiences of tobacco users in general, these findings are not necessarily generalizable to psychological processes occurring within any single person at a given time (Molenaar, 2004; Fisher et al., 2018; Soyster & Fisher, 2019). Put another way, individuals often deviate from what is “generally true”, but most between-subjects research is unable to determine if and to what degree a specific individual may differ from the model-estimated average. An idiographic, or person-specific, approach attempts to overcome this issue by quantifying how an individual’s behaviors and subjective experiences occur in real-world situations and how they change over time (Molenaar, 2004; Fisher, 2015; Piccirillo & Rodebaugh, 2019). Idiographic studies repeatedly sample individuals over time, and analyze each participant’s data independently (i.e., without combining data between participants) to identify each individual’s potentially unique relationships between variables of interest. After idiographic analyses have been completed, it is possible to compare results across subjects to identify similarities and differences in results that may be attributable to demographic or other factors. Since tobacco use is a behavior that occurs at the individual level, recognizing the heterogeneity of smoking experiences may ultimately allow for personalized conceptualizations of use and improved cessation treatment outcomes.

Studying Affect and Tobacco Use

The reasons for smoking likely vary significantly at the individual level, and theories of smoking behavior have long posited that affect is closely linked to smoking behavior on a day-to-day basis and that these relationships vary from person to person (e.g., Ikard, Green, & Horn, 1969; Ikard & Tomkins, 1973; Leventhal et al., 2010; Shiffman, 2005). Affective states have also been shown to vary substantially over time, but group-aggregate analyses (where multiple individuals are measured cross-sectionally) do not reflect this time-dependent variability as it occurs within a specific individual (Verduyn et al., 2009).

In light of this, the present study took an idiographic approach to quantify the extent to which variation in affect over time within individuals accounted for the variance in those individuals’ smoking and craving. We sought to understand the extent to which an individuals’ moment-to-moment changes in NA and PA were related to their smoking and tobacco craving over the subsequent several hours.

One method for idiographic sampling is ecological momentary assessment (EMA), which utilizes pen and paper, mobile computers, or smartphones to allow participants to answer questions in non-laboratory settings over many time points. Despite robust evidence for neurobiological adaptation in addiction processes, indicating negative reinforcement from a biological perspective, previous EMA studies with varying sampling strategies and time intervals that measured self-reported affect have found inconsistent relationships between NA and increased smoking (e.g.: Carter et al., 2008; Shiffman, et al., 2004; Chandra, Scharf, & Shiffman, 2011). This complicates findings from global questionnaire data that self-reported NA motivates smoking at the group level and indicates a need for further research on the relationships between emotions and smoking (Shiffman et al., 2009). Lower PA has also been found to predict worse long-term cessation outcomes, and some studies have even found that lower PA may be a stronger motivation for smoking during withdrawal than higher NA (Cook, Spring, McChargue, & Hedeker, 2004; Doran, et al., 2006).

The Present Study

The present study used EMA sampling to examine whether individual, self-reported NA and PA at a given time point (t) could reliably predict whether an individual smoked a cigarette during the subsequent several hours (t+1). We also examined the relationships between NA, PA, and tobacco craving.

Based on finding from previous between-subjects research, we investigated four hypotheses regarding the individual relationships between affect and smoking behavior and craving: (1) that NA at a time point (t) would be positively associated with smoking between times t and t+1; (2) that NA at a time point (t) would be positively associated with craving between times t and t+1, after controlling for smoking at time t; (3) that PA at a time point (t) would be negatively associated with smoking between times t and t+1; and (4) that PA at a time point (t) would be negatively associated with craving between times t and t+1, after controlling for smoking at time t. We expected these hypotheses to hold for the majority of participants surveyed.

At the between-subjects level, tobacco users who are more nicotine dependent often report higher NA while abstaining from smoking when compared to less dependent users, and increases in NA have been shown to predict tobacco use in the following hours (Robinson, et al., 2011; Shiffman, 2005). As one develops increased tolerance to nicotine, withdrawal effects caused by nicotine deprivation worsen, meaning NA is higher during withdrawal states (George & Koob, 2017). Thus, we were interested in investigating whether there was evidence for group-level moderation of individual model results. Although the present study’s between-subject sample size was relatively small, we sought to describe moderating effects of tobacco use and demographic variables on the idiographic model results. We hypothesized that there would be a group-level moderating effect of the experiences of NA at baseline (measured by the depression and anxiety subscales of the Depression, Anxiety, and Stress Scale [DASS; Lovibond & Lovibond, 1995]), such that the relationships between NA and smoking would be more positive in general for participants with greater baseline symptoms of NA.

Methods

This study used self-report data collected as part of a parent study (see Soyster & Fisher, 2019), which sought to model individualized predictors of smoking behavior. We analyzed the data independently for the present hypotheses.

Participants

The final analyses included a total of 52 adult participants (Mean Age = 25.63, SD = 11.22, range = 19 – 62) who lived in the San Francisco Bay Area and self-identified as current tobacco users. Participants reported a median of 4 years smoking (Mean 8.15, SD = 10.15, range = 1 – 48). Among participants, 36.5% (N=19) identified as female, 61.5% (N=32) identified as male, and 1.9% (N=1) responded “prefer to self describe”. Thirty eight percent (N=19) of participants identified as Asian, 33% (N=17) White, 2% (N=1) Black, 2% (N=1) as Native Hawaiian or Pacific Islander, 23% (N=12) multiracial or ‘other’, with 4% who preferred not to disclose. The ‘other’ category had an optional free response and included 6 responses of “Middle Eastern”, 1 response of “Persian”, and 1 response of “Biracial, Black and Latina.”

Procedure

All study procedures were approved by the University of California, Berkeley Committee for the Protection of Human Subjects.

Recruitment.

Participants were recruited via an undergraduate research participation pool, flyers posted in the community, as well as digital advertisements on Craigslist. In order to participate, individuals were required to be fluent in English, be 18 years or older, have regular access to a smartphone capable of receiving text-messages and accessing the internet, report having smoked at least 100 cigarettes in their lifetime, and report smoking at least 1 cigarette per week at baseline. Eligible individuals were invited to the laboratory where, after providing informed consent, they completed a baseline survey and received instructions for completing the EMA surveys.

Measures

DASS.

Administered at baseline, the DASS contains 42 items designed to assess the related negative emotional states of depression, anxiety, and stress. The four subscales corresponding to each negative emotional state each consist of 14 items each rated on a 4-point Likert scale ranging from 0–3 with anchors “did not apply to me at all” and “applied to me very much or most of the time.” For each subscale, we calculated scores by summing across the 14 items.

EMA items.

After completing the baseline questionnaires, lab staff collected each participant’s cell phone number and entered it into a secure web-based survey system. The system pushed surveys to participants’ phones at block-randomized intervals. Participants received surveys on their phones four times per day, randomly during four-hour windows set to start at their reported wake-time (e.g., for a reported wake time of 8am, a survey would come randomly between 8am - 11:59am; 12pm - 3:59pm; 4pm - 7:59pm; and 8pm - 11:59pm) for 30 days. Participants were required to complete at least 80% of the surveys over the length of the assessment period in order to receive study reimbursement. EMA surveys assessed the number of cigarettes consumed since the previous survey, as well as current (i.e., “right now”) craving, smoking cues, social context, nicotine withdrawal symptoms, positive emotions, negative emotions, and an additional 3 questions for the first survey each day assessing the previous night’s sleep. At each observation, participants rated their current experience of each item on a 0–100 visual analog slider, anchored at 0 with not at all and at 100 with as much as possible. Based on focus-group participant input in previous studies, Soyster and Fisher (2019) developed these survey items to mirror the language used by current tobacco users and found in academic literature for each construct (Soyster & Fisher, 2019). The final questions included in the survey are attached in Appendix 1.

We operationalized NA as a latent factor, indicated by the ratings of four discrete negative emotions: angry, down/depressed, bored, and stress at each time point. We similarly operationalized PA, indicated by ratings of happy, enthusiastic, and calm/relaxed. To identify the NA and PA factor loadings for each participant, we conducted an exploratory factor analysis (EFA) using the Psych package in R (Revelle, 2022). To assess model fit, we then conducted a confirmatory factor analysis (CFA) using the Lavaan package in R (Rosseel, 2012). Both the NA and PA factor models were determined to have good fit as indicated by the between-participant average comparative fit index (CFI; CFINA = 0.95; CFIPA = 0.99), the root mean square error of approximation (RMSEA; RMSEANA = 0.04; RMSEAPA = 0.02), and the standard root mean square residual (SRMR; SRMRNA = 0.07; SRMR; SRMRPA = 0.04; Hu and Bentler, 1999). See Table 1 for person-specific factor model fit statistics. From these models, we generated NA and PA factor scores for each timepoint for each individual. Smoking was analyzed as binary (smoking or non-smoking) and craving was analyzed continuously from 0–100 for each time point, with higher scores indicating a greater degree of craving.

Table 1:

RMSEA, CFI, and SRMR outputs from CFA for NA and PA scales for each participant

ID RMSEANA CFINA SRMRNA RMSEAPA CFIPA SRMRPA
3 0.106 0.750 0.090 0 1 0.047
4 0.096 0.851 0.076 0 1 0.023
5 0.056 0.966 0.067 0.057 0.983 0.074
6 0.000 1.000 0.080 0 1 0.036
7 0.104 0.812 0.079 0 1 0.014
8 0.000 1.000 0.054 0.247 0.686 0.126
9 0.000 1.000 0.040 0 1 0.025
10 0.058 0.967 0.060 0 1 0.029
11 0.000 1.000 0.032 0 1 0.002
12 0.030 0.978 0.063 0 1 0.029
13 0.000 1.000 0.053 0 1 0.021
14 0.000 1.000 0.058 0.087 0.903 0.087
15 0.000 1.000 0.053 0 1 0.008
16 0.055 0.960 0.104 0 1 0.039
17 0.000 1.000 0.035 0 1 0.039
18 0.090 0.907 0.074 0 1 0.025
19 0.000 1.000 0.029 0 1 0.022
20 0.045 0.980 0.067 0 1 0.029
21 0.019 0.987 0.073 0.069 0.933 0.065
22 0.000 1.000 0.044 0 1 0.034
23 0.000 1.000 0.076 0 1 0.057
24 0.000 1.000 0.041 0 1 0.005
25 0.064 0.941 0.074 0 1 0.021
26 0.000 1.000 0.070 0 1 0.019
27 0.000 1.000 0.049 0 1 0.02
28 0.000 1.000 0.066 0 1 0.005
29 0.000 1.000 0.038 0.03 0.998 0.063
30 0.000 1.000 0.072 0 1 0.019
31 0.000 1.000 0.024 0 1 0.019
32 0.049 0.884 0.082 0.015 0.999 0.078
33 0.104 0.755 0.090 0 1 0.044
34 0.000 1.000 0.050 0 1 0.036
35 0.000 1.000 0.041 0 1 0.008
36 0.074 0.913 0.076 0 1 0.024
37 0.084 0.895 0.075 0.115 0.954 0.083
38 0.000 1.000 0.033 0 1 0.054
40 0.000 1.000 0.065 0 1 0.032
41 0.069 0.956 0.076 0.163 0.906 0.112
42 0.000 1.000 0.088 0 1 0.049
44 0.000 1.000 0.049 0.069 0.982 0.104
45 0.000 1.000 0.051 0 1 0.025
46 0.307 0.083 0.200 0 1 0.014
47 0.000 1.000 0.044 0 1 0.017
48 0.056 0.948 0.077 0 1 0.023
49 0.041 0.991 0.071 0 1 0.014
50 0.142 0.799 0.121 0 1 0.05
51 0.070 0.965 0.076 0 1 0.013
52 0.000 1.000 0.053 0 1 0.01
53 0.000 1.000 0.038 0 1 0.039
55 0.055 0.966 0.060 0 1 0.029
56 0.000 1.000 0.047 0 1 0.017
57 0.054 0.961 0.090 0 1 0.032
Average 0.035 0.946 0.065 0.016 0.987 0.037
Open in a new tab

Data Preparation and Analysis

All analyses were conducted using R version 3.5.3 (R Core Team, 2019). Hypotheses 1–4 were all tested in fully idiographic models (i.e., each participant’s data were modeled separately). After completing these analyses, idiographic results were aggregated into a between-subjects dataset, which we used to investigate our between-subjects hypotheses.

To account for autoregressive trends in affect, we regressed each observation of NA and PA at t+1 onto their respective observations at t and then extracted the residuals. The residuals were then used as the NA and PA variables for the present analyses. In order to determine the time-lagged relationships among NA, PA, smoking, and craving over time, lagged versions (i.e., the variable’s value at t-1) of the smoking and craving variables were created and included as independent variables to account for possible autoregressive effects in the dependent variables.

Hypotheses 1 and 2: NA and Smoking/Craving.

To test whether participants’ self-reported NA at a given time point was positively correlated with self-reported cigarette smoking at the following time point, we ran a generalized linear model for each individual predicting smoking at t+1 from smoking and the residuals of the NA autoregression at time t. To test whether NA was positively related to self-reported craving, we ran a similar model predicting craving at time t+1 from craving, the residuals of the NA autoregression, and smoking at t. We included lagged smoking to control for any effect that recent smoking may have had on craving.

Hypothesis 3 and 4: PA and Smoking/Craving.

To test whether self-reported PA at a given time point was negatively correlated with self-reported smoking or craving a cigarette at the next time point, we ran similar models as above but with PA. First, we predicted smoking at time t+1 from smoking and the PA autoregression residuals at t. Then, we predicted craving at time t+1 from craving, the residuals of the PA autoregression, and smoking at t.

Between-Subject Hypotheses.

Aggregating the coefficients from the idiographic analyses described above as outcomes, we ran a series of linear regression models to assess whether there were any moderating variables which systematically predicted differences in the individual relationships between NA or PA and smoking or craving. We tested the following variables: self-reported number of years smoked; number of cigarettes smoked in past 24 hours at study baseline; depression at baseline (as assessed by the DASS, a composite of the depression and stress subscales); anxiety at baseline (from a composite of DASS anxiety and stress subscales); and sex assigned at birth. The stress subscale of the DASS was included in the measure of depression and anxiety as it is strongly associated with NA and distress is a “more accurate index of overall severity of negative emotion” (Brown, et al., 1997).

Results

Individual-level Hypotheses:

Participants (N=52) responded to 82.2% of survey prompts on average (range = 28.0% - 97.6%). Each participant varied in the amount they reported smoking, with smoking reported on average at 51.06% of observations (range = 2.83% - 99.1%; SD = 29.2%).

Hypotheses 1 and 2: NA and Smoking/Craving.

After running the model for each of the 52 participants, two had significant positive predictive relationships between NA and smoking, and five had significant negative predictive relationships. The individual coefficients are listed in Table 2, with significant values noted with asterisks. Significant beta coefficients ranged from −0.208 to 0.079. Additionally, twenty-one participants had significant relationships between NA and craving, of which 19 were positively predictive and two were negatively predictive of smoking. Table 3 shows the coefficients for this model. Significant beta coefficients ranged from −0.862 to 0.865.

Table 2:

NA and Smoking Beta Coefficients

ID β Z P-value (*= <0.05) AUC
3 0.032 1.892 0.058 0.645
4 −0.008 −0.52 0.603 0.612
5 −0.05 −1.287 0.198 0.577
6 −0.076 −0.902 0.367 0.415
7 −0.009 −0.408 0.683 0.531
8 0.046 1.914 0.056 0.593
9 −0.009 −0.362 0.717 0.603
10 −0.012 −0.594 0.552 0.593
11 0.013 1.415 0.157 0.672
12 0.021 1.319 0.187 0.617
13 −0.072 −3.598 <0.001* 0.753
14 0.011 0.76 0.447 0.573
15 −0.005 −0.475 0.634 0.596
16 −0.023 −1.458 0.145 0.599
17 −0.007 −0.879 0.379 0.569
18 0.057 2.639 0.008* 0.682
19 0.011 0.914 0.361 0.566
20 0.079 2.311 0.021* 0.712
21 −0.024 −1.991 0.046* 0.644
22 0.008 0.543 0.587 0.541
23 −0.047 −1.053 0.292 0.865
24 −0.002 −0.172 0.863 0.532
25 0.045 1.373 0.17 0.598
26 −0.022 −0.671 0.502 0.67
27 0.009 0.858 0.391 0.645
28 −0.02 −0.525 0.599 0.551
29 −0.042 −2.193 0.028* 0.738
30 0.004 0.289 0.773 0.635
31 −0.025 −0.936 0.349 0.56
32 −0.208 −2.347 0.019* 0.854
33 0.015 0.786 0.432 0.636
34 −0.011 −1.276 0.202 0.599
35 −0.005 −0.345 0.73 0.488
36 −0.011 −0.65 0.516 0.55
37 −0.009 −0.65 0.516 0.558
38 0.015 1.051 0.293 0.421
40 −0.027 −0.996 0.319 0.659
41 0.008 0.358 0.72 0.542
42 −0.019 −0.303 0.762 0.596
44 −0.016 −0.814 0.415 0.659
45 −0.002 −0.176 0.86 0.658
46 0.972 1.964 0.05 0.623
47 −0.02 −1.03 0.303 0.589
48 −0.023 −1.049 0.294 0.727
49 −0.012 −0.257 0.797 0.631
50 0.016 0.63 0.528 0.663
51 −0.155 −1.971 0.049* 0.9
52 0.027 1.013 0.311 0.455
53 −0.001 −0.055 0.956 0.518
55 0.034 1.451 0.147 0.668
56 0.003 0.38 0.704 0.572
57 −0.012 −0.625 0.532 0.6
Open in a new tab
Table 3:

NA and Craving Beta Coefficients

ID β T P-value (*= <0.05) R2
3 0.675 4.084 <0.001* 0.15
4 0.447 3.281 0.001* 0.106
5 0.37 2.338 0.021* 0.177
6 0.683 4.319 <0.001* 0.204
7 −0.105 −0.49 0.625 −0.009
8 0.062 0.328 0.743 0.008
9 −0.018 −0.12 0.905 −0.023
10 −0.05 −0.423 0.673 −0.016
11 0.413 4.019 <0.001* 0.243
12 −0.109 −0.818 0.415 0.06
13 0.298 1.977 0.051 0.015
14 0.073 0.366 0.715 −0.008
15 −0.067 −0.396 0.693 0.063
16 −0.826 −5.232 <0.001* 0.262
17 0.136 1.422 0.158 −0.005
18 −0.11 −1.221 0.225 0.002
19 0.818 9.097 <0.001* 0.459
20 0.855 7.207 <0.001* 0.387
21 0.262 1.8 0.075 0.111
22 −0.009 −0.062 0.951 −0.017
23 0.017 0.258 0.797 −0.024
24 0.074 0.486 0.628 0
25 0.418 3.025 0.003* 0.092
26 0.319 0.973 0.338 −0.043
27 0.263 1.449 0.152 0.006
28 0.865 5.32 <0.001* 0.255
29 0.357 2.654 0.009* 0.039
30 0.289 1.249 0.216 0.118
31 0.307 2.367 0.02* 0.038
32 0.019 0.098 0.922 0.373
33 0.177 0.867 0.388 0.075
34 −0.276 −2.086 0.04* 0.031
35 0.43 2.447 0.016* 0.048
36 −0.06 −0.359 0.72 0.025
37 0.009 0.049 0.961 0.149
38 0.611 4.154 <0.001* 0.148
40 0.216 1.455 0.149 0.025
41 −0.309 −0.919 0.36 −0.003
42 −0.115 −0.657 0.512 −0.014
44 0.191 1.365 0.175 0.04
45 0.239 1.456 0.15 0.188
46 0.185 0.164 0.87 −0.017
47 0.431 2.366 0.02* 0.102
48 −0.022 −0.125 0.901 0.041
49 0.266 2.32 0.023* 0.028
50 0.486 3.378 0.001* 0.255
51 0.485 4.138 <0.001* 0.276
52 0.125 0.969 0.335 0.022
53 0.595 3.753 <0.001* 0.204
55 0.274 1.583 0.117 0.04
56 −0.138 −1.211 0.229 0
57 0.506 2.393 0.019* 0.03
Open in a new tab

Figures 1 and 2 show the distributions of all beta coefficients from the intra-individual models of the relationship between (1) NA and smoking and (2) NA and craving.

Figure 1:

Figure 1:

Open in a new tab

Frequency of β coefficients between Smoking & NA

Figure 2:

Figure 2:

Open in a new tab

Frequency of β coefficients between Craving & NA

Hypotheses 3 and 4: PA and Smoking/Craving.

Eight participants had significant relationships between PA and smoking. Three participants had negative relationships, while five had positive relationships. Significant coefficients are listed in Table 4. Twenty-one participants had significant relationships between PA and craving, of which 16 were negative and five were positive. Table 5 shows the coefficients for this model. Significant beta coefficients ranged from −0.837 to 0.636.

Table 4:

PA and Smoking Beta Coefficients

ID β T P-value (*= <0.05) R2
3 −0.017 −1.584 0.113 0.653
4 0.007 0.443 0.658 0.616
5 0.018 0.502 0.615 0.544
6 −0.003 −0.167 0.867 0.502
7 −0.01 −0.692 0.489 0.533
8 −0.008 −0.639 0.523 0.491
9 0.077 2.444 0.015* 0.696
10 −0.015 −0.619 0.536 0.563
11 −0.014 −1.665 0.096 0.696
12 0.001 0.098 0.922 0.533
13 0.051 2.946 0.003* 0.709
14 −0.02 −1.485 0.138 0.427
15 −0.004 −0.463 0.644 0.57
16 0.017 0.996 0.319 0.602
17 −0.004 −0.447 0.655 0.511
18 −0.017 −1.09 0.276 0.56
19 −0.038 −2.368 0.018* 0.715
20 −0.041 −2.093 0.036* 0.654
21 0.047 1.689 0.091 0.621
22 −0.014 −0.778 0.436 0.551
23 0.018 0.319 0.75 0.631
24 0.012 1.016 0.31 0.549
25 −0.03 −1.365 0.172 0.597
26 0.053 2.112 0.035* 0.712
27 −0.008 −0.891 0.373 0.642
28 0.011 0.387 0.698 0.514
29 0.024 1.556 0.12 0.659
30 0.001 0.027 0.978 0.564
31 0.011 0.345 0.73 0.594
32 0.017 0.62 0.536 0.667
33 0 0.033 0.974 0.611
34 0.008 1.077 0.281 0.571
35 −0.004 −0.408 0.683 0.567
36 0.019 1.592 0.111 0.604
37 0.005 0.328 0.743 0.571
38 −0.023 −1.475 0.14 0.667
40 0.025 0.805 0.421 0.664
41 0.017 1.031 0.303 0.575
42 0.248 1.36 0.174 0.981
44 0.028 1.243 0.214 0.669
45 0 0.029 0.977 0.635
46 −0.016 −0.271 0.786 0.604
47 0 −0.024 0.981 0.566
48 0.036 2.14 0.032* 0.768
49 −0.012 −0.209 0.835 0.551
50 0.001 0.044 0.965 0.62
51 0.31 2.302 0.021* 0.948
52 −0.049 −2.175 0.03* 0.67
53 0.014 0.833 0.405 0.46
55 0.001 0.037 0.97 0.618
56 −0.003 −0.394 0.694 0.593
57 0.008 0.564 0.573 0.649
Open in a new tab
Table 5:

PA and Craving Beta Coefficients

ID β T P-value (*= <0.05) R2
3 −0.412 −4.204 <0.001* 0.158
4 −0.301 −2.061 0.042* 0.046
5 −0.213 −1.137 0.258 0.143
6 0.052 0.44 0.661 0.041
7 0.372 2.423 0.017* 0.041
8 −0.144 −0.896 0.372 0.015
9 −0.096 −0.614 0.541 −0.019
10 0.115 0.975 0.331 −0.009
11 −0.237 −2.23 0.029* 0.144
12 0.068 0.472 0.638 0.056
13 −0.313 −2.234 0.028* 0.026
14 0.176 0.999 0.32 0
15 0.121 0.932 0.354 0.07
16 0.636 3.232 0.002* 0.142
17 0.072 0.585 0.56 −0.02
18 0.184 2.186 0.031* 0.031
19 −0.374 −2.882 0.005* 0.062
20 −0.488 −4.756 <0.001* 0.223
21 −0.67 −2.092 0.039* 0.12
22 −0.254 −1.431 0.156 0.004
23 0.076 1.047 0.297 −0.014
24 −0.184 −1.279 0.203 0.012
25 −0.193 −2.059 0.042* 0.051
26 0.105 0.505 0.618 −0.067
27 −0.384 −2.488 0.015* 0.06
28 −0.405 −2.776 0.007* 0.104
29 −0.361 −2.423 0.017* 0.03
30 −0.837 −3.002 0.004* 0.203
31 0.475 3.499 0.001* 0.095
32 0.006 0.04 0.968 0.373
33 −0.066 −0.412 0.681 0.07
34 0.391 3.425 0.001* 0.106
35 −0.327 −2.758 0.007* 0.063
36 0.14 1.297 0.198 0.043
37 0.396 1.823 0.071 0.175
38 −0.46 −2.974 0.004* 0.086
40 −0.195 −1.159 0.249 0.018
41 0.445 1.799 0.075 0.017
42 −0.13 −0.536 0.593 −0.015
44 −0.183 −1.161 0.248 0.035
45 −0.026 −0.17 0.866 0.165
46 −0.109 −0.245 0.807 −0.017
47 −0.234 −1.555 0.123 0.072
48 0.208 1.793 0.076 0.07
49 0.129 0.973 0.333 −0.022
50 −0.586 −4.464 <0.001* 0.309
51 −0.503 −4.146 <0.001* 0.277
52 0.081 0.984 0.327 0.022
53 −0.167 −1.021 0.31 0.097
55 −0.196 −1.106 0.272 0.026
56 0.129 1.097 0.275 −0.002
57 −0.012 −0.061 0.951 −0.034
Open in a new tab

Figures 3 & 4 show the distributions of all the beta coefficients from the intra-individual models of the relationships between (3) PA and smoking and (4) PA and craving.

Figure 3:

Figure 3:

Open in a new tab

Frequency of β coefficients between Smoking & PA

Figure 4:

Figure 4:

Open in a new tab

Frequency of β coefficients between Craving & PA

Group-level Hypotheses:

There was no evidence of significant moderating effects of self-reported number of years smoked, number of cigarettes smoked in 24 hours at baseline, or sex assigned at birth on the relationships reported in Hypotheses 1–4. Further, none of these variables had statistically significant relationships with the percentage of smoking observations or mean craving during the sampling period. All p-values for these group-level hypotheses were above the 0.05 p-value threshold of significance. Across participants, the relationship between PA and smoking was moderated such that participants who had greater baseline anxiety (β = 0.30, p = 0.03) or baseline depression (β =0.32, p = 0.02), had an increasingly positive relationship between PA and smoking. There were no significant relationships between baseline anxiety or depression and PA and craving, nor NA and smoking or craving.

Discussion

This paper sought to address whether there was evidence of a significant relationship between momentary affective experiences and subsequent smoking or craving a cigarette within individuals over time. Additionally, we sought to evaluate the extent to which the relationships between NA, PA, smoking, and craving varied among individuals. Finally, we tested whether select group-level variables moderated the presence or strength of these within-person relationships when aggregated at the group level.

Self-report studies at the group level have found that affective states are related to both smoking and craving (e.g., Shiffman, Paty, Gwaltney, & Dang, 2004). However, EMA studies have complicated this narrative by indicating the momentary relationships between emotions and affect or craving may not be as strong in real-life situations as they are in retrospective self-reports (e.g., Shiffman & Waters, 2004; Shiyko et al., 2013). Consistent with other EMA studies, the present aggregate analyses failed to find evidence that NA or PA were significantly related to smoking or craving in most individuals. Importantly, though, some individuals in the present study did have significant relationships between their momentary affective state and subsequent smoking/craving. By utilizing person-specific analyses of ecological data, we were able to quantify these relationships for each participant.

Hypotheses 1&2.

The results indicated that, for two individuals (3.85%), a momentary increase in NA predicted increased subsequent smoking and an increase in momentary NA predicted increased subsequent craving for 19 individuals (36.5%). One individual had significant positive relationships for both smoking and craving, indicating that they smoked and craved smoking more after experiencing higher NA.

Decreased NA predicted an increase in subsequent smoking for five individuals (9.62%) and higher levels of craving for two individuals (3.84%). These individuals smoked or craved smoking more after observations when they reported lower NA, which is the opposite of what previous group-level research has reported based on self-reports. None of these individuals had negative relationships for both smoking and craving, and two who had negative relationships between NA and smoking had positive relationships between NA and craving. This split in affective motivations for smoking and craving highlights how complex these interactions can be, even within a single person. For these two individuals, they reported craving to smoke more after they reported higher NA, but actually smoked less.

These findings supported the hypotheses that some individuals have significant positive relationships between NA and smoking and craving. Additionally, this latter finding complicates the narrative that increased NA leads to increased smoking in general. This was somewhat unexpected, given that previous studies have found higher NA during withdrawal -- which can reinforce smoking-- and that people who smoke often describe smoking in response to increased NA (e.g., Baker, et al., 2004; Koob, 2015; Piper et al., 2008). A study by Chandra, et al. (2011) found similar results for the effects of NA on smoking, however, they also found that increased smoking was associated with a decrease in NA. Our findings suggest that while this affective relationship is likely an important reason for continued smoking for some, it may be less relevant than other factors for many people who smoke.

Hypotheses 3&4.

The results suggest that higher PA predicted an increase in subsequent smoking for five individuals (9.62%) and increased subsequent craving for five individuals (9.62%). None of the participants both smoked more and craved more following increased PA. Two of the five with positive relationships between smoking and PA had negative relationships with craving following increased PA. This was somewhat surprising, given that higher craving is often reported prior to increased smoking (for a review, see: Wray, et al., 2013). Some data suggest that PA and craving may differ before and during a quit attempt, but the present analyses did not test for this (Bold, Witkiewitz, & McCarthy, 2016).

We also found that PA was negatively associated with subsequent smoking for three individuals (5.77%) and with subsequent craving for 16 individuals (30.77%), indicating that lower PA tended to precede craving for a minority of participants (and smoking, in two cases). Previous studies have found that anhedonia (reduced PA gain from enjoyable experiences) may be a more important factor in motivating smoking than other NA-related symptoms of depression (Cook, Spring, & McChargue, 2007; Leventhal, Piper, Japuntich, Baker, & Cook, 2014). This finding suggests that some individuals indeed experience stronger desire to smoke after experiencing lower PA, though increased smoking following decreased PA might be less common.

Fifteen participants (28.9%) had significant relationships between both NA and craving and PA and craving, with NA being positively and PA being negatively related to craving in all but two cases. This may indicate that changes in affect are more related to craving than to smoking for these individuals, meaning that when they feel more negatively, they crave a cigarette more without necessarily smoking. Alternatively, this could indicate that they experience these patterns of affect in situations where they cannot smoke even if they want, like at school or work. Only two of the thirteen participants with relationships between both NA and PA and craving also had significant relationships for both NA and PA and smoking. Despite not holding true for the majority of participants in this study, the experiences of these individuals fit with previous research indicating that increased NA and decreased PA are related to greater rates of smoking (e.g., Shiffman, 2005; Cook, Spring, McChargue, & Hedeker, 2004).

Overall, the AUC for NA and PA relationships with smoking ranged from 0.415 to 0.981, and the R2 values for the NA or PA relationships with craving in each hypothesis ranged from −0.07 to 0.46, indicating that affective states may explain up to just under half of the variation in smoking behavior and craving in particular individuals. The heterogeneity of AUC and R2 values between individuals is consistent with and supports prior findings that the affective mechanisms for maintaining cigarette smoking (Leventhal, 2010; Zheng et al., 2013) and substance use in general (Soyster, Ashlock, and Fisher, 2021) are different between individuals. The aggregated results did not support any single hypothesis at the group level, but significant effects of both NA and PA in individuals indicate that continued person-specific approaches to understanding cigarette use patterns may further improve our understanding of how smoking is reinforced and maintained over time within individuals.

By providing information about the emotional context in which smoking occurs, idiographic sampling may be a valuable tool to support individual behavior change. The relationships between affect and craving have been shown to change over time during smoking cessation, so using information about the emotional context that precedes smoking and/or craving may improve clinicians’ abilities to suggest person-specific strategies (Lanza, Vasilenko, Liu, Li, & Piper, 2013). For example, if a person’s craving is strongly associated with their NA but not with their PA, this person may smoke less after learning new approaches for down-regulating NA but not for increasing PA. However, a person who only has relationships between PA and smoking/craving may benefit more from addressing anhedonia in situations where they normally smoke and do not need any assistance regulating NA.

In this sample the lack of connection between affect and smoking may be less informative than the presence of a connection between affect and craving. For example, a person who only had relationships between NA and craving felt stronger craving for cigarettes when they felt more NA, even if they rarely smoked. Since craving is informative of moments when they are predisposed to smoke, craving responses may still be clinically relevant for smoking cessation (Lanza, et al., 2013).

Group-level.

We also examined whether group-level variables predicted the nature or direction of the relationships between affect and smoking. There was no significant evidence of any group-level moderating effects of self-reported number of years smoked, number of cigarettes smoked in 24 hours, or sex assigned at birth on these findings. These results do not support the hypotheses that these categories represent homogeneous sub-populations of tobacco users who have reliably similar relationships between affect and tobacco use. We did find evidence of significant positive relationships between baseline depression and anxiety severity and the PA and smoking relationships. Given our modest between-subject sample size, we hesitate to speculate about a population-level estimate of these effects. With that said, that an increased level of baseline anxiety or depression was associated with a more positive relationship between momentary PA and subsequent smoking is consistent with the possibility that tobacco use is differentially rewarding for these populations. Perhaps these participants were using cigarettes to maximize or extend their PA in a way others were not. It is also possible that these participants were more likely to experience PA in situations where tobacco use is also common (e.g., at a party). Future EMA studies that assess situational contexts and subjective smoking motives would be well-suited to investigate these possibilities.

Limitations:

The findings of the present study should be evaluated in the context of several limitations. There was considerable range in the goodness of fit indices for the NA and PA measurement models among participants (Table 1), indicating that these constructs, as defined in the present analyses, may not fully capture the nuances of person-specific affective experiences. For example, the NA model exhibited good fit for P12 (RMSEANA = 0.03) but poor fit for P3 (RMSEANA = 0.106). These differences, while in some ways reinforcing the need for individualized modeling of psychological constructs (see Fisher et al., 2018), limit confidence in the results of some of the idiographic models. Although we allowed for person-specific factor loadings for NA and PA, which did not appear to work well for every participant, we did not allow for person-specific indicators of affect. In future idiographic studies, researchers may consider creating person-specific models of independent variables (e.g., a person-specific measure of NA as represented by the individual’s experience) to overcome this limitation.

Overall, while our within-subject sample sizes were relatively large, the between-subjects sample size for these data was relatively small, meaning our statistical power to detect group-level differences was small, and any generalized conclusions from the group-level analyses were limited. Thus, our between-subjects results should be interpreted more as descriptive of this sample than representative of the population.

In the present study, EMA sampling took place approximately every four hours during individuals’ reported waking schedules, yet the precise time scales on which the hypothesized mechanisms proceed are not known. Even if a participant completed all of the EMA pings, we only had observations approximately every four hours. If the relationships between momentary affect and smoking/craving actually unfold faster (i.e., a brief spike in NA leads to craving minutes, rather than hours, later) our sampling frequency would not be fast enough to capture these relationships.

Though EMA has the unique strength of being able to acquire large amounts of within-subject data points in real time, there is limited information on how environmental contexts might influence these measures. For example, when a person completed an observation, we did not gather data about who they were with or what they were doing at that time that could have influenced both their affect and their craving/smoking behavior. Analysis data were only available for times when people were able to answer the surveys, which means there is limited data on situations where they were unable to use their phones but may have had changes in affect or smoking/craving (e.g., at work or at school). Some individuals may have had better social support or mechanisms for coping with their affective states which could not be reflected in their EMA responses.

These missing environmental factors may offer some explanation as to why more individuals showed relationships between affect and craving than affect and smoking. Smoking is often limited by certain environments (e.g., at school or work), while craving is an experience that can occur regardless of the situation and is not necessarily followed by smoking in all cases. For some individuals, stronger relationships between craving and affect may have been the result of such constraints.

Future Directions:

Studies with larger between-subjects sample sizes would be better suited to address whether moderating variables influence these findings at the group-level. An EMA study with similar constructs by Gunter, et al., (2020) suggests that these relationships may differ across cultures. Even within the US, smoking prevalence varies significantly between groups (USDHHS, 2020). The present study did not include analyses of race, national origin, culture, or socioeconomic status. Future research would be useful to understand the extent to which these relationships may change across different groups.

Future research could also determine the precise time scale of intraindividual emotional changes in tobacco users to determine the optimal sampling frequency to get more representative data to analyze the connections between affect and use. The duration of the EMA sampling period could be increased in order to analyze how individuals’ emotional experiences, smoking habits, and the relationships between them change over a longer term of their life. For example, it would be interesting to repeat this EMA collection with the same group of participants to see if their individual relationships between affect and smoking/craving have changed significantly over months or years of use.

Additionally, while evidence fails to support the existence of the hypothesized relationships in most individuals, future work could try to identify any commonalities which would make individuals with significant affective relationships easier to identify. Some of the individuals who had significant relationships between affect and smoking had the opposite relationship between smoking and craving. Future research could investigate how common such experiences are at the group level and try to understand what additional factors influence this divergence. For example, someone who has increased craving following a change in affect without increasing their smoking may provide valuable insight into the conscious control of smoking behavior.

Conclusion

Smoking is an avoidable health risk and preventable cause of death. Many factors influence a person to start smoking, but nicotine addiction plays a key role in maintaining smoking long-term (e.g., Benowitz, 2010; De Biasi & Dani, 2011). Changes in affect have been theorized to be related to craving and smoking cigarettes, and evidence from group-level studies supports this (Baker et al., 2004). The present findings-- that NA and PA at a given time point are related to smoking and craving over time for some individuals-- indicates the complexity of how variation in group-level findings manifests within individuals. Ultimately, this suggests that there is potential for future research to quantify variation in affect and tobacco-use patterns within and between individuals as a step towards personalized clinical interventions for smoking cessation.

Acknowledgements

We would like to thank Dr. Aaron Fisher, Dr. Hannah Bosley, and Apoorva Polisetty for their feedback on early versions of this manuscript.

Funding Statement

The project described was supported by Award Number T32AA007240, Graduate Research Training in Alcohol Problems: Alcohol-Related Disparities and P50AA005595, Epidemiology of Alcohol Problems: Alcohol-related Disparities, from the National Institute on Alcohol Abuse and Alcoholism. The content is solely the responsibility of the authors and does not necessarily represent the official view of the National Institute on Alcohol Abuse and Alcoholism or the National Institutes of Health.

Appendix

Appendix 1: List of questions included on the EMA Survey.

  1. Smoked (yes/no)

  2. How strong is your urge/craving to smoke right now?

  3. I enjoyed my last cigarette

  4. People or situations are triggering me to smoke

  5. I feel irritable

  6. I feel angry

  7. I am having difficulty concentrating

  8. I am hungry

  9. I feel restless

  10. I feel down/depressed

  11. I feel bored

  12. I feel stressed

  13. I feel calm/relaxed

  14. I feel happy

  15. I feel enthusiastic

  16. I feel nervous/tense

  17. I feel tired

  18. I feel frustrated

  19. I feel impulsive

  20. My smoking is hurting my health

  21. I am motivated to quit smoking

  22. I want to quit smoking

  23. A cigarette would improve my mood or make me feel better

  24. I am embarrassed/ashamed that I am a smoker

  25. The amount I smoke is within my control

  26. If I tried to quit smoking right now; I would be successful

  27. I can delay gratification

  28. My health would improve if I quit smoking

  29. I feel comfortable in my current location/situation

  30. I am enjoying my interactions with other people

Footnotes

Declaration of Interest Statement

The authors report no conflict of interest.

Data Availability Statement

Data and files available at the following link: https://osf.io/uyr48/?view_only=71dae47d245e492c8c75590937487d74

References

  1. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). 10.1176/appi.books.9780890425596 [DOI]
  2. Agaku IT, King BA, & Dube SR (2014). Current Cigarette Smoking Among Adults—United States, 2005–2012. MMWR. Morbidity and Mortality Weekly Report, 63(2), 29–34. [PMC free article] [PubMed] [Google Scholar]
  3. Baker TB, Piper ME, McCarthy DE, Majeskie MR, & Fiore MC (2004). Addiction Motivation Reformulated: An Affective Processing Model of Negative Reinforcement. Psychological Review, 111(1), 33–51. 10.1037/0033-295X.111.1.33 [DOI] [PubMed] [Google Scholar]
  4. Benowitz NL (2009). Pharmacology of nicotine: addiction, smoking-induced disease, and therapeutics. Annual review of pharmacology and toxicology, 49, 57–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Benowitz NL (2010). Nicotine Addiction. New England Journal of Medicine, 362(24), 2295–2303. 10.1056/NEJMra0809890 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bold KW, Witkiewitz K, & McCarthy DE (2016). Multilevel factor analysis of smokers’ real-time negative affect ratings while quitting. Psychological assessment, 28(9), 1033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Brown TA, Chorpita BF, Korotitsch W, & Barlow DH (1997). Psychometric properties of the Depression Anxiety Stress Scales (DASS) in clinical samples. Behaviour research and therapy, 35(1), 79–89. [DOI] [PubMed] [Google Scholar]
  8. Carter B, Lam C, Robinson J, Paris M, Waters A, Wetter D, & Cinciripini P (2008). Real-time craving and mood assessments before and after smoking. Nicotine & Tobacco Research, 10(7), 1165–1169. 10.1080/14622200802163084 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chandra S, Scharf D, & Shiffman S (2011). Within-day temporal patterns of smoking, withdrawal symptoms, and craving. Drug and alcohol dependence, 117(2–3), 118–125. 10.1016/j.drugalcdep.2010.12.027 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Changeux J-P (2010). Nicotine addiction and nicotinic receptors: Lessons from genetically modified mice. Nature Reviews Neuroscience, 11(6), 389–401. 10.1038/nrn2849 [DOI] [PubMed] [Google Scholar]
  11. Christakis NA, & Fowler JH (2008). The Collective Dynamics of Smoking in a Large Social Network. New England Journal of Medicine, 358(21), 2249–2258. 10.1056/NEJMsa0706154 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cook JW, Spring B, McChargue D, & Hedeker D (2004). Hedonic capacity, cigarette craving, and diminished positive mood. Nicotine & Tobacco Research, 6(1), 39–47. 10.1080/14622200310001656849 [DOI] [PubMed] [Google Scholar]
  13. Cook JW, Spring B, & McChargue D (2007). Influence of nicotine on positive affect in anhedonic smokers. Psychopharmacology, 192(1), 87–95. 10.1007/s00213-006-0688-5 [DOI] [PubMed] [Google Scholar]
  14. De Biasi M, & Dani JA (2011). Reward, Addiction, Withdrawal to Nicotine. Annual Review of Neuroscience, 34(1), 105–130. 10.1146/annurev-neuro-061010-113734 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Doran N, Spring B, Borrelli B, McChargue D, Hitsman B, Niaura R, & Hedeker D (2006). Elevated positive mood: A mixed blessing for abstinence. Psychology of Addictive Behaviors, 20(1), 36–43. https://doi-org.libproxy.berkeley.edu/10.1037/0893-164X.20.1.36 [DOI] [PubMed] [Google Scholar]
  16. Dunbar MS, Scharf D, Kirchner T, & Shiffman S (2010). Do smokers crave cigarettes in some smoking situations more than others? Situational correlates of craving when smoking. Nicotine & tobacco research, 12(3), 226–234. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fisher AJ, Newman MG, & Molenaar PCM (2011). A quantitative method for the analysis of nomothetic relationships between idiographic structures: Dynamic patterns create attractor states for sustained posttreatment change. Journal of Consulting and Clinical Psychology, 79, 552–563. 10.1037/a0024069 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Fisher AJ (2015). Toward a dynamic model of psychological assessment: Implications for personalized care. Journal of Consulting and Clinical Psychology, 83(4), 825–836. 10.1037/ccp0000026 [DOI] [PubMed] [Google Scholar]
  19. Fisher AJ, Medaglia JD, & Jeronimus BF (2018). Lack of group-to-individual generalizability is a threat to human subjects research. Proceedings of the National Academy of Sciences, 115(27), E6106–E6115. 10.1073/pnas.1711978115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Fox J, &Weisberg S (2019). An {R} Companion to Applied Regression, Third Edition. Thousand Oaks CA: Sage. URL: https://socialsciences.mcmaster.ca/jfox/Books/Companion/ [Google Scholar]
  21. George O, & Koob GF (2017). Overview of Nicotine Withdrawal and Negative Reinforcement (Preclinical). In Negative Affective States and Cognitive Impairments in Nicotine Dependence (pp. 1–20). Elsevier. 10.1016/B978-0-12-802574-1.00001-6 [DOI] [Google Scholar]
  22. Gunter RG, Szeto EH, Suh S, Kim Y, Jeong SH, & Waters AJ (2020). Associations between Affect, Craving, and Smoking in Korean Smokers: An Ecological Momentary Assessment study. Addictive Behaviors Reports, 100301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hu L, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. 10.1080/10705519909540118 [DOI] [Google Scholar]
  24. Hu SS (2016). Tobacco Product Use Among Adults—United States, 2013–2014. MMWR. Morbidity and Mortality Weekly Report, 65. 10.15585/mmwr.mm6527a1 [DOI] [PubMed] [Google Scholar]
  25. Ikard FF, Green DE, & Horn D (1969). A scale to differentiate between types of smoking as related to the management of affect. International Journal of the Addictions, 4(4), 649–659. [Google Scholar]
  26. Ikard FF, & Tomkins S (1973). The experience of affect as a determinant of smoking behavior: A series of validity studies. Journal of Abnormal Psychology, 81(2), 172–181. 10.1037/h0034555 [DOI] [PubMed] [Google Scholar]
  27. Koob GF (2015). The dark side of emotion: the addiction perspective. European journal of pharmacology, 753, 73–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Koob GF, & Le Moal M (1997). Drug abuse: hedonic homeostatic dysregulation. Science, 278(5335), 52–58. [DOI] [PubMed] [Google Scholar]
  29. Lanza ST, Vasilenko SA, Liu X, Li R, & Piper ME (2013). Advancing the understanding of craving during smoking cessation attempts: A demonstration of the time-varying effect model. Nicotine & Tobacco Research, 16(Suppl_2), S127–S134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Leventhal AM, Waters AJ, Moolchan ET, Heishman SJ, & Pickworth WB (2010). A quantitative analysis of subjective, cognitive, and physiological manifestations of the acute tobacco abstinence syndrome. Addictive Behaviors, 35(12), 1120–1130. 10.1016/j.addbeh.2010.08.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Leventhal AM, Piper ME, Japuntich SJ, Baker TB, & Cook JW (2014). Anhedonia, depressed mood, and smoking cessation outcome. Journal of consulting and clinical psychology, 82(1), 122. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lovibond PF, & Lovibond SH (1995). The structure of negative emotional states: Comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and Anxiety Inventories. Behaviour research and therapy, 33(3), 335–343. [DOI] [PubMed] [Google Scholar]
  33. McLaughlin I, Dani JA, De Biasi M (2015) Nicotine Withdrawal. In: Balfour D, Munafò M (eds) The Neuropharmacology of Nicotine Dependence. Current Topics in Behavioral Neurosciences, vol 24. Springer, Cham: [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Molenaar PC (2004) A Manifesto on Psychology as Idiographic Science: Bringing the Person Back Into Scientific Psychology, This Time Forever, Measurement: Interdisciplinary Research and Perspectives, 2:4, 201–218, DOI: 10.1207/s15366359mea0204_1 [DOI] [Google Scholar]
  35. Piccirillo ML, & Rodebaugh TL (2019). Foundations of idiographic methods in psychology and applications for psychotherapy. Clinical Psychology Review, 71, 90–100. 10.1016/j.cpr.2019.01.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Piper ME, Federmen EB, McCarthy DE, Bolt DM, Smith SS, Fiore MC, & Baker TB (2008). Using mediational models to explore the nature of tobacco motivation and tobacco treatment effects. Journal of Abnormal Psychology, 117(1), 94–105. 10.1037/0021-843X.117.1.94 [DOI] [PubMed] [Google Scholar]
  37. R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. [Google Scholar]
  38. Revelle W (2022). psych: Procedures for Psychological, Psychometric, and Personality Research (2.2.5). [Computer software]. https://CRAN.R-project.org/package=psych.
  39. Robinson JD, Lam CY, Carter BL, Minnix JA, Cui Y, Versace F, Wetter DW, & Cinciripini PM (2011). A multimodal approach to assessing the impact of nicotine dependence, nicotine abstinence, and craving on negative affect in smokers. Experimental and Clinical Psychopharmacology, 19(1), 40–52. 10.1037/a0022114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Rosseel Y (2012). lavaan: An R Package for Structural Equation Modeling. Journal of Statistical Software, 48(2), 1–36. 10.18637/jss.v048.i02 [DOI] [Google Scholar]
  41. Schork NJ (2015). Personalized medicine: Time for one-person trials. Nature, 520(7549), 7549. 10.1038/520609a [DOI] [PubMed] [Google Scholar]
  42. Shiffman S, Paty JA, Gwaltney CJ, & Dang Q (2004). Immediate antecedents of cigarette smoking: an analysis of unrestricted smoking patterns. Journal of Abnormal Psychology, 113(1), 166. [DOI] [PubMed] [Google Scholar]
  43. Shiffman S (2005). Dynamic Influences on Smoking Relapse Process. Journal of Personality, 73(6), 1715–1748. 10.1111/j.0022-3506.2005.00364.x [DOI] [PubMed] [Google Scholar]
  44. Shiffman S (2009). Ecological momentary assessment (EMA) in studies of substance use. Psychological assessment, 21(4), 486. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Soyster PD, & Fisher AJ (2019). Involving stakeholders in the design of ecological momentary assessment research: An example from smoking cessation. PLOS ONE, 14(5), e0217150. 10.1371/journal.pone.0217150 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Swan GE, Ward MM, & Jack LM (1996). Abstinence effects as predictors of 28-day relapse in smokers. Addictive behaviors, 21(4), 481–490. [DOI] [PubMed] [Google Scholar]
  47. U.S. Department of Health and Human Services. (2020) Smoking Cessation. 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]
  48. Verduyn P, Van Mechelen I, Tuerlinckx F, Meers K, & Van Coillie H (2009). Intensity profiles of emotional experience over time. Cognition and Emotion, 23(7), 1427–1443. [Google Scholar]
  49. Wang TW, Asman K, Gentzke AS, Cullen KA, Holder-Hayes E, Reyes-Guzman C, Jamal A, Neff L, & King BA (2018). Tobacco Product Use Among Adults—United States, 2017. 67(44), 8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Watkins S, Koob GF, Athina S (2000). Neural mechanisms underlying nicotine addiction: Acute positive reinforcement and withdrawal. Nicotine & Tobacco Research, 2(1), 19–37. 10.1080/14622200050011277 [DOI] [PubMed] [Google Scholar]
  51. Wray JM, Gass JC, & Tiffany ST (2013). A systematic review of the relationships between craving and smoking cessation. Nicotine & Tobacco Research, 15(7), 1167–1182. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Zhang D (2018). rsq: R-Squared and Related Measures. R package version 1.1. https://CRAN.R-project.org/package=rsq [Google Scholar]
  53. Zheng Y, Wiebe RP, Cleveland HH, Molenaar PC, & Harris KS (2013). An Idiographic Examination of Day-to-Day Patterns of Substance Use Craving, Negative Affect and Tobacco Use among Young Adults in Recovery. Multivariate behavioral research, 48(2), 241–266. 10.1080/00273171.2013.763012 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Data Availability Statement

Data and files available at the following link: https://osf.io/uyr48/?view_only=71dae47d245e492c8c75590937487d74

RESOURCES