Abstract

Introduction

The study of happiness and life satisfaction at an individual and societal level has gained importance not only for psychological science (Diener and Chan, 2011; Forgeard et al., 2011; Kahneman, 2011), but also for social and economic policy (e.g., Organization for Economic Cooperation and Development [Oecd], 2001, 2011). Consequently, the optimal methods for the routine assessment of these factors in real-life situations require examination. Previous research examining happiness and satisfaction has typically assessed such variables using questionnaires (Diener et al., 1985; Fordyce, 1988) administered face to face with respondents. In addition, the day reconstruction method (DRM) has also been used to retrospectively measure the frequency and intensity of positive and negative emotions over a period of 24 h (Kahneman et al., 2004). Relatedly, experience sampling methods (ESM; Csikszentmihalyi et al., 1977; Scollon et al., 2007) use electronic devices (e.g., pager and mobile phone) to signal respondents during the day to answer questions regarding their current feelings; such studies typically last 1–2 weeks (Reis and Gable, 2000). ESM overcomes the memory biases associated with retrospective methods, such as the DRM, by asking respondents to describe their current psychological state when answering the questions. Furthermore, by collecting longitudinal data from a sample with repeated measures ESM facilitates detailed examination of within-person fluctuations in happiness and satisfaction. A number of previous studies (Emmons, 1986; Freeman et al., 1986; Moneta et al., 2001) have used ESM to measure participant mood states over time but such data collection was done using written forms and web-forms. Although collecting such data was initially very costly and time-consuming, ESM can now be easily implemented using mobile phone technology.

Short message service (SMS) texting provides the potential to routinely collect data from large-scale representative samples from the population at little cost. As noted by Miller (2012), due to their ubiquity, phones can facilitate the gathering of ecologically valid data on the experiences of millions of people in a convenient and unobtrusive manner. The widespread adoption of mobile technology by populations opens up new research programs and provides enhanced understanding of the everyday experience of individuals (Reiss et al., 2014). For the past two decades researchers have used mobile electronic devices to collect research data and although SMS texting has been used by psychologists in time-based studies to look at mood variations (Killingsworth and Gilbert, 2010), the method has not been used to measure the happiness and satisfaction of a sample of people over time.

In addition to collecting data, texting can also be used to provide information to research participants. For example, participants could receive feedback on the average levels of happiness and satisfaction for the sample. Such information may increase participants’ motivation to continue to provide data, which may overcome well-noted problems arising from high levels of attrition in longitudinal research (Patel et al., 2003). Previous studies have established that providing personalized feedback can successfully reduce attrition rates in longitudinal interventions and surveys. Marcus et al. (2007) found that providing personalized feedback to participants in a randomized controlled trial (RCT) to enhance physical activity maintained a high level of participation. Bewick et al. (2008) reported that providing personalized feedback on alcohol consumption was effective for retaining participants in the research. Fumagalli et al. (2013) found that providing salient feedback was more successful in reducing survey attrition rates than financial incentives. Providing participants with personalized information that is highly salient to the individual can maintain high levels of participation rates; however, the effect of giving non-personalized information is unclear. Research indicates that provision of even minimal feedback can give a task purpose, which consequently increases motivation of participants to perform at levels above those for individuals not receiving feedback (Ariely et al., 2009). Furthermore, while general feedback may enhance motivation, awareness of the average level of happiness and satisfaction among a general sample might provide a reference point for comparison that influences the individuals’ subsequent ratings.

Social comparison theory states that when objective standards do not exist, people compare their beliefs, achievements and abilities with others’ beliefs, achievements and abilities (Festinger, 1954). The referent group for such comparisons varies across time and context; however, research indicates that social comparisons are associated with levels of happiness (Gutek et al., 1983; Emmons and Diener, 1985; Diener and Shigehiro, 2005). Once presented with the mean level for others, participants may be motivated to increase their own reported levels of happiness and satisfaction to be “better-than-average”, a pervasive self-enhancement bias noted in the literature (Alicke and Govorun, 2005). Alternatively, individuals may use such values as anchors that may subsequently bias their subsequent happiness and satisfaction ratings toward the mean values (Furham and Boo, 2011).

In summary, the study contributes to the literature by using SMS to collect weekly data on happiness and satisfaction in a large national sample. The provision of general feedback offers a novel distinction between the present study and the previous studies, which have used personalized feedback typically related to a salient health behavior. In addition, the effects of receiving feedback on the average level of happiness and satisfaction from the previous week on (i) participant retention, and (ii) subsequent levels of happiness and satisfaction were assessed.

Materials and Methods

Results

The sample was predominantly female (76%) and the average age was 45.3 years (SD = 13.4, range 18–84). One half of the sample was aged between 40 and 59 years; over half were parents (59%). The weekly ratings of happiness and satisfaction are presented in Table ​Table11.

Over the course of the 6 weeks the mean level of happiness was 6.80 (SD = 1.38) and was 6.66 (SD = 1.54) for satisfaction. There were no significant gender differences for either average happiness [M_female = 6.79 vs. M_male = 6.82, t(2,011) = -0.39, p = 0.66, d = 0.01] or satisfaction [M_female = 6.69 vs. M_male = 6.87, t(1,806) = -0.001, p = 0.93; d < 0.01]. There was a significant non-linear relationship between age and both happiness (Spearman’s ρ = 0.16, p < 0.001) and satisfaction (Spearman’s ρ = 0.16, p < 0.001); the mean levels for each variable across different age categories displayed a U-shaped curve (Figure ​Figure11).

Open in a separate window

FIGURE 1

M (SE) happiness and satisfaction levels by age category.

Discussion

Text messaging provided data on national levels of happiness and satisfaction for a large sample in a novel longitudinal study. The present results are similar to those reported using traditional pen-and-paper methods to assess happiness and satisfaction; national levels of happiness and satisfaction are in line with the values reported using standard survey methodology (Organization for Economic Cooperation and Development [Oecd], 2011). In addition, the analysis revealed a U-shaped relationship between age and both happiness and satisfaction. Such a finding is congruent with recent research on happiness; for example, Blanchflower and Oswald (2008) used a variety of data sets that could distinguish age effects from cohort effects to suggest that the U-shaped structure of the curve is broadly consistent across countries throughout the globe. Overall, the present findings are consistent with those reported using standard methods and provide evidence for the validity of the method to obtain meaningful psychological data in an efficient manner using mobile phone technology.

Those who received feedback were more likely to provide complete data across repeated assessments; this pattern is consistent with Ariely et al.’s (2009) argument that feedback provides ‘minimal meaning’ to a task and enhances motivation regarding task performance. Those who received feedback did not remain in the study due to feeling happier or more satisfied in general. Provision of summary data on happiness and satisfaction did not systematically bias subsequent ratings in relation to sample mean normative anchors; this finding, if replicated, indicates that use of such minimal feedback does not distort individual responses. However, it should be noted that although the effect sizes were small (-0.07 < d < 0.03), the differences at week 5 would be considered statistically significant using a less stringent value of 0.05. Clearly, addition research is required to determine whether such differences represent systematic and substantive effects of generalized feedback.

Furthermore, it must be acknowledged that participants may not have perceived the generalized feedback (i.e., the mean ratings) as being salient; given the diversity of the sample (e.g., in terms of age range) used to calculate the means participants may not have regarded the mean value as being relevant. For example, the significant association between age and happiness/satisfaction ratings (Figure ​Figure11) indicates that different age groups may have different normative levels of happiness and satisfaction. Given such an association, it is plausible that an effect of feedback could be significant if such feedback was tailored based on the participant’s demographic profile (e.g., age or gender). Consequently, although general feedback had little effect on ratings, it is possible that salient personalized feedback would serve as an anchor to subsequently bias individuals’ subsequent happiness and satisfaction ratings toward the mean values (Furham and Boo, 2011). Such a possibility requires empirical investigation.

In comparison to standard longitudinal research methods (Patel et al., 2003), the relatively low drop-out rate, particularly among those receiving generalized feedback, is a novel finding that highlights the value of texting as a simple and effective means to provide longitudinal data on psychological phenomena. Previous research had documented that personally salient feedback impacted upon participation; the present study adds to this literature by providing evidence that generalized feedback may have a similar effect. This study also indicates one of the means by which texting data can contribute to producing ‘Big Data’ and addressing macropsychology questions (MacLachlan, 2014).

Conclusion

The present study provides novel data to indicate that collecting longitudinal happiness and satisfaction data from a large sample is feasible using SMS text messaging. Provision of summary feedback to participants may diminish attrition rates without biasing their individual responses relative to potential normative anchors or social comparison targets.

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

Footnotes

References