This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Formative Research, is properly cited. The complete bibliographic information, a link to the original publication on https://formative.jmir.org, as well as this copyright and license information must be included.
Digital tools may help to address social deficits in schizophrenia, particularly those that engage social comparison processes (ie, evaluating oneself relative to others). Yet, little is known about social comparison processes in schizophrenia or how best to capture between- versus within-person variability, which is critical to engaging comparisons in digital interventions.
The goals of this pilot study were to (1) better understand affective responses to social comparisons among individuals with schizophrenia, relative to healthy controls, using a validated global self-report measure; and (2) test a new brief, mobile assessment of affective responses to social comparison among individuals with schizophrenia, relative to the full measure. This study was conducted in 2 phases.
We first compared self-reported affective responses to social comparisons between individuals with schizophrenia (n=39) and healthy controls (n=38) using a traditional self-report measure, at 2 time points. We examined the temporal stability in responses and differences between groups. We then evaluated the performance of brief, mobile assessment of comparison responses among individuals with schizophrenia, completed over 12 weeks (n=31).
Individuals with schizophrenia showed greater variability in affective responses to social comparison than controls on traditional measures and completed an average of 7.46 mobile assessments over 12 weeks. Mobile assessments captured within-person variability in affective responses in the natural environment (intraclass correlation coefficients of 0.40-0.60). Average scores for mobile assessments were positively correlated with responses to traditional measures.
Affective responses to social comparison vary both between and within individuals with schizophrenia and capturing this variability via smartphone surveys shows some evidence of feasibility. As affective variability is a potential indicator of poor outcomes among individuals with mental health conditions, in the future, a brief, mobile assessment of affective responses to social comparisons may be useful for screening among individuals with schizophrenia. Further research on this process is needed to identify when specific comparison messaging may be most effective in digital interventions and could suggest new therapeutic targets for illnesses such as schizophrenia.
Schizophrenia currently affects approximately 1% of the US population [
Across health care, digital technologies have the potential to increase access to and improve quality of care. Digital treatments for mental health conditions, such as those delivered via smartphone apps, are popular and over 10,000 already exist [
In addition to processes such as social support, the efficacy of social networks to drive engagement (and consequent behavior change) rests in part on
Specifically, the Identification/Contrast Model [
In addition, there is ample evidence that people with mental health conditions (eg, major depression, anxiety disorders) may use and respond to comparisons differently than people without these conditions [
Although both upward and downward comparisons are common in illnesses such as cancer [
Effective assessments should be ecologically valid and respond to known contextual influences on social comparison processes [
How do self-reported responses to social comparisons among individuals with schizophrenia compare with those of healthy controls? We expected to observe stronger negative or weaker positive responses to comparisons among those with schizophrenia.
A. How much within-person variability is there in self-reported responses to social comparisons? We expected to observe meaningful within-person variability in affective responses to social comparisons.
B. Does variability differ between individuals with schizophrenia and healthy controls? We expected to observe greater within-person variability in affective responses to social comparisons among individuals with schizophrenia.
Among individuals with schizophrenia, does a brief mobile assessment of self-reported responses to social comparisons show convergent validity with the full scale? We expected to observe moderate to strong correlations between scores on the full and mobile versions of this measure.
Procedures were approved by the institutional review board at the supporting institution and all participants provided written informed consent. Adults with schizophrenia were recruited from outpatient clinics in a large city in the northeastern United States region, where diagnosis was verified through clinical records. Control participants were recruited via online social media postings targeting college students in the same city. Control participants were assessed in person and were eligible if they did not screen positive for mental illness based on the Mini International Neuropsychiatric Interview [
Participants were 39 patients with schizophrenia (20/39, 51%, men; mean age 37.45 [SD 14.86] years) and 38 healthy controls (17/38, 45%, men; mean age 30.50 [SD 16.65] years;
During 12 weeks of mobile assessment that occurred between clinic visits, the mobile version of the Identification-Contrast Scale developed for this study (also described below) appeared twice per week, among patients with schizophrenia only. A total of 24 patients completed mobile assessments during the 12-week window. Participants were oriented to the questions in person. When using the app between sessions, participants were free to ignore any mobile surveys and were not compensated on the basis of completion.
Demographic information for individuals with schizophrenia and healthy controls.
Demographic | Individuals with schizophrenia (n=39) | Healthy controls (n=38)a | |
Age, mean (SD) | 37.45 (14.86) | 30.50 (16.65) | |
|
|
|
|
|
Men | 20 (51) | 17 (45) |
|
Women | 19 (49) | 19 (50) |
|
|
|
|
|
American Indian or Alaskan native | 4 (10) | 0 (0) |
|
Asian American | 1 (3) | 25 (66) |
|
Black or African American | 11 (28) | 3 (8) |
|
Multiracial or other | 1 (3) | 2 (5) |
|
White | 21 (54) | 6 (16) |
|
|
|
|
|
Four-year college graduate or higher | 14 (36) | 30 (79) |
|
Some college | 11 (28) | 3 (8) |
|
High-school graduate/General Educational Development | 11 (28) | 3 (8) |
|
Some high school | 3 (8) | 0 (0) |
aTwo participants did not provide complete demographic information.
bOne participant did not specify their race.
The Identification-Contrast Scale [
The mobile version of the Identification-Contrast Scale was designed to maximize the power of the full scale while limiting the number of items to be completed in the natural environment. To achieve this balance, the item on each scale with the highest factor loadings was selected for delivery via mobile app [
Descriptive statistics for each subscale of the full Identification/Contrast measure included means and SDs for each group (individuals with schizophrenia vs. healthy controls) at each time point. To address the first research question, independent
The third research question was addressed in 2 ways. First, descriptive information was examined to determine how often individuals with schizophrenia completed mobile assessments of social comparison responses and how much variability in their responses was between- versus within-person. Second, bivariate correlations (
The Institutional review board at Beth Israel Deaconess Medical Center has approved this study (institutional review board protocol number: 2017P000359).
Descriptive statistics for each group by time point are presented in
Descriptive statistics for traditional self-report measures and differences between individuals with schizophrenia and healthy controls.
Response to comparison | Individuals with schizophrenia, mean (SD) | Healthy controls, mean (SD) | Differences between samples | ||||
|
|
|
|
||||
|
Upward identification | 4.06 (1.08) | 4.19 (0.99) | t76=–0.57, |
|||
|
Upward contrast | 2.54 (1.35) | 1.81 (0.92) | t76=2.82b, |
|||
|
Downward identification | 2.06 (1.26) | 1.37 (0.55) | t76=3.10b, |
|||
|
Downward contrast | 3.36 (1.26) | 3.78 (0.97) | t76=1.69, |
|||
|
|
|
|
||||
|
Upward identification | 3.92 (1.11) | 4.01 (0.86) | t56=0.33, |
|||
|
Upward contrast | 2.57 (1.29) | 2.07 (1.18) | t56=1.54, |
|||
|
Downward identification | 2.22 (1.26) | 1.49 (0.71) | t56=2.66d, |
|||
|
Downward contrast | 2.88 (1.42) | 3.62 (0.91) | t56=–2.35d, |
an=39 and 38 for columns 2 and 3, respectively.
b
cn=31 and 28 for columns 2 and 3, respectively.
d
Across time points and participant groups, ICCs for upward and downward identification were 0.40 and 0.41, respectively, indicating that approximately 40% of variability in these tendencies was due to stable, between-person differences, whereas 60% was due to within-person variation (and error). Stability estimates for upward and downward contrast were slightly higher (ICCs 0.60 and 0.57, respectively), though within-person variation components for all 4 scales were statistically significant (
Among individuals with schizophrenia, there was considerable between-person variability in the number of mobile assessments of social comparison responses completed during the 12-week assessment window. These individuals completed assessments between 1 and 28 times, with an average of 7.46 times per person (SD 6.47). ICCs showed that 40%-60% of variability in response to each item was attributable to stable, between-person differences (
Variability estimates for mobile social comparison response measure and relations with traditional self-report measure among individuals with schizophrenia (n=24).
Response to comparison | Variability estimate (intraclass correlation coefficient) | Relation with time 1 score ( |
Relation with time 2 score ( |
Upward identification | 0.40 | 0.38a | 0.24 |
Upward contrast | 0.60 | 0.53b | 0.33 |
Downward identification | 0.41 | 0.40a | 0.76c |
Downward contrast | 0.57 | 0.50b | 0.74c |
a
b
c
Individuals with schizophrenia experience meaningful deficits in social integration and perception that may be targeted with digital interventions, though patient engagement with these interventions is modest. The opportunity to make social comparisons may help to address these problems, though this concept has received little attention in schizophrenia. As an initial step, the results of this study provide necessary, if preliminary, insight into this process at multiple levels. The limited existing work on social comparisons among individuals with schizophrenia focused on the use of upward versus downward comparisons [
Specifically, this study captured reports of affective responses to upward and downward social comparisons (as indicators of identification and contrast processes), which are better longitudinal predictors of clinical outcomes among individuals with chronic medical conditions than the reported direction of comparisons [
Further, although most studies of social comparison focus on stable differences between people [
Importantly, findings from this study also provide preliminary support for the feasibility of collecting real-time data on social comparison responses through digital tools such as apps, and suggest the potential for these data to inform the tailoring of digital interventions for schizophrenia. For example, although there were considerable between-person differences in the number of social comparison smartphone assessments completed (and considerable variability in item responses), smartphone assessments showed 3 important features. These assessments were voluntarily completed throughout the assessment period; they captured both between- and within-person variability in affective responses to comparisons; and responses to mobile items correlated with those completed with traditional self-reports from the original measure. Thus, a brief, smartphone-based assessment of social comparison responses appears to perform well for its intended purpose, and additional work is needed to confirm and extend these findings.
Overall, the observed variability in affective response to comparisons among patients with schizophrenia suggests that there are times when negative (and positive) affective responses are stronger than others. In future studies of this kind, smartphone-based assessment may enable modeling of moderators of social comparison response, such as comparison dimension (ie, what about the self is being compared), mode of comparison (ie, face-to-face vs. via social media), or motivation for comparison (ie, self-selected from a range of options for a particular purpose, or in response to exposure to a single target) [
In the current era of socially connected digital health tools, where patients with schizophrenia engage at rates equal to the general population [
Strengths of this study include its recruitment of both individuals with schizophrenia and healthy controls, both with equal proportions of men and women, and the use and comparison of both traditional self-report measures and brief versions modified for mobile assessment. Further, the emphasis of this study was on differentiating between- and within-person variability in a critical but understudied aspect of social comparison (ie, affective response), using appropriately sophisticated statistical methods.
As this was a formative pilot study, however, there were noteworthy limitations. Our sample sizes were modest, particularly at time 2, and participants were predominantly White or Asian American. We also did not have the opportunity to include a clinical control group. Given that participants had flexibility in their completion of mobile assessments, compliance with these assessments was inconsistent across participants. Modest compliance with mobile assessments is common among individuals with schizophrenia and other severe and persistent mental illnesses [
In addition, despite reviewing instructions with participants at orientation to specify the time window they should use for reference when completing the mobile version of the Identification-Contrast Scale, it is possible that participants with schizophrenia responded with more global than contextually sensitive impressions of their affective responses. The considerable within-person variability observed in their responses suggests that the measure was sensitive to context, but future studies should consider adding more specific instructions to the mobile version of the measure.
Finally, given the complexity of social comparison and the emphasis on general affective responses in this study, assessments also did not capture all of the aspects of this process that may be relevant to its role in daily life. For example, measures used in this study did not assess individual instances of social comparison, and thus, did not capture the dimension or mode [
intraclass correlation coefficient
This work was supported by the US National Institute of Mental Health under grant number K23 MH116130 (PI: JT) and the US National Heart, Lung, and Blood Institute under grant number K23 HL136657 (PI: DA).
Data are available by request to the second author.
None declared.