Abstract
This study shows that the current form of the Turkish Sensory Responsiveness Questionnaire can contribute to understanding the sensory processing abilities of Turkish people with and without mental disorders.
According to a report from the World Health Organization (2022), the prevalence of mental disorders is around 13% worldwide. According to the Turkey Mental Health Profile Study, conducted in 1998, the cumulative prevalence of mental disorders in adults was reported to be 18% in Turkey (Erol et al., 1998). A recent study reported that this prevalence has increased to 43% (Keskin et al., 2013). Therefore, early diagnosis, management, and treatment of early psychosis and mental disorders are crucial, worldwide and in Turkey.
In the literature, many people with mental disorders have reported difficulties in sensory processing (Brown et al., 2020; Halperin & Falk-Kessler, 2020; Pfeiffer et al., 2014). People with mental disorders showed elevated levels of low registration, sensory sensitivity, and sensory avoidance and lowered sensory-seeking behavior (van den Boogert et al., 2022). Moreover, the relationship between mental disorders symptoms and sensory processing difficulties is frequently mentioned. For instance, Serafini, Gonda, et al. (2017) found that sensory sensitivity and avoidance are associated with depression, impulsivity, and alexithymia in people with depression. Brown et al. (2002) stated that the appeal of predictable and low-demand environments and negative symptoms characterizes sensation avoidance in schizophrenia. King (1974) and Horváth et al. (2019) emphasized that the problem in proprioceptive processing of people with schizophrenia may cause perceptual problems, problems with body image, and motor coordination disorder.
Interventions that are based on sensory integration theory as a consumer-centered humane approach for people with mental disorders is regarded mainly as the profession of occupational therapy’s response (Champagne & Koomar, 2012; Champagne & Stromberg, 2004). Occupational therapists have developed several instruments to detect sensory processing difficulties in people of different ages. Many instruments have been developed for the adult population (Kinnealey & Oliver, 2002; Kinnealey et al., 1995; Blanche et al., 2014 ; Brown and Dunn, 2002;). The two most widely tested tools used to identify sensory processing difficulties are the Adolescent/Adult Sensory Profile (AASP; Brown et al., 2001) and the Sensory Responsiveness Questionnaire (SRQ; Bar-Shalita et al., 2009; Machingura et al., 2018). The SRQ is based on the work by Bar-Shalita et al. (2009), whereas the AASP is theoretically based on Dunn’s (1997) model of sensory processing. AASP has Turkish reliability and validity (Üçgül et al., 2017). However, the Turkish AASP is only culturally adapted, reliable, and valid for healthy people and patients with schizophrenia. Moreover, a license from the Pearson Corporation is required to use the AASP. The SRQ is an easy-to-use and free questionnaire. There is a lack of Turkish-language measurement tools to evaluate sensory processing disorders, and currently there are limitations in Turkish-language tools to assess sensory modulation disorder in people with mental disorders. Therefore, we aimed to adapt the SRQ into Turkish and evaluate its cultural adaptation, reliability, and validity in participants with and without mental disorders.
The SRQ is used to measure the severity of sensory processing difficulties. The SRQ assesses the intensity and frequency of responsiveness to stimuli in several sensory modalities. This self-report questionnaire presents a set of 58 typical scenarios encountered throughout daily life. Each scenario involves one sensory stimulus in one modality, including auditory, visual, gustatory, olfactory, vestibular, and somatosensory stimuli (excluding pain). The items are presented in a manner that attributes a hedonic or aversive valence to the situation. The participant ranks the intensity of pleasure (associated with sensory-seeking behaviors or sensory underresponsiveness) or aversion (associated with sensory overresponsiveness) that these daily sensory experiences cause and the frequency of their occurrence. Two parallel rating measures are used, one for the perceived intensity of the response to the sensory stimuli and the other for the frequency of their occurrence. Participants rate the intensity of the hedonic or aversive responses to the situation on a 5-point Likert scale ranging from 1 (not at all) to 5 (very much). The frequency of response is also rated on a 5-point Likert scale ranging from 1 (never) to 5 (always). The interaction between the two questionnaires (Intensity × Frequency) represents the combined SRQ score (Bar-Shalita et al., 2009).
The internal consistency and reliability of the different SRQ questionnaires for intensity (Cronbach’s αs =.90–.93; rs = −.71–.84, respectively) and frequency (Cronbach’s αs = .93–.98; rs = .93–.94, respectively) are high. The SRQ has further been administered twice to 24 adults 2 weeks apart. The correlations for both the intensity (r = .71, p < .001) and frequency (r = .74, p < .001) at the two timepoints were high, indicating the stability of SRQ over time (Bar-Shalita et al., 2009, 2012).
Method
Ethical Considerations
Ethics approvals were obtained from the ethics commission of the Hacettepe University Non- Interventional Clinical Researches Ethics Board (14.05.2019-GO19/352). The research conforms to the provisions of the Declaration of Helsinki. Written informed consent was obtained from the participants.
Participants
This study was conducted at Hacettepe University Faculty of Health Sciences and Ankara University Faculty of Medicine Department of Mental Health and Disorders.
The adequacy of the sample size was measured with the Kaiser-Meyer-Olkin test. The inclusion criteria of the participants with a confirmed mental disorder (schizophrenia, bipolar disorder, or depression) were being age 18 to 60 yr, being able to understand (spoken or written) Turkish, and having a desire to participate. Participants who were in the flare-up or acute episode of the disease, with mental retardation or having any treatment-resistant physical or neurological disease, were excluded from the study. The inclusion criterion for participants without mental disorders was being age 18 to 60 yr. Exclusion criteria included the presence of severe systemic chronic diseases (e.g., cancer), nervous system impairments (e.g., stroke or epilepsy), and daily medication use. All participants were recruited for participation in this study in a face-to-face interview.
Six hundred people were invited to each group. Among the people with mental disorders who were invited to the study, 10 were excluded because they refused to participate in the study, 50 were excluded because they did not meet the age criteria, and 11 were excluded because they had a chronic disease. The response rate for participants with mental disorders was 88.1%. One hundred fifty people who were invited to join the healthy group did not want to participate in the study, and 66 did not meet the age criteria. The response rate for healthy participants was 64%. A total of 1,200 people were called to participate in the study, and 287 were rejected.
Of the 913 participants, 274 were randomly selected to be reinterviewed, and they completed the questionnaire once more. A total of 529 participants ages 18 to 60 yr with mental disorders (schizophrenia, n = 149; bipolar disorder, n = 155; depression n = 225) were included in the study. A total of 384 participants without mental disorders, ages 18 to 60 yr, were involved in the study as a comparison group.
Data Collection
During the face-to-face interview, participants completed the paper-based sociodemographic form, the SRQ, and the AASP. The sociodemographic form contained items regarding age and gender.
We demonstrated the concurrent validity of the SRQ using the AASP as an external criterion (Bar-Shalita et al., 2009). Therefore, the external validity of the SRQ was examined with the AASP. The AASP was developed and built on Dunn’s sensory processing model and is suitable for those ages ≥11 yr (Brown & Dunn, 2002). It consists of 60 items and is used to evaluate sensory processing abilities in daily life. It evaluates the responses to sensory input through six subsections that refer to everyday activities, including taste and smell, movement, visual, touch, auditory, and activity level. The participants’ responses to this questionnaire are evaluated in four quadrants: low registration, sensory sensitivity, sensory avoiding, and sensory seeking. Low registration refers to behaviors such as missing stimuli or taking longer than other people to respond to them. Sensory sensitivity is defined as being easily distracted from stimuli, feeling discomfort with sensation, and responding more than normal to stimuli but not actively limiting exposure to uncomfortable sensations. Sensory avoiding is defined as actively limiting exposure to sensations and avoiding distracting situations, and sensory seeking refers to the tendency to create additional stimuli or look for an environment that provides sensory stimuli (Dunn, 1997). Responses are provided on a Likert scale system on which 5 = almost always, 4 = often, 3 = sometimes, 2 = rarely, and 1 = almost never, and the scores of the related items are summed up to provide the score for each quadrant. Üçgül et al. (2017) completed the tests to settle the validity and reliability of the Turkish AASP. The Turkish AASP has high internal consistency and test–retest reliability (rs = .66–.82 and .67–.82, respectively; Üçgül et al., 2017).
Translation and Cultural Adaptation Process
We performed cross-cultural adaptation of the SRQ by following the guidelines provided by Beaton et al. (2000). Permission to translate and adapt the SRQ was obtained from Tami Bar-Shalita. The SRQ was translated into Turkish by a committee of three therapists specializing in sensory integration and mental health. Both semantic equivalence and cultural relevance were considered to ensure conceptual equivalence in the Turkish translation. The translator, with no knowledge of the SRQ, then back-translated the Turkish translation into English. The committee then compared the original and back-translations, discussed discrepancies, and revised the Turkish translation. This process was iterated until we agreed on a satisfactory Turkish translation.
Piloting Process
The translated questionnaire was then used for subsequent cognitive debriefing interviews with 120 participants (30 participants in each group) to assess further the clarity, comprehension, cultural relevance, and appropriateness of expression, words, or phrases. On the basis of the participants’ input, the committee modified the Turkish translation, which was then used to evaluate its psychometric properties in the reliability and validity process of the study. For example, we changed the first translation term ocean in Item 25 to sea. In addition, although Item 41 (shaving bothers me) was expressed only for men in the original questionnaire, it was considered as an item suitable for both genders in line with the returns received from the pilot study. To concretize the answer in line with the feedback received from participants with mental disorders, the percentile value was added in consultation with Tami Bar-Shalita.
Reliability and Validity Process
The statistical analysis was performed using IBM SPSS Statistics (Version 23) and AMOS 24.0. The Shapiro– Wilks test was used to evaluate the distribution of the collected data (normal, p > .05; Razali & Wah, 2011). Descriptive analysis was used to describe questionnaire results. Numerical values are represented as mean, standard deviation, minimum and maximum; categorical data are represented as frequencies.
We used confirmatory factor analysis (CFA) to measure the construct validity of the instrument. Model fit indices were used to determine the best suitable analysis to model the dataset and theoretical model in CFA.
Reliability was determined using internal consistency and test–retest reliability. Cronbach’s α internal consistency coefficient evaluated internal consistency for each factor. Cronbach’s α coefficient reliability increases as it approaches 1 and decreases as it approaches 0. The Cronbach’s α values were interpreted as <.40, not reliable; .40 to .60, low reliability; .60 to .80, moderate reliability; and >.80, high reliability (Alpar, 2013). Test–retest reliability was assessed by comparing responses for the first and second administrations of the SRQ using the intraclass correlation coefficient (ICC; Koo & Li, 2016). ICC rates were accepted as poor (<.40), moderate (.40–.60), good (.60–.75), and excellent (.75–1.00; Hayran & Hayran, 2011).
A total of 274 participants completed the SRQ again 2 weeks later for test–retest reliability. The Spearman correlation coefficient was used for test–retest reliability. The levels of significance according to Spearman’s correlation coefficient were determined as r = .75–1.00, excellent; r = .70–0.75, very strong; r = .60–.70, strong; r = .40–.60, moderate; r = .30–.40, lower moderate; and r = .05–.30, weak or an insignificant relationship (Hayran & Hayran, 2011). In examining the difference between categories in quantitative variables, we used the Mann–Whitney U and Kruskal–Wallis H tests for variables with two categories and more than two categories, respectively. A significance level of .05 was used for all statistical tests.
Results
A total of 529 participants with mental disorders (schizophrenia, n = 149; bipolar disorder, n = 155; depression, n = 225) and 384 participants without mental disorders, ages 18 to 60, were included in the study. The baseline characteristics of participants are presented in Table 1. The completion time of the questionnaire was between 30 and 45 min for participants with mental disorders and 15 to 30 min for participants without mental disorders.
Sociodemographic Characteristics of Participants
Construct Validity
Confirmatory Factor Analysis
Factor loadings should not be below .30 (Harrington, 2009). Fifty-eight items did not provide structural validity with the factor analysis method. Therefore, 22 items were removed. We used the following steps while removing the items from the questionnaire: ▪ consultation with Tami Bar-Shalita ▪ consultation with a psychiatrist working with mental disorders and an occupational therapist working in the field of sensory integration ▪ rating the items in order of clinical importance ▪ elimination of the most clinically insignificant and interchangeable items according to experts.
Because the factor loadings of 10 items (Items 3, 13, 16, 17, 27, 35, 43, 47, 53, and 58) in the hedonic subfactor and 12 items (Items 2, 5, 6, 7, 11, 18, 21, 23, 26, 36, 40, and 54) in the avoidance subfactor in the original version were below .30, these items were removed from the questionnaire. The factor loading of the remaining 36 items are presented in Table 2.
Confirmatory Factor Analysis Factor Loadings for the Sensory Responsiveness Questionnaire, Turkish and English Versions
Note. SRQ = Sensory Responsiveness Questionnaire.
aItems removed from the Turkish version of the SRQ after confirmatory factor analysis.
The goodness-of-fit indices was used to analyze the construct validity of the two-factored structured questionnaire. The questionnaire’s χ2/df ratio shows good agreement with 4.200. The goodness-of-fit indices showed acceptable fit: goodness-of-fit index = .921, comparative fit index = .910, and root-mean-square error of approximation = .063. We examined the convergent validity of the questionnaire by evaluating the correlations between the SRQ subfactors and AASP quadrants. Correlations between the SRQ subfactors and the AASP quadrant scores are shown in Table 3.
Correlations Between the Subfactors of the Turkish Version of the Sensory Responsiveness Questionnaire and the Quadrants of the Turkish Version of the Adolescent/Adult Sensory Profile
Note. Correlations are expressed as Spearman correlation coefficient (r).
*p < .05.
As can be seen in Table 3, a positive moderate and statistically significant correlation was found between the avoidance subfactor of the SRQ and the low registration quadrant of the AASP (r = .411, p < .001). A positive lower moderate statistically significant relationship was found between the hedonic subfactor of the SRQ and the sensory seeking quadrant of the AASP (r = .302, p < .001). Moreover, there was also a positive lower moderate statistically significant relationship between the avoidance subfactor of SRQ and the sensory sensitivity and sensory avoiding quadrants of AASP (rs = .398 and .353, respectively). A statistically significant weak relationship was found between the hedonic subfactor of the SRQ and the low registration quadrant of the AASP (r = .291, p < .001). Furthermore, a negative, weak, and statistically significant correlation was found between the avoidance subfactor of the SRQ and the sensory seeking quadrant of the AASP (r = −.099, p < .05).
Testing Construct Validity
The score distribution of hedonic and avoidance subfactors according to sociodemographic information is presented in Table 4. Regarding hedonic and avoidance subfactors, a statistically significant difference was found between participants without mental disorders and those with schizophrenia, depression, and bipolar disorder. Scores of both subfactors were higher in groups of participants with a mental disorder. In terms of gender variables, males had higher scores than females on both subfactors. There was no statistically significant difference in the hedonic subfactor score regarding the education level variable between the groups. As the education level increased, sensory avoidance decreased in the avoidance subfactor score. In terms of the marital status variable, married participants had high scores on both subscales. Regarding the working status variable, retirees had higher scores on the hedonic subfactor and part-time workers had higher scores in the avoidance subfactor. There was no significant difference on the hedonic subscale in terms of the family income variable, whereas the avoidance subfactor score decreased as the family income increased (Table 4).
Comparison of Scores on the Turkish Version of the SRQ According to Sociodemographic Information
Note. N = 913. SRQ = Sensory Responsiveness Questionnaire.
aMann–Whitney U test.
bKruskal–Wallis H test.
*p < .05. **p < .01.
Reliability Analysis
Internal Consistency
Cronbach’s α was determined for each subfactor. Internal consistency was high for the hedonic subfactor (.812) and the avoidance subfactor (.815), as well as the total score (.837).
Test–Retest Reliability
Test–retest reliability of the SRQ was excellent for the hedonic and avoidance subfactors and the total score (ICCs = .973, .988 and .974, respectively).
Discussion
This study was carried out to establish the validity and reliability of the Turkish version of the SRQ. The sample in this study consisted of 913 participants, of whom 529 were participants with mental disorders and 384 were participants without mental disorders. Our study indicated that the Turkish version of the 36-item SRQ is reliable and valid for evaluating the sensory processing abilities of Turkish people ages 18–60 yr with and without mental disorders. This validation study was essential to provide a standardized, validated tool available for therapists to identify sensory problems, plan specific intervention programs, and conduct research in the Turkish population.
The cross-cultural adaptation process produced minor challenges. Some modifications were made to some of the items to ensure semantic, experiential, and conceptual equivalence of the terms. In Item 25, the term ocean was replaced with sea. Moreover, Item 41 (“shaving bothers me”) was suitable for both genders in the Turkish population. The percentile value was added in consultation with the questionnaire developer. This addition follows the suggestion by Beaton et al. (2000) that experiential equivalence should be ensured between the original and target languages during cross-cultural adaptation. I participants were informed that the questionnaire evaluated the intensity and frequency of sensory processing abilities. However, they tended to focus on the behavioral responses to the items rather than on the intensity and frequency.
Confirmatory factor analysis was performed to test the construct validity. Two subfactors defined by Bar-Shalita et al. (2012) in the original questionnaire were also supported for all groups in the present study. Harrington (2009) highlighted that the factor loading should be at least .30. Confirmatory factor analysis did not approve the structure of the Turkish SRQ with 58 items. Therefore, 22 items were excluded from the Turkish SRQ. The English version of the SRQ has 26 items in the hedonic subfactor and 32 items in the avoidance subfactor, whereas the Turkish version of the SRQ has 16 items in the hedonic subfactor and 20 items in the avoidance subfactor. After the analysis, the Turkish version of the SRQ was found to be valid and reliable with 36 items.
The items that were removed from the Turkish SRQ evaluated similar sensory-processing skills from different perspectives. For instance, the 11th item removed from the Turkish SRQ— “Arka fondan gelen uğultu sesleri beni rahatsız eder (örneğin; klima, buzdolabı, bilgisayar fanı gibi)” (“I am bothered by background humming noises [e.g., air conditioner, refrigerator, computer fan, etc.])”—and the 12th item remaining on the Turkish SRQ—“Arka fondan gelen sesle çalışmak ve konsantre olmak hoşuma gider” (“I enjoy working and concentrating with background noises”)—were related to auditory stimuli from different perspectives. In the validity study of the AASP conducted in Israel by Engel-Yeger (2012), items that were not statistically and culturally significant and had a low factor load, as in this study, were excluded from the Israeli version of the AASP.
The scales’ Cronbach’s α values should be >.40 to be accepted as reliable (Alpar, 2013). Cronbach’s α values for the Turkish and English versions of the questionnaire were .812 and .90 for the hedonic subfactor and .815 and .93 for the avoidance subfactor. Similar to Bar-Shalita et al. (2009), we found positive statistically significant moderate- and high-level differences between the test–retest applications that were conducted to assess time variance at 2-wk intervals; ICCs = .973 (hedonic) and .988 (avoidance). This finding indicates that the Turkish version of the SRQ has high reliability, and the internal consistency and test–retest reliability are comparable with those of the English version.
We conducted a correlation analysis between the SRQ subfactors and AASP quadrants to analyze the convergent validity of the questionnaire. In the present study, we found correlations between the hedonic subfactor of SRQ and the low registration and sensory seeking quadrants of AASP. There are also correlations between the avoidance subfactor of SRQ and sensory sensitivity and sensory avoiding quadrants of AASP. These findings are consistent with those of the study conducted by Bar-Shalita et al. (2009). On the other hand, we also found a correlation between the avoidance subfactor of the SRQ and the low registration and sensory seeking quadrants of the AASP. Bahadir et al. (2022) found that the low registration and sensory seeking quadrants of AASP were correlated with the overresponsiveness to vestibular input subfactor of the Turkish Adult Sensory Processing Scale. Bahadir et al. (2022) also suggested that the factors affecting the sensory responses of Turkish people should be investigated in detail and that reference values in Turkey should be defined.
We assessed the construct validity of the questionnaire by hypothesizing distinguishing group differences between participants with and without mental disorders. The participants with mental disorders had higher scores on the subfactors than those without mental disorders. Similarly, Brown et al. (2002) compared the sensory modulation skills of people with schizophrenia, bipolar disorder, and other mental disorders with the skills of people without mental disorders. Likewise, they concluded that the group with mental disorders scored higher than the participants without mental disorders. Therefore, we concluded that the SRQ is a specific tool for detecting sensory processing difficulties in people with and without mental disorders.
To our knowledge, no study exists on how sociodemographic variables affect sensory processing skills in people with mental disorders. Related studies have stated limitations regarding this issue and have emphasized that a study should be conducted to examine this (Bahadir et al., 2022; Serafini, Engel-Yeger, et al., 2017). It was emphasized that, in healthy adults, sociodemographic variables affect sensory processing skills (Engel-Yeger, 2012; Engel-Yeger & Dunn, 2011; Hebert, 2015; Machingura et al., 2020). In our study, sensory processing data was analyzed for all participants, not only those with mental disorders. For this reason, we cannot comment on how sociodemographic variables affect the sensory processing skills of participants with mental disorders. We suggest that future studies should examine how sensory processing skills are affected by sociodemographic variables, especially in people with mental disorders.
Furthermore, although the completion time of the questionnaire was between 30 and 45 min in participants with a mental disorder, it took 15–30 min in participants without mental disorders. Slow reaction time, constantly changing answers because of erroneous information, or thinking long may cause this difference (Brown et al., 2002; Mino, 2004; Üçgül et al., 2017).
Because this study only included people with or without a mental disorder at Ankara University and people without a mental disorder at Hacettepe University, the sample group was limited. Sensory processing abilities can vary according to sociodemographic information such as gender, age, education level, marital status, occupation, and family income for people without a mental disorder or certain diseases such as schizophrenia, depression, and bipolar disorder for people who had a mental disorder. Conducting validity and reliability studies of the Turkish SRQ in different mental disorders may help expand the scope of applicability of the questionnaire.
Implications for Occupational Therapy Practice
The results of this study have the following implications for occupational therapy practice: ▪ Occupational therapists can use the SRQ, a 36-item measurement tool with a 5-point response scale, to measure the severity of sensory processing difficulties. ▪ The SRQ is a standardized, reliable, and valid scale. ▪ Occupational therapists can use the SRQ with people with and without mental disorders. Future research is necessary to explore further the psychometric properties of the SRQ for use in occupational therapy practice and research.
Conclusion
The Turkish version of the 36-item SRQ is a reliable and valid assessment tool in Turkish and can be used by therapists who specialize in mental health services. The current form of the Turkish SRQ can contribute to understanding sensory processing abilities in people with and without mental disorders.
Footnotes
Acknowledgments
We thank the participants in this study. This article is based on the Recep Yükçü’s master’s degree dissertation.