Data on demographic characteristics, such as age, gender, occupation and employment status, were collected at the start of the study. In addition to the TSRQ (Diet and Exercise), we assessed the stage of behavior change, which is based on the transtheoretical model (TTM), to measure the convergent validity of the scale.
The Treatment Self-Regulation Questionnaire (TSRQ)
The TSRQ was used to measure participants’ motivation to maintain diet and exercise-related behaviors. According to the Center for Self-Determination Theory (CSDT), the original version of the scale consists of 15 items each on diet and exercise, and each domain additionally includes four subscales (autonomous motivation, introjected regulation , external regulation and amotivation) . All items are rated on a 7-point Likert scale ranging from 1 (not at all true) to 7 (very true). With the exception of items 5, 10 and 15, the higher the score, the higher the autonomous motivation. Existing research suggests that the validity of the TSRQ and the internal consistency of each subscale are adequate (most α values > 0.73) .
Translation of the TSRQ in Japanese
To translate the scale into Japanese, we first obtained permission from the CSDT to use the TSRQ. A licensed Japanese physician, who was a native Japanese speaker and fluent in English, and also familiar with Japanese and Western healthcare systems, translated the scale into Japanese. Consistency between the Japanese and English versions of the scale was ensured by (1) using simple sentences, (2) using nouns rather than pronouns, (3) avoiding metaphors and colloquial phrases, (4) ) avoiding passive expressions and (5) avoiding hypothetical expressions . In addition, exchanges took place between the researchers, public health nurses and doctors, to verify whether the wording of an item was appropriate to the field of health orientation and whether the participants could understand the item; corrections have been made where necessary. Back-translation into English was performed by a bilingual Japanese expert, and the CSDT confirmed the conceptual integrity of the translated version of the scale by reviewing the items.
Stages of behavior change
The TTM, which is the theory behind the stages of behavior change, was developed in the 1980s . It was introduced to Japan in the late 1990s when the country began to focus on measures to prevent and manage lifestyle-related illnesses. . Since 2000, studies have applied the TTM to Japanese individuals. The theory is widely used and the Japanese Ministry of Health, Labor and Welfare also recommends using the behavior change stages in health counseling. . The Stages of Change model posits that individuals go through five stages of behavior change: pre-contemplation, contemplation, preparation, action, and maintenance. Therefore, we asked participants to fit their health behavior to one of five stages through the following questions: “I have no intention of acting at all” “I plan to act in the future ” “Sometimes I act” “In the 6 months since I played” “Over 6 months since I played”. The Stages of Behavior Change Scale used in Japan has been verified: Cronbach’s alpha coefficients for diet items are 0.74 and its reliability and validity have been confirmed. . Research on exercise elements has also been reported .
We calculated Cronbach’s alpha for internal consistency, item-total correlation to examine reliability, performed correlational analyzes to test convergent validity, and performed confirmatory factor analysis for structural validity. SPSS version 24 was used for each analysis.
According to the COSMIN criteria, the sample size for any internal consistency analysis is considered “good” if it is five times as many items and greater than 100. Since the Diet and Exercise questionnaires in this study together consist of 30 items, the minimum sample size required was 150. Therefore, the sample size in this study was sufficient and met the COSMIN criteria.
From previous studies  confirmed that the TSRQ has a four-factor structure (autonomic motivation, introjected regulation, external regulation, and amotivation), the total score on the Japanese version of the TSRQ and Cronbach’s alpha for each factor were calculated to assess internal consistency . In addition, item-total correlations (hereafter referred to as “computer correlations”) were calculated to examine reliability. In the Japanese version of the TSRQ, items 5, 10, and 15 measure lack of motivation and were scored backwards. After performing the computational correlation, unreliable and unsuitable items were excluded.
Convergent validity was assessed by calculating Pearson’s correlation coefficients between TSRQ and stage of behavior change. The effect size detected in this study was 0.3 . Sample size was calculated using G*Power 3.1. For an alpha error of 0.05 and a power of 0.8, it was estimated that a minimum of 352 participants would be needed. Therefore, the sample size for this study was sufficient and met the COSMIN criteria.
A confirmatory factor analysis (CFA) was performed to assess structural validity. Based on previous studies, a four-factor model was assumed . The COSMIN criterion for minimum sample size for factor analysis was met. The maximum likelihood estimation method was used, with the value of the chi-square (χ2), the goodness of fit of the comparative fit index (CFI) and the root mean square error of approximation (RMSEA). Goodness-of-fit and RMSEA thresholds were 0.90 or greater and 0.08 or less, respectively .