1 Introduction

Happiness is considered as a positive emotion that influences one’s behaviour and feelings [1]. However, it is an ambiguous and complex concept; which has constituted, if not the ultimate, a very important goal for philosophers, thinkers, researches, ordinary people of all epochs, ages, gender, and races. The oldest approach comes from ancient Greece where, initially, it is referred to as ‘hedonism’ with the main pursuit of pleasure, bodily pleasure, and avoidance of pain. Socrates states that happiness is achieved by having ‘Know Thyself’ self -knowledge [2]. Plato presents well-being or happiness as the most noteworthy goal of moral thought and behaviour [3]; while for Aristotle the search for happiness is summed up in ‘eudaimonia’, the activity of the soul which is driven by virtue [4].

In current psychological research regarding happiness, various terms have been dealt with, usually found as well-being, a state characterized by feelings of pleasure [5]. Particularly, Seligman defined the three ways of well-being: pleasure, engagement, and meaning [6]. A life of ‘’pleasure’’ is highly related to hedonism, by means of pleasure maximization and pain minimization at the same time [7]. Regarding the second way, a life of ‘’engagement’’, it means the absorption in engaging activities and experiencing flow [8], while a life of ‘’meaning’’ is related to the Aristotelian idea of ‘’eudaimonia’’. According to [9], the aforementioned approaches of well-being are distinguishable but not incompatible [10]. Apart from the well-being state, the World Database of Happiness mentions that the term of happiness acts as an umbrella for all good and defines it as the degree to which an individual experiences overall life satisfaction [11]. Specifically, a previous study considers that synonyms of happiness are not only the ‘subjective well-being’ but also the ‘life-satisfaction’ [12]. Nevertheless, in a recent meta-analysis [13], note that the terms happiness, life satisfaction and well-being differ; where happiness is related to emotional state and the description of momentarily pleasant feelings, while life satisfaction to cognitive evaluation of life, and well-being is related to overall quality of life. In addition to that, an examination of World Values Surveys from 1981 to 2005 found that happiness and satisfaction were moderate correlated, indicating that the two notions, while connected, actually differ.

The latest effort to understand the nature of happiness and well-being comes from the field of Positive Psychology, as a person feels happy when his/her life, relative to the standards he/she has chosen, is going well [4]. According to Seligman, well-being can be taught, defined, and measured, while greater well-being also depends on optimism [6]. Additionally, a person's moods and emotions are influenced by culture; thus, happiness can appear in diverse forms in different cultures, and these emotional experiences can affect one's degree and impression of happiness [14, 15], so the degree of individuality in a culture influence how we comprehend and experience wellbeing [16]. Furthermore, to explain the type and degree of happiness, the nation’s features are also crucial [17].

1.1 Existing measures of happiness

There are many tools that measure happiness. Up to now, qualitative and quantitative methods have been used to evaluate happiness [18], while the self-report measures are also commonly used [19]. Scales that measure happiness vary from single-questions to multi-item scales, either in unidimensional or multidimensional structures.

Despite the fact that many current measures have been evaluated, there is still a lingering question over the precise structures or components that best characterize happiness. To this direction, most researchers have focused on identifying several aspects or factors that lead to an individual’s happiness [20], but there are also scales, such as a singular self-reporting item assessing individual happiness [21]. Nevertheless, a single item scale has several limitations in its use. Particularly, it may lack the potential for internal consistency reliability assessment or factorial analysis implementation, rendering it inadequate for factor analysis requirements [20]. Some authors argue that single-item measures tend to result in diminished predictive capability, leading easily to Type II errors [22], and are presenting inferior predictive validity, construct validity, and criterion validity, limiting their application in scientific research [22, 23].

Regarding the unidimensional tools, the Subjective Happiness Scale (SHS) is a widely used scale consisting of 4-items on a 7-point Likert [24]. SHS measures global subjective happiness. It was validated in general population showing excellent psychometric properties, while its construct validity correlates highly with measures of happiness and well-being [25]. It is translated and adapted in a Greek adults’ sample, aged 18–79, indicating satisfactory reliability and validity [26, 27]. Moreover, the Authentic Happiness Inventory is another unidimensional tool, (AHI), which in an earlier version is referred as Steen Happiness Index—20items) measures the subjective assessment of happiness describing individuals’ feelings in the past week [28]. It comprises 24 sets of five statements as possible answers ranging from a negative to an extreme positive, from which the person has to choose the statement that best describes his/her feelings. Its use is basically made in the Intervention of Positive Psychology (PPI) properties and is also used in a broad range of other fields showing satisfactory psychometric properties, such as high internal consistency. Although no factorial validity or test–retest reliability have been provided by the developers, further research revealed that the AHI is adequately represented by a one-dimensional model, stated as a reliable and valid tool in its construct validity and internal reliability (discriminant as well convergence validity, test- retest correlations) [29]; additionally, more evidence has shown that it might be utilized trustworthily as a single-item scale, cause question 24 is indicated as a global happiness item [30]. Happiness measure (HM) based on [31], extended and refined by Fordyce [32, 33] is another tool consisting of one self-reporting item: “In general, how happy or unhappy do you usually feel?”, measuring two dimensions of happiness: quality (intensity) and quantity (frequency). The first dimension is a statement on an 11-point Likert unhappy/happy scale and in the second dimension individuals have to fill the percentage of time spending on feeling happy/unhappy/neutral. A review if 15 years evaluation on HM indicates good reliability, exceptional stability, and an extremely broad evidence base which supports the convergent, construct, and discriminative validity [34].

On the other hand, the majority of instruments are multidimensional, as happiness is considered to be multidimensional concept by many researchers [35]. A commonly used scale is Positive and Negative Affect Scale (PANAS), consisting of two 10-item sub-scales which examine on a 5-point Likert the two primary dimensions of mood: Positive Affect (PA) and Negative Affect (NA) states; reporting high internal reliability, stability, convergent and discriminant validity (students and general population) [36]. The validation of PANAS scale has been confirmed in multiple languages, showing excellent internal consistency and discriminant validity [37]. In Greece this scale was adapted indicating satisfactory validity and satisfactory internal reliability in adolescent and young adult population [38]. Orientation to Happiness (OTH) is a combined tool which consists of 18-items in total, having three sub-scales (factors) of life of engagement (6-items), meaning (6-items), and pleasure (6-items) measuring in a 5-point scale the degree to which the item applied (from “very much unlike me’’ to “very much like me’’) [9]. Several studies in undergraduate, university students and adult workers indicate OTH as a reliable and valid measure regarding its psychometric properties with a satisfactory internal reliability [39,40,41,42,43]. A newly developed test is Measure of Happiness (MH), a 14-item on a 10-point Likert which contains 5 factors: Psychophysics Status (3-item), Financial Status (3-item), Relational Private Sphere (3-item), Socio-Relational Sphere (2-item), and Life Perspective (3-item) [25]. It is a multidimensional scale which provides satisfactory psychometric properties in adults of general population, aged 22 to 84, with adequate interval consistency, and excellent construct validity [25].

One of the most widely used scales for personal happiness is Oxford Happiness Inventory (OHI) [44, 45]. OHI is a 29-items measure suitable for assessing happiness, consisting of seven factors: well-being, self-esteem, satisfaction with life, cheerfulness, positive outlook, efficacy, empathy-sociability. Each of 29 items was represented into 4 incremental levels (0 to 3) (4 response options demonstrating different levels in aspects of happiness). The tool showed satisfactory internal consistency and validity, allowing for endorsement over an extended range [45, 46]. It is already adapted in Greek adults, private sector workers, with an average age of 41.57 years, from various educational levels, and presents satisfactory reliability and validity [47]. Despite the fact that the OHI had shown good psychometrics properties, this measure had one notable drawback, it was a lengthy questionnaire due to its unique structure with 4 incremental levels for each of 29 items [48]. To resolve the above issue, the authors modified the measurement method to a 6-point Likert scale to enable respondents to choose from a wider range improving the statistical properties of the individual items [45]. This alternative scale, the Oxford Happiness Questionnaire (OHQ) contains a similar number of items (29-items) of similar wording to the OHI with the same amount of positive and negative items. OHQ indicates construct validity, and high scales and item reliability. It is noteworthy that the OHQ shows a stronger correlation with personality variables than the OHI; this leads to the conclusion that it is probably the preferred tool for measuring happiness [45]. Additionally, it constitutes a more comprehensive tool than other scales and less prone to respondent bias [49]. A shorter version of the OHQ, the Oxford Happiness Questionnaire Short Form (OHQ-SF) was created from the existing OHQ, containing items of similar wording [45]. It is a unidimensional, 8-item on a 6-point Likert scale, compact, easy to administer tool that respectively allows endorsements across a wide range which can be employed in circumstances where in temporal constraints are present. The correlation among the full and short scales results was highly significant, while it is worth noting that the individual differences variables such as happiness, satisfaction with life, self-esteem, personality, life regard and orientation among OHQ and OHQ-SF had also significant and strong correlation, demonstrating satisfactory validity [45, 50], internal consistency and test–retest reliability [50].

Recent systematic reviews regarding the measures of happiness revealed that the most common used multiple item scales were OHI, SHS, OHQ and OHQ-SF [51, 52]. The SHS has already been validated in Greek sample, similarly ΟΗΙ has been adapted and validated too [27, 47]. Due to the fact that the most happiness scales are relatively long, which may be an obstacle for administration, the choice of a short scale or a single item it would be more feasible to use for research or clinical purposes. Thus, the prospect of applying one of the most widespread scales, with good psychometric properties, compact and easy to administer which allows endorsements across a wide spectrum, lead us to choose the OHQ-SF to validate it in a Greek sample. The aim of this quantitative cross-sectional study will be to examine the Psychometric Properties of Oxford Happiness Questionnaire Short Form in Greek undergraduate student sample.

2 Method

2.1 Design

In the current research, a quantitative cross-sectional study design was applied to validate the translated OHQ-SF.

2.2 Participants

The sample of the current study was consisted of adult students of various field of studies who enrolled into a full-time undergraduate degree program in the Public Educational Training Institution, Evosmos Thessaloniki. The inclusion criteria for the participants were designated to be native Greek speakers aged over 18 years. The study aimed to recruit a sample size of at least 10 students per item, as an often-encountered threshold in the literature is that the ratio of sample size over the items should exceed the number ten [53,54,55]. Thus, a minimum sample size of 80 students was considered as sufficient in proportion to the 8-item questionnaire OHQ to be adapted. Furthermore, by using an a priori sample size analysis, where a = 0.05, power = 0.8, and the target Goodness-of-fit Index (GFI) is at least 0.95, the estimated number of participants is equal to 89 [7]. In the first assessment 101 participants were recruited voluntarily by signing a written consent and 88 participants followed-up with the test procedure 1 week later. One participant from the first assessment procedure requested data withdrawal during the first week (100 final sample of test assessment). The final sample with non-missing values in both assessments (test–retest) consisted of 88 participants which was considered as an adequate sample. The final sample consisted of 65 females, 20 males and 3 not specified, aged 18–57 (M = 21.58, SD = 7.54). The participants were recruited in 2 days for test and re-test assessment, respectively.

2.3 Procedure

The sample was applied to Public Educational Training Institution of Evosmos. The director of the Institution was asked for a written consent of conducting the study (Ν: 1912/12-12-2022). Then, an approval to conduct research by the Greek Ministry of Education and Religious Affairs (N: K5/2736/11-01-2023) was conceded. After discussion with the director regarding the convenient days and hours, the researcher visited the classes to proceed the assessment, where in the first time informed the participants about the purpose of the study. The scales were completed before class started, on the same day and time, at 1 week interval.

The participants initially were required to read the information sheet where all the appropriate information regarding the current study was provided, then to create their unique participant code by noting the initial letter of their name, address, number of their address, and initial letter of their father’s name. Afterwards the participants had to consent in all requirements for data collection in order to participate in the study, and to complete demographic data. This was followed by the completion of the two questionnaires (OHQ-SF/GR 8-item; LOT-R-GR 10-item) (first test assessment denoted as T1). The questionnaires assessment was repeated after 1 week (re-test assessment denoted as T2).

Previous studies have supported that optimism exhibits a moderate to strong positive correlation with happiness, therefore we utilized the Greek version of the LOT-R scale to explore the convergence validity [56]. The estimate average time for students to complete them was 10–15 min, respectively. The completed documents, on both occasions, were handed back to the researcher and were remained locked with access only to the supervisor and the researcher. The data collection was anonymous and confidential. The personal code was associated with each participant and was used again in the re-test assessment. The completed documents of the test assessment procedure were 101 which were completed appropriately, where one participant draw back the data 1 week after the submission. Correspondingly, the same procedure was followed in the re-test assessment; where 88 completed documents were submitted. Then, the data were identified based on the unique participant code and analysed.

2.4 Measures

2.4.1 Oxford Happiness Questionnaire Short Form (OHQ-SF)—an 8-item questionnaire

The OHQ-SF constructed by Hills and Argyle in 2002 [45]; it is a short version (8-items) of the Oxford Happiness Questionnaire (OHQ) (29-item questionnaires) [45]. OHQ-SF (Appendix) is a unidimensional tool including the 1st, 3rd, 12th, 13th, 16th, 18th, 21st, and 29th questions from OHQ. The 1st, 13th, and 29th are reversed items in scoring; this is a 6-point Likert scale instrument and is ranging from 1 = strongly disagree to 6 = strongly agree, scoring from 8 to 48 where higher scores indicate higher levels of happiness. The total score is obtained through the sum of the items. According to [45], the OHQ shows high reliability (\(a\) = 0.91); and follow-up studies have confirmed the internal consistency and test–retest reliability of the instrument [57]. The inter-item correlation recorded identical results, establishing that OHQ can accurately replace OHI, and the internal consistency designate that according to overall happiness, all the items contribute to a valid measurement. Correlation among the OHQ and OHQ-SF results was highly significant, r (168) = 0.93 (p < 0.001), and indicated the validity of the measurement [57].

2.4.2 Life Orientation Test Revised (LOT-R)—a 10-item questionnaire

The ‘Life Orientation Test’ was first constructed by Scheier and Carver in 1985 and then was republished in 1994 as Life Orientation Test—Revised (LOT-R) [58, 59]. LOT was designed to examine a person’s degree of optimism as an element of their personality. The tool contained 10 questions, four of which reflect expectations for positive outcomes in life (optimism factor), four for negative (pessimism factor), and the remaining four were filler items which were deleted in the calculation of the total score. The internal consistency reliability (Cronbach’s Alpha Coefficient) was 0.76 and the test–retest correlation was 0.79 [58]. LOT-R, is an improved two-factor structure instrument, which includes 10 questions, three from optimism factor (1st, 4th and 10th), three for pessimism (3rd, 7th, and 9th), and four are the filler that do not contribute to the outcome. The 3rd, 7th, and 9th are reversed items in scoring; this is a 5-point Likert scale instrument and is ranging from 0 = strongly disagree to 4 = strongly agree [59]. LOT-R has been translated, culturally adopted, and validated in a Greek population (hospital nurses, aged 22–65) by demonstrating good convergent validity (r = 0.73) with the one-item visual analogue optimism scale by Lai and Yue in 2000 [60], good internal consistency (Cronbach’s a = 0.71), and stability over a 3-months period [61].

2.4.3 Adaptation of the Greek Version of OHQ-SF

In the specific study, the English version of the OHQ-SF was translated into Greek using the forward and backward translation [62]. In particular, two Greek bilinguals translated separately the original English version into Greek. Then, an expert identified and resolved any discrepancies between alternative forward translations, and reconciled them into a single version. Then, another two bilinguals carried out the same procedure in the back translation from Greek to English. The originally translated and the back-translated Greek versions were then compared by an expert for consistency, relevance and meaning of the content [63, 64]. Before administering the final research questionnaire, the pre-final version was administered to seven individuals, to identify potential comprehension issues. The researcher delivered each question separately to determine how well the respondents comprehended them in order to be able to respond. All respondents reported that the terms were easily understood, the language was straightforward, and the questions were not confusing. As one of the seven identified some difficulty with the sixth item of the questionnaire, an additional opinion was sought from a bilingual, who provided the final version of this question: “Ταιριάζω σε όλα τα περιβάλλοντα.” (Appendix).

2.5 Statistical analysis

Descriptive statistics for the eight items include the means, standard deviations and pairwise correlations. Boxplots were used to identify the presence of possible outliers and the Shapiro–Wilk test examined if the variables under investigation followed the normal distribution. The pilot study assessed the external reliability using Pearson’s correlation coefficient (\(r\)) and kappa (\(k\)) coefficient, while for the instrument’s internal consistency, the Cronbach’s alpha (\(\alpha\)) and McDonald’s omega (\(\omega\)) coefficients were calculated for both test and retest conditions. The corrected item-total correlations were calculated through Pearson’s \(r\). The correlation between the OHQ-SF/GR and LOT-R Greek Version was used as an indicator of the construct validity of the OHQ-SF/GR.

Confirmatory Factor Analysis (CFA) was performed using the unweighted least squares (ULS) estimator to examine the univariate structure of the eight items of the OHQ-SF/GR. The ULS estimators tends to overperform the maximum-likelihood (ML) estimators, especially when the parametric assumptions are violated [65]. Furthermore, previous studies have found that they provide more accurate and less variability in parameter estimates compared to the ML or the diagonally weighted least squares estimators [66]. The Kaiser–Meyer–Olkin (KMO) statistic was used to estimate the applicability of factor analysis and Bartlett’s test of sphericity was used to assess the significance of item correlations. For the factor model, loadings greater than 0.30 were assumed to be acceptable [67]. Based on previous recommendations, different fit indexes were estimated and presented [68]. The model fit indices included the χ2 test statistic, Root Mean Square Error of Approximation (RMSEA), the Tucker-Lewis Index (TLI), the Goodness-of-fit Index (GFI), and the Comparative Fit index (CFI). For the ratio of the χ2 statistic over the number of degrees-of-freedom (df), a score below 2 indicates good fit [69]. For the other indexes, various cut-off points have been determined – RMSEA < 0.08 [15], CFI > 0.95 [62], GFI > 0.95 [53], TLI > 0.95 [30]. Data analysis was carried out using the statistical analysis program JASP (Version 0.18.3) [70].

3 Results

3.1 Pilot study

A pilot study was administered to 20 volunteer participants within an intermediate period of 2 days to complete the OHQ-SF Greek version (OHQ-SF/GR) and the original English version of OHQ-SF in order to examine the equivalency of the two instruments. The participants of the pilot study were all Greek native speakers with a Certificate in English language in Proficiency level, the formal requirement of Greek Ministry of Education and Religious Affairs for teaching English. The external reliability of the OHQ-SF/GR was tested in the pilot study. OHQ-SF/GR had a high positive correlation r = 0.98, p < 0.001 with OHQ-SF original version. By these results, the OHQ-SF/GR translated version was considered as equivalent to the original version. Additionally, the English version of the OHQ-SF instrument indicated weak to moderate internal reliability (Cronbach’s α = 0.50, McDonald’s ω = 0.49), and highlighting the presence of heterogeneous items or the existence of weakly connected items or a result of a limited number of questions [71].

The descriptive statistics (mean values and standard deviations), the kappa correlation of agreement between the two measurements, and the item correlations for the eight items are presented in Table 1. Items showed low to moderate correlations, except for item 5, which did not exhibit significant correlations with any of the other items.

Table 1 Descriptive statistics, kappa coefficient, and item correlations

A preliminary check was performed on the eight items for the presence of outliers using boxplots, and it was found that six low-score outliers were associated with the 5th item. After removing those outliers, a Shapiro Wilk test showed that there was a significant deviation from the normal distribution (p < 0.001).

According to the original version of the tool, the expected result was the unidimensionality of the questionnaire (one-factor for happiness). Confirmatory Factor Analysis (CFA) was performed using the ULS estimator to examine the univariate structure of the eight items. For the first iteration, the 5th item “I find beauty in some things—Βρίσκω ομορφιά σε κάποια πράγματα” had a non-significant factor loading (b = 0.06, p = 0.393); hence it was removed from the subsequent factor solution.

The final factor model consisted of seven items, and Cronbach’s alpha coefficient was equal to 0.75, while McDonald’s omega coefficient was equal to 0.76. Bartlett’s test of sphericity was significant (p < 0.001) and the KMO statistic (0.74) showed that the sample size was adequate, thus the data were regarded suitable for factor analysis. The corrected item-total correlations ranged from 0.31 (6th item) to 0.63 (1st item). The unidimensional factor structure provided a good fit to the data (\({\chi }^{2}\) = 22.78, p = 0.064; χ2/df = 1.63; RMSEA = 0.079, p = 0.193; CFI = 0.96; TLI = 0.95; GFI = 0.99), and all the factor loadings were significant and above 0.35 (Table 2). The average variance extracted (AVE) was equal to 34.4% and the Composite Reliability (CR) coefficient was 0.78, for the seven-item univariate factor model.

Table 2 Standardized factor loadings

3.2 Test–Retest reliability

Test–retest reliability was determined by the administration of the OHQ-SF/GR twice, with a 1-week intermediate time interval (T1 and T2), for the participants that were recruited in both the first assessment and the follow-up period (N = 88). The results indicated good internal reliability a = 0.72, ω = 0.73, and a = 0.69, ω = 0.71. for T1 and T2, respectively. This signifies that each item on the scale correlates positively and significantly with the sum of the other items. The correlation with Pearson’s r value showed a strong positive significant correlation r = 0.78 (p < 0.001) between the two measurements of OHQ-SF/GR. By removing the 5th item, the internal consistency coefficients slightly increased for both T1 (a = 0.77, ω = 0.75) and T2 (a = 0.73, ω = 0.73), suggesting higher internal consistency, while Pearson’s r without this item showed a stronger positive correlation r = 0.80 (p < 0.001).

3.3 OHQ-SF/GR scale’s structural validity

Structural validity of OHQ-SF/GR was evaluated with the use of LOT-R Greek version separately in test–retest (T1 and T2). OHQ-SF/GR had an adequate positive correlation r = 0.66 (p < 0.001) and r = 0.68 (p < 0.001) with LOT-R Greek version for T1 and T2, demonstrating that happiness is positive correlated with optimism, showing evidence that OHQ-SF/GR has adequate convergent validity. After removing item-5, the OHQ-SF/GR had an adequate positive correlation r = 0.69 (p < 0.001) and r = 0.70 (p < 0.001) with LOT-R Greek version in T1 and T2.

4 Discussion

The aim of the current study was to examine the psychometric properties of Oxford Happiness Questionnaire Short Form (OHQ-SF) in Greek undergraduate student sample. The Greek version of OHQ-SF was found for both test and retest assessment, to have an adequate internal reliability over 1 week period, satisfactory structural and adequate construct validity. The results of CFA have shown the unidimensionality of the OHQ-SF/GR, while the fit indexes demonstrated that the one-factor model had a good fit. However, the 5th item presented poor correlation with the one-dimensional factor (happiness) and it was removed. After removing the 5th item, the model provided a better fit, with an increase in the extracted variance which was explained and OHQ-SF/GR, for both test and retest, presented satisfactory internal reliability, structural validity, and convergent validity. Thus, the final 7-item scale, OHQ-SF/GR showed very satisfactory psychometric properties and being a useful instrument to measure happiness levels in Greek adult students that can be utilized in psychological field of research. It is worth noting that the OHQ-SF contained reverse scoring items, which, according to the literature, may have a negative effect on the psychometric properties of the scale [72]; nevertheless, it is questionable to what extent the reversed scoring items control the answer sets [73].

Providing similar findings with the original version [45], the OHQ-SF/GR certified the unidimensionality of the tool identifying happiness. Nevertheless, the internal consistency reliability and test–retest reliability of OHQ-SF are not mentioned by the authors in the original version, while another study evaluated the original version at 2 weeks interval [50]. In our study, the final seven-item OHQ-SF/GR presents higher Cronbach’s alpha to the original version of eight items [50], where the sample in both studies were adult students. Furthermore, there is no information about the loadings of the items in the original scale.

Regarding the adaptation and validation in other languages, the OHQ-SF has been translated and validated so far in Spanish (nursing students’ sample), Turkish (university students’ sample), Chinese (12-year-old children sample), Farsi (university students’ sample), and Portuguese (general population sample) demonstrating proper psychometric properties for research. Specifically, OHQ-SF Spanish version presented satisfactory internal consistency and temporal stability (α = 0.73), similar with the current study, while the scale correlated positively with LOT-R, similar with the Greek version [74]. Concerning the Turkish version [75], after some modifications to the original scale, the authors reported a one-dimensional factor achieving satisfactory construct validity and internal reliability (α = 0.74), quite similar with the current study. Of particular interest is the Chinese version [76], which was carried out on underaged population (12-year-old children), the items were separated in two dimensions and the internal consistency was quite lower than in the current study (α = 0.63). A possible explanation would be the different, non-adult sample population. Farsi and Portuguese versions presented both appropriate psychometric properties, where in comparison to OHQ-SF/GR they had lower, but again adequate Cronbach’s alpha (α = 0.71, α = 0.70, respectively) [77, 78].

Consonant with the above, and even though happiness is universal, it is observed that cultural differences may reflect differences in the measurement scales, which, according to the literature, lead to a seven-item construct of the scale for several versions (language adaptations) [15, 16]. For instance, removing item- 8 in the English version [50], item-4 in Turkish [75], item-7 in Chinese [76], but also in our study-Greek version, removing item 5 was the best solution for the final versions of all the above adaptations.

Additionally, from the existing translated OHQ-SF versions, the tool is reported as unidimensional except for the Chinese version; where [76] suggested that the two dimensions refer to both societal and individual fulfilment facets of happiness. Accordingly, the findings present that optimism is positively correlated with happiness, which is consistent with the literature [79].

The psychometric properties of OHQ-SF/GR can compare favourably with the psychometric properties of other unidimensional scales measuring happiness. In specific, the SHS in its original version, was used in high school students and adults, presenting adequate to excellent internal consistency, and the test–retest reliability ranged from 0.55 to 0.90 [24]; the SHS Greek version, which carried out in general population [27], showed satisfactory internal reliability (α = 0.76), quite similar with OHQ-SF/GR. On the other hand, the SHS Greek version presented weak positive correlation (r = 0.11) to optimism (LOT), while for the original version of OHQ-SF it had moderate levels (r = 0.47–0.60), while the current study exhibited moderate to strong convergent validity (r = 0.69). One previous study provided an explanation to the specific finding, referring that while subjective happiness is typically regarded as a state, optimism is often viewed as a trait despite its positive emotional nature [27]. It is noteworthy to mention a significant structural contrast among the two questionnaires, wherein the OHQ-SF incorporates three reverse items, whereas the SHS does not include any reverse items.

The psychometric properties of OHQ-SF/GR compared to the multidimensional tools which have been adapted in Greece present interesting results. The commonly used scale PANAS adapted from [38], in high school students and adult students indicated satisfactory validity and internal reliability for the positive and negative affect scale (α = 0.71 and α = 0.79, respectively), correspondingly, similar with the OHQ-SF/GR; whereas the most widely applied scale Satisfaction with Life Scale (SWLS) adapted from [80], in adult students, presented slightly higher internal consistency in test–retest (α = 0.87, α = 0.82, respectively) than the OHQ-SF/GR. The latter comparison comes in line with a part of literature which suggests that the correlation between happiness and life satisfaction exists, however the two constructs differ [13].

In summary, the seven-item version of the OHQ-SF/GR has demonstrated sound psychometric properties, thereby establishing its validity and reliability as a measurement tool for measuring happiness in Greek adult students. The findings showed a positive correlation between optimism and happiness; thus, it is expected that happy individuals are more optimistic or vice versa. According to the recommendation provided by the original author [45], the more compact and straightforward version can be substituted for the longer-standing one in any and all circumstances, including those in which there is a limited amount of time. Therefore, this instrument holds potential as, compact and easy to administer, a valuable research tool for psychological research in Greek context.

4.1 Limitations and future research

A limitation of this study is that it relied on self-report questionnaires to collect the subjective judgments of students, who were mostly comprised of young people who lived in urban areas. Due to the fact that the findings are contingent upon the specific set of participants who took part in the study, there is a pressing need for additional research with participants of varying demographic characteristics, belonging to a wide variety of subgroups, and hailing from a number of distinct geographic regions. Another limitation could be identified in the convenience sample which is considered as an unrepresentative sample of the population and the result lacks wide generalisability as a result of sample bias [81]. However, it is a widely used non-probabilistic method on which research is based [82]. Thus, a generalisation of the findings from more extensive investigations is something that can further contribute to this discipline.