The Social Awareness Index as a Tool to Determine the Role of a Companion Robot in the Lives of People with Reduced Mobility

Abstract

Robots are becoming an increasingly popular element of the developing technological reality. A companion (social) robot can be defined as an autonomous machine that recognises other robots and people and engages in social interactions. Robots of this type are designed to serve people, so they often act as guides, assistants, companions, guardians, teachers and domestic pets. This study aims to define and operationalise the components and synthetically measure the social awareness (index) concerning the usability of a companion robot in everyday life. A comprehensive new research tool was developed as part of the study in the form of an original standardised interview questionnaire by the CATI/CAWI technique. Empirical studies show that an increase in social acceptance requires that potential consumers be provided with the necessary information to know the technologies better and improve their digital skills related to social robots.

1. Introduction

The rapid development of information technologies, including various robotic solutions, has increased the interest in human cooperation with companion robots. One of the major areas of interest pursued in various research initiatives includes the effectiveness of these devices and their cooperation with humans [1,2,3,4,5].

The literature on the subject reveals a distinct orientation towards analysing these issues from the perspective of various age groups, particularly older adults [6,7] as well as children and adolescents [8,9,10]. The potential for the use of companion robots in the therapy of various illnesses and ailments and care of the sick [11,12,13,14], as well as various ethical aspects associated with human cooperation with a companion robot [15,16,17], are popular thematic areas.

The use of such solutions can be very helpful for people who lose the ability to make autonomous decisions and/or those who have lost their independence due to illness or age. Support in the form of robot companions will be particularly valuable for people with limited mobility because it can help improve their well-being and mental condition. In this context, considerations and work on this issue are, therefore, part of the Third Sustainable Development Goal, which assumes “Ensuring healthy lives and promoting well-being for all at all ages” (SDG 3). Notably, several studies have examined the extent of social awareness of the current and future role of companion robots and the level of acceptance for this technological solution [18,19,20,21], with attitudes varying significantly among different social groups.

Social awareness is a widely discussed concept. It is defined in various contexts, mainly in the philosophical, psychological, educational, and sociological areas. French sociologist E. Durkheim [22] defined social awareness as a set of beliefs, values and norms common to all the members of a community, which shape their attitudes and actions, and A. Giddens [23] specified that social awareness includes both the understanding of social structures and the ability of individuals to reflect on their roles in these structures and on the structures’ impact on their lives. According to the definition proposed by P. Sztompka, social awareness is a set of beliefs, ideas and convictions, common and widely accepted in a community, which become patterns of thinking instilled in its members and enforced by social pressure [24]. According to S. Kuśmierski’s definition, social awareness comprises the body of knowledge, a set of values and ideas of a social group [25]. Following S. Greenspan, social awareness can be defined as an individual’s ability to understand people, social events and processes associated with regulating social events [26].

Along with social sensitivity, human awareness is a cognitive component of human competence. It should be indicated that the term “awareness” has more than one meaning. It can either mean “focusing on something”, “realising what is happening inside us or in our closest surroundings or “the highest level of a human mental development” [27]. As understood in research practice, the term “social awareness” usually includes opinions, convictions, knowledge, views, preferences, beliefs, ways of thinking, hierarchies of assessments and elements of attitudes [28]. Interestingly, studies concerning social attitudes are equally often initiated in social studies of relations with social robots and social awareness. The sociological concepts represented by W. Thomas and F. Znaniecki refer to the relationship between awareness and attitude. In their view, an attitude is a process of individual awareness, which determines both the current and potential responses of each person to the social world [29]. According to S. Nowak, an attitude can be defined as a set of relatively permanent dispositions to assess an object and to respond to it emotionally and of accompanying emotional and judging dispositions for relatively permanent convictions concerning the nature and properties of an object and relatively permanent dispositions to behave towards the object [30]. Scholars point to a structural nature of attitudes, which comprise two elements: intellectual (cognitive) and emotional (affective) [29], p.17. After Mikon J. Rosenberg [31], the following three components of attitudes can be identified: cognitive (knowledge, suppositions, convictions about an object), affective (feelings and emotions towards an object) and conative (intentions concerning behaviours towards the object, dispositions to behave positively or negatively towards the object) [32]. Recognising human attitudes towards robots is important in further studies and projection of interactions. It can allow forecasting consequences resulting from human interactions with real robots. Research results can be used to improve the effectiveness of designing new robots by programming conflict-free and safe interactions with robots [33] and adapting them to the cultural context, users’ age, and preferences. The purely cognitive aspect should also be appreciated [34]. Wykowska [34] indicates that, apart from acting as helpers and companions in various areas of human lives, social robots can be used as scientific tools for studying the human system of social cognition [34].

Reliable tools used to measure attitudes towards robots are an important methodological issue in studies of social attitudes. These tools include the following six scales [35] tested psychometrically and diverse in terms of content and purpose that measure attitudes towards robots: the Negative Attitudes towards Robots Scale (NARS) [36], the Robotic Social Attributes Scale (RoSAS) [37], the Ethical Acceptability Scale (EAC) [38], the Technology-Specific Expectations Scale (TSES) with the Technology-Specific Satisfaction Scale (TSSS) [39], the Frankenstein Syndrome Questionnaire (FSQ) and the Multi-Dimensional Robot Attitude Scale [40].

Robots are becoming an increasingly popular element of the developing technological reality. A companion (social) robot can be defined as an autonomous machine that recognises other robots and people and engages in social interactions [41,42]. Robots of this type are designed to serve people, so they often act as guides, assistants, companions, guardians, teachers and domestic pets [32]. This study aims to define and operationalise the components and synthetically measure the social awareness concerning the usability of a companion robot in everyday life.

There is a shortage of proposals for a synthetic measurement of social awareness in the literature. For this reason, based on the results of the study within the VALET project, it was decided to construct such a measure and diagnose the social awareness of people with limited mobility in this regard.

2. Materials and Methods

The survey questionnaire developed as part of this study is an original tool used to measure selected aspects of the social awareness of persons with reduced mobility regarding the use of companion robots in everyday life. This study is part of the VALET project (navigating a companion robot as a tool for improving the quality of life of people with reduced mobility).

Physical disability is generally understood as being limited to dysfunction of motor organs. However, similar to Wolski [43], p. 8, this study assumed that physical disability encompasses a range of dysfunctions unrelated directly to the motor organs.

The term range of “a person with reduced mobility”, used in article, was the closest to the assumptions adopted by the authors of the current study. Article 2 of Regulation (EC) No. 1107/2006 of the European Parliament and of the Council of 5 July 2006 concerning the rights of disabled persons and persons with reduced mobility when travelling by air [44] states that a person with reduced mobility denotes “any person whose mobility when using transport is reduced due to any physical disability (sensory or locomotor, permanent or temporary), intellectual disability or impairment, or any other cause of disability, or age, and whose situation needs appropriate attention and the adaptation to his or her particular needs of the service made available to all passengers”.

Individuals with physical disabilities do not make a uniform group. Various factors can cause disability. It is characterised by a considerable diversity of dysfunctions and a degree of mobility reduction, and it can cause various psychological and social effects depending on age, phase of life and previous lifestyle [43], p. 10. The diversity of the effects of disability linked with the socio-demographic and economic situation can cause considerable disproportions in the subjective opinion of disabled people about their quality of life.

In the study, it was assumed that the unit sampling frame for the survey would ensure the random selection of a representative sample of respondents. The sampling for the survey was conducted in two stages, the first of which consisted of generating telephone numbers by selecting the successive digits of those numbers. The telephone numbers thus obtained were the basis for the pre-recruitment of respondents to the survey using a filter question concerning the presence of mobility limitations of the person drawn for the survey or another person in the household. Apart from verifying respondents with respect to mobility constraints, the Contractor was obliged to verify the demographic particulars of such persons, with a view to preserving the structure of the realised sample that was to reflect the demographic structure of the population of Poland as of 2021. The achieved research samples were characterised by quota sizes assumed by the Contractor and approved by the Contracting Entity. Differences (by +/−1.06 percentage points) in the percentage shares of individual categories within the demographic structure result from rounding the quota sizes to unity.

Surveys were carried out using two complementary statistical observation techniques: Computer-Assisted Telephone Interviewing (CATI) and Computer-Assisted Web Interviewing (CAWI). The choice of observation method (interview form) was left to the respondents who qualified for the survey after pre-recruitment over the telephone. Thus, the respondent had the option of taking part in the survey via telephone or completing the questionnaire available on the website on their own. Indicating the latter option involved providing an email address to which a link to the survey questionnaire was sent. About 80% of the responses came from the CATI study.

Prior to conducting the main part of the survey, the interviewers underwent training with the Contractor (BST Group) on 30 January 2023; the training was attended by the research team. The survey was conducted from 31 January 2023 to 30 March 2023. Before launch, the Contractor submitted the sampling scheme to the Contracting Entity for approval, describing the structure of the research sample (percentage distribution between the indicated categories of characteristics) and the assumed number of questionnaires for each category of demographic characteristics of the sample, which added up to the total number. Subsequently, the study commenced with the random generation of telephone numbers to potential respondents. After consenting to take part in the survey and obtaining a positive outcome from pre-recruitment, the respondents would choose the form of the interview. After selecting the online survey, the respondent was required to provide an email address. If they failed to complete the survey within two calendar days, the Interviewer had to contact the Respondent again via telephone to conduct a telephone interview. If the person contacted by telephone stated that they were not a person with reduced mobility, they were asked whether they lived with such a person. If the answer was affirmative, the possibility of conducting an interview with such a person was requested, the person’s consent was sought, and the demographic characteristics were verified.

On 10 March 2023, the Contractor in charge of the survey submitted the databases containing the individual answers of the respondents. Following an analysis of the databases, return comments on the structures of the realised survey samples—in terms of the socio-demographic characteristics as well as completeness and correctness of coding—were provided on 24 March 2023. On 30 March 2023, the Contractor submitted the corrected and supplemented database. Each database contained the individual responses of 1064 respondents. A total of 17,163 telephone calls were made in order to effectively meet the planned survey sample size. The survey response rate was 6.2%.

This study involved people aged 15+ who declared reduced mobility, which is associated with movement problems arising from dysfunctions of the motor organ (lower limbs, upper spine). A comprehensive research tool was developed as part of the study in the form of an original standardised interview questionnaire by the CATI/CAWI technique (six pre-recruitment questions, sixteen questions concerning the study object, divided into seven thematic blocks, four questions on the respondents’ particulars and seven technical questions—altogether thirty-three questions). This study assumed a random selection of a representative respondent sample. In total, a sample of 1064 people participated in the study, corresponding to the socio-demographic profile of Poland.

An analysis of the definition ranges and available empirical studies of social awareness and social attitudes towards robots helped to determine the components that define selected elements of social awareness in this study. Seven areas were taken into account in order to describe the level of social awareness of the role of a companion robot: (1) knowledge of a companion robot, (2) perception of a companion robot, (3) barriers preventing the use of a companion robot, (4) projection of a role, (5) function, and (6) time period before a companion robot is used in everyday life by people with reduced mobility, and (7) the level of a robot acceptance. The adopted thematic scope, divided into these components, resulted in the development of a data acquisition tool in the form of a standardised survey questionnaire. The questions concerning the study object were divided into thematic blocks that addressed selected aspects characterising components of social awareness.

By adopting the commonly used methodology of measuring complex socio-economic processes and phenomena [45], p. 99, the questionnaire construction was based on sets of statements assessed by the respondents according to a scale. An ordinal scale of relatively high intensity, from 0 to 10, was used to improve the measurement accuracy. Scores for individual statements (items) are simultaneously values that allow for constructing diagnostic variables. Owing to these variables, it was possible to construct aggregate sub-measures that synthesise scores for the seven components of social awareness and to construct the final synthetic measure of the social awareness level.

Verification of reliability is one of the requirements for the construction of internally coherent synthetic measures based on sociometric or psychometric scales. Scale assessment is usually made for original survey questionnaires or if existing questionnaires are modified. The Cronbach alpha was used to measure the reliability of scales, i.e., categories of responses to questions divided into thematic blocks. This method helps to assess the internal coherence of scales used in test questionnaires. Internal coherence describes the level of measuring a term by all items. This measure is applied with the dichotomous response format, and responses are measured on polynomial scales (e.g., Likert scale [46]). The Cronbach alpha can lie between 0 and 1. Higher values denote higher scale reliability. It is assumed that values above 0.7 are characteristic of a reliable scale. To this end, the following formula was applied:

$$\alpha ={\displaystyle \frac{k}{k-1}}\left(1-{\displaystyle \frac{\sum _{i=1}^k{s}_i^2}{s_c^2}}\right)$$

(1)

where:

k—number of test items;

s_i²—variance of the i-th item (assessed categories/items);

s_c²—variance of overall results (variance of sums of assessments for all assessed categories/items).

An assessment of the scale reliability leaves out the block concerning the respondents’ knowledge of the possibility of using a companion robot (questions with diverse scales of response) and a question about an estimated time period before such robots are used widely by society (question about the number of years). An analysis of the scale reliability for the other thematic areas, based on the Cronbach alpha, demonstrated the high internal coherence of the scales. All the coefficients had values above 0.7 (cf.

Table 1. Cronbach alpha values for measurement scales for social awareness components.

	Perception of a Companion Robot	Barriers to Robot Use	Projection of Robot Role	Projection of Robot Function	Level of Robot Acceptance
Cronbach alpha	0.889	0.953	0.915	0.982	0.909
N	796	915	896	905	895
Scale items	6	11	8	35	5

N—sample size after “no responses” were excluded. Source: prepared by the author in the R Cran program, based on the results of a survey among persons with reduced mobility, conducted as part of the VALET project, purpose: C.S.2.2.

).

Methods of Data Analysis

The scope of the statistical description and inference based on the survey results depended on the measurement scales for the individual variables that made up the detailed series of responses to the socio-demographic questions (concerning the respondents’ particulars) and those concerning the survey’s subject matter. The socio-demographic questions include respondent’s gender, age (divided into age groups: 15–18, 19–25, 26–44, 45–64 and 65+), town/village, place of residence (according to its size: village, town of up to 25 thousand residents, town of up to 50 thousand residents, town of up 100 thousand residents, town of up to 250 thousand residents, town of up to 500 thousand residents, town of over 500 thousand residents), voivodship (NUTS 2 regions in Poland), education (incomplete primary, primary, junior high school, vocational, vocational secondary, general secondary, post-secondary, bachelor’s degree, engineer’s degree, master’s degree, post-graduate degree), living together with children aged 0–14 years and being married. The Pearson linear correlation coefficient was used to assess the relationship between individual components of the synthetic measure of social awareness and their impact on the aggregate variable. The following statistical tests were used in order to verify the statistical relationship between individual variables:

Pearson’s independence chi-square (χ²) test [47] was the base for verification of conformity or independence for variables measured on the nominal level. Pearson’s chi-square test was used to determine relationships between any two variables. The Frank Yates continuity correction test was used to calculate chi-square tests for two-value variables. [48], p. 12.

Cramer’s V measures the correlation strength between two qualitative variables measured at the nominal levels. It supplements the chi-square test in a dual sense: first, calculations are based on it, and second, it takes information from the calculated chi-square to determine whether the correlation is statistically significant. Therefore, calculating chi-square and then Cramer’s V provided two items of information: about the existence or exclusion of a statistically significant correlation between variables (chi-square) and the strength of this correlation (Cramer’s V) [49], pp. 231–232.

The Kruskal–Wallis test expands the Mann–Whitney U test for two independent samples. It is applied when differences in the distribution of more than two compared groups are verified. The test is a non-parametric equivalent of a one-way analysis of variance used when the series under analysis fails to meet the assumption of a normal distribution. Regarding computations, the test statistics are based on the rank analysis. It detects differences in the distribution position [50].

Pearson’s correlation coefficient allows for determining the strength and direction of a linear relationship between two quantitative (interval or ratio) variables. The minimum sample in the Pearson r test should have 30 observations (N ≥ 30). The Pearson r coefficient lies between −1 (perfect negative correlation) and +1 [49], p. 204.

A descriptive analysis of the diagnostic attributes that characterise the level of social awareness was conducted in division into thematic ranges that correspond to components of social awareness adopted in this study. The discussion of the study results focused on demonstrating differences in response distributions and assessing the statistical significance of the impact of socio-demographic attributes, which can potentially influence differences in response distributions and characteristics of the distribution patterns. The relationships between the socio-demographic attributes were examined independently. It is noteworthy that these attributes can be inter-related. The age of a respondent and living with children can be examples. Half of the respondents who lived with children were aged 25–44 years, and respondents aged 54–64 years accounted for another 31% of the respondents. Only 7% of respondents were aged 65+ years.

Variation of the social awareness level concerning the role of companion robots in everyday life—the concept of an aggregate measure

A thematically diverse range of areas that define social awareness in the study provided the base for constructing seven separate sub-measures (components of the aggregate measure) (cf. Figure 1). An awareness measure consists of sub-measures related to the component areas that build social awareness concerning companion robots.

The first stage of the measure construction involved quantifying the nominal categories (dichotomous categories of responses and statements) and scores measured on the ordinal scale (from 0 to 10). To this end, an algorithm was developed for transforming the responses to the values of diagnostic variables (cf. Appendix A). Subsequently, program codes were developed to re-encode the responses in the individual response database. Before making calculations, a transformation of the destimulants (scores for categories of negative responses) into stimulants was required.

The construction of synthetic sub-measures involved calculating the share of the total score (the number of points gained by a respondent) in the maximum possible score. The sums were divided by the maximum score to facilitate interpretation and ensure that the measure variation ranges are comparable. Each measure lies between 0 and 100. To this end, the following formula was used:

$${subWss}_{i,k}={\displaystyle \frac{\sum _{j=1}^m{x}_{ijk}}{\sum _{j=1}^{m}max\left({\widehat{x}}_{jk}\right)}}\ast 100$$

(2)

where:

$x_{ijk}$—i-th observation of j-th variable, range of variables, attributed to the k-th sub-measure [knowledge, perception, barriers, role, functions, time, acceptance],

$max\left({\widehat{x}}_{jk}\right)$—maximum, theoretical value of the j-th variable of the k-th sub-measure;

Construction of the aggregate synthetic measure of social awareness reproduced the synthetic measure construction method proposed by Perkal [51], p. 14. The measure form is an arithmetic average of synthetic sub-measures according to the following formula:

$${Wss}_i={\displaystyle \frac{1}{p}}\sum _{k=1}^p{subWss}_{i,k}$$

(3)

where:

p—number of sub-measures (partial measures);

3. Results and Discussion

The calculations for the sub-measures and the aggregate synthetic measure (WSS) show that a medium level of social awareness characterised half of the respondents (cf.

Table 2. Descriptive statistics for sub-measures and the aggregate measure of the social awareness level.

Responses	N	$\overline{\mathbf{x}}$	sx	Me	Q1	Q3
WSS	1064	44.82	18.18	48.72	35.39	58.11
Knowledge	1064	6.55	16.72	0.00	0.00	0.00
Perception	1064	36.52	24.34	43.33	0.00	53.33
Barriers	1064	32.65	27.38	29.09	9.09	49.09
Role	1064	56.86	32.95	66.25	36.25	82.50
Functions	1064	64.43	33.89	76.57	49.43	91.43
Time	1064	57.95	40.11	80.00	0.00	90.00
Acceptance	1064	58.81	35.08	66.00	35.50	90.00

Source: prepared by the author in the R Cran program, based on the results of a survey among persons with reduced mobility, conducted as part of the VALET project, purpose: C.S.2.2.

). A very low level of the respondents’ knowledge of companion robots and high scores for the barriers and fear of using a companion robot probably contributed to the greatest extent to a decrease in this score. Despite the destimulating factors for assessment of the social awareness level, one should note the respondents’ positive opinions about the relatively short time period before companion robots become widely used in everyday life by persons with reduced mobility (half of the respondents gave this factor 80/100 points). High scores for the companion robot functions are another factor with a positive impact on the assessment of the level of social awareness.

Statistical inference based on the Shapiro–Wilk test for distribution normality was performed to assess the distribution of the measures. A normal measures distribution is necessary [49], p. 325 for the assumed application of the analysis of variance (ANOVA) to verify the relationship between the measures and the socio-demographic variables. The zero hypothesis (made when the test probability, p-value > 0.05) determines the normality of the distribution of an attribute under study (the model residues in the case under analysis), whereas the alternative hypothesis indicates that the attribute distribution is not normal (p-value < 0.05). The measure distributions are illustrated with histograms (Figure 2). They represent the distribution of values of aggregate synthetic measure (WSS) and sub-measures (divided into class interval groups) through plotted bars. The bar height reflects the frequency—the number of respondents. Due to differences in respondents’ assessments in terms of the analysed measures, for a clearer illustration, the axes of individual charts have different scale ranges. Histograms show the lack of symmetricity of distributions of most of the measures under analysis (most distributions are leftward skewed). The analysis showed the absence of normality for the synthetic measure distribution. It was, therefore, unjustified to use the analysis of variance to verify the relationship between the socio-demographic characteristics (measured on nominal and ordinal scales) and the values of synthetic measures. For this reason, the Kruskal–Wallis test—a non-parametric equivalent of the one-way analysis of variance ANOVA—was used to verify the relationships between these attributes.

An assessment of the variation of aggregate synthetic measures (WSS) concerning the socio-demographic attributes showed a significant relationship (p-value < 0.05) between the social awareness level and age and living with children. On average, a higher level of social awareness was observed in respondents aged 19–44 years and people who lived with their children (as was shown earlier; there is a strong correlation between these two attributes). A higher level of knowledge was observed in respondents aged 15–18 years and those who lived with children (

Table 3. Results of the Kruskal–Wallis test used for verification of the statistical significance of differences in synthetic measures concerning the category of socio-demographic attributes.

		Gender	Age	Town/Village	Place of Residence (by Size)	Voivodeship	Education	Children	Marriage
WSS	K-W stat	0.064	16.793	0.412	5.855	11.036	9.397	3.865	0.021
	p-value	0.801	0.002	0.521	0.440	0.750	0.495	0.049	0.884
Knowledge	K-W stat	0.471	15.674	0.930	1.341	19.818	5.034	10.081	0.017
	p-value	0.493	0.003	0.335	0.969	0.179	0.889	0.001	0.896
Perception	K-W stat	1.077	7.163	0.119	9.407	13.906	10.838	1.774	0.365
	p-value	0.299	0.128	0.730	0.152	0.533	0.370	0.183	0.546
Barriers	K-W stat	11.305	13.594	0.583	8.416	21.449	9.642	1.086	1.404
	p-value	0.001	0.009	0.445	0.209	0.123	0.472	0.297	0.236
Role	K-W stat	6.020	13.857	1.407	7.466	8.219	9.499	2.306	0.893
	p-value	0.014	0.008	0.235	0.280	0.915	0.485	0.129	0.345
Functions	K-W stat	13.396	11.787	3.451	8.515	19.578	9.919	2.276	0.309
	p-value	0.000	0.019	0.063	0.203	0.189	0.448	0.131	0.578
Time	K-W stat	2.726	4.025	3.630	9.487	9.870	25.390	0.078	0.145
	p-value	0.099	0.403	0.057	0.148	0.828	0.005	0.781	0.703
Acceptance	K-W stat	6.727	11.352	1.492	5.328	9.973	10.550	0.047	0.057
	p-value	0.009	0.023	0.222	0.503	0.821	0.394	0.829	0.811

Source: prepared by the author in the R Cran program, based on the results of a survey among persons with reduced mobility, conducted as part of the VALET project, purpose: C.S.2.2.

). The variation in scores for perception was not significantly affected by any sociodemographic attribute. Women tended to express more concern and mention barriers, but they also saw robots as highly important in projecting future roles and the range of tasks they performed. This perception resulted in a higher score given for a companion robot’s acceptance level. People at non-mobility age (45–64) and post-working age (65+) were assigned high scores for fear and barriers. The highest scores for society’s openness to robots performing specific roles in society and scores for functions and social acceptance were given by respondents aged 25–44 years.

The analysis of the correlation between the measures showed that despite the statistical significance (p-value < 0.05) of the impact of the knowledge of a companion robot on the variation of the aggregate measure and the other sub-measures, this impact was positive and low (

Table 4. Pearson correlation coefficients with statistically significant values marked.

	WSS	Knowledge	Perception	Barriers	Role	Functions	Time	Acceptance
WSS	1	0.26 ***	0.55 ***	0.31 ***	0.73 ***	0.73 ***	0.66 ***	0.73 ***
Knowledge	0.26 ***	1	0.12 ***	0.01	0.09 ***	0.1 ***	0.1 ***	0.09 ***
Perception	0.55 ***	0.12 ***	1	0.45 ***	0.27 ***	0.18 ***	0.19 ***	0.27 ***
Barriers	0.31 ***	0.01	0.45 ***	1	−0.02	−0.03	0.06	0.03
Role	0.73 ***	0.09 ***	0.27 ***	−0.02	1	0.59 ***	0.37 ***	0.5 ***
Functions	0.73 ***	0.1 ***	0.18 ***	−0.03	0.59 ***	1	0.37 ***	0.57 ***
Time	0.66 ***	0.1 ***	0.19 ***	0.06	0.37 ***	0.37 ***	1	0.31 ***
Acceptance	0.73 ***	0.09 ***	0.27 ***	0.03	0.5 ***	0.57 ***	0.31 ***	1

***—p-value < 0.05. Source: prepared by the author in the R Cran program, based on the results of a survey among persons with reduced mobility, conducted as part of the VALET project, purpose: C.S.2.2.

). The strongest correlation with the scores for social awareness was observed for the scores given for the role that should be played by a robot, projection of function and the level of a robot acceptance. A negative but non-significant correlation was observed between the scores for barriers and the role of a robot and its proposed functions. High scores for the robot functions resulted in positive opinions on the role that a robot should play and a higher level of social acceptance.

Negative scores for the possibility of using a companion robot in everyday life, which characterise mainly respondents aged 15–19 years and those at a post-working age (65+), suggest a need to propagate knowledge of technological solutions that enable the use of social robots to assist people with disabilities. Empirical studies show that an increase in social acceptance requires that potential consumers be provided with the necessary information to know the technologies better and improve their digital skills related to social robots [52], p. 160.

The high cost of a robot was the main barrier to a companion robot’s availability and common use. Important factors that limit the use of robots also include fear of entrusting one’s relatives to machines, which is associated with a lack of trust in machines and devices and reluctance to use them. The respondents were less afraid of damaging a robot or unwilling to learn to operate a new device. The attitude of potential users, including older adults, especially women and people living in small towns, may be a significant issue in propagating the use of robots, as this group was the most frequent to indicate the importance of most barriers that limit the use of a companion robot in everyday life.

Assisting persons with reduced mobility, adapted to their preferences and needs, with the possibility to turn the robot off, was among the most preferred roles to be performed by a companion robot. The possibility of interacting with humans enjoyed much less interest. People at a post-working age mentioned the need to gather information by robots about their users more often than younger respondents. This could be caused by a lower awareness and fear of cyber threats associated with gathering and sharing information, as well as the need to monitor the health and behaviours of potential users to stay independent (from the family and carers).

The score for the usefulness of the proposed functions to be performed by robots showed that they should largely monitor the surroundings to ensure safety and alarm about dangers to health or life. Despite the theoretically high importance of monitoring physiological parameters, the possibility of using these functions was given a lower score, i.e., the same as reminding the user about important activities or planned actions. The lowest scores were given for assistance in performing basic activities and providing company, entertainment, and communication, which confirms opinions showing a low trust in companion robots and reluctance to interact with them. Studies conducted in Denmark and Austria also showed society’s aversion to robots’ social behaviours [53]. It was emphasised that the social expectations towards robots are focused largely on utility functions, and they less frequently focus on robots as socially interactive devices [54]. This conviction is associated with an opinion that robots should serve people rather than replace them [55].

When asked to estimate the time before a companion robot can be used in everyday life by persons with reduced mobility, the respondents said it would be possible in ten years (on average).

The scores in categories characterising the level of acceptance for a companion robot showed that although the common use of companion robots in nursing homes and hospitals is desirable, the respondents were reluctant to use them themselves or to have robots used by their parents. The level of acceptance for a companion robot was higher among the respondents aged 65+ years. This relationship aligns with studies which found elderly persons to be prepared to accept devices assisting in healthcare to maintain independence [56]. Failure of some telemedical projects is believed to be partially a result of wrongly identified needs and expectations. A careful assessment of the needs and mapping them into technologies can result in a higher level of acceptance [57]. For this reason, the importance of pre-implementation studies conducted among potential beneficiaries of the technological solutions to be implemented and their formal or informal carers is emphasised. To this end, information can be gathered by conducting a pre-implementation survey [58].

The analysis of relationships between the synthetic measures showed the scores for barriers to have a negative impact on the perception of the role and preferred functions of robots. However, the relationship was not statistically significant. The score for the proposed functions was associated with a positive perception of a robot’s role and social acceptance. These elements were responsible for the variation in social awareness aggregate measure to the greatest extent. The above relationship demonstrated in this study confirms the study’s conclusions that usefulness is a condition for the acceptance of social robots [52], p. 163. According to previous studies, the presence of robots in everyday life does not increase their chances of acceptance or the will to interact with them [59]. Since it has also been indicated that robots’ ability to interact with people is still poorly developed regarding the technological development of robots’ basic social skills and the social acceptance of social interactions with interactive technologies, robot designers should try to improve robots’ social skills.

4. Conclusions

Disseminating knowledge of the applications of companion robots and gradually familiarising people with new, useful technologies are necessary conditions for building social awareness of companion robots. This especially concerns older adults, i.e., those with a lower level of knowledge, trust in new technologies and ability to operate modern devices, who are potential support beneficiaries. Awareness of robots’ functions and intuitive operation can increase their social acceptance and use in everyday life.