South African citizens’ self-assessed knowledge about the fourth industrial revolution

Abstract

This study assessed the levels of knowledge of South African citizens about the fourth industrial revolution (4IR) to determine how organizations can assimilate the implications of 4IR to leapfrog the development of the country. The findings show that knowledge was higher among citizens with better socio-economic status, higher education, with in excess of R6 500 (US$464.29) household income per month, are employed full-time, and have access to uncapped internet data or could afford to spend more than R150 (US$10.72) on data per month (more than 1 GB). There were low levels of knowledge among unemployed citizens (excluding tertiary institution students and unemployed graduates) and entrepreneurs. This was concerning as these groups are the main target for the country’s economic development and growth. The study recommends increasing knowledge sharing about the 4IR and related technologies and raising awareness or advocating appropriate skills for future careers. This study serves as a baseline for the policymakers and other development agencies or stakeholders to effectively set a national agenda to respond to opportunities and challenges emerging with the 4IR.

Keywords:

• citizens
• fourth industrial revolution
• knowledge
• self-assessment
• South Africa

Introduction

The fourth industrial revolution (4IR) concept is relatively new in the knowledge space. Due to global gaps in development and resources, the levels of understanding of 4IR within a country, region, and continent differ across various sectors and population groups. It is evident from recent studies that the 4IR is a reality, and its effects of societal disruptions are continuously being realized (Brits et al. 2017; Deloitte 2018; Naude 2017; PWC 2016; Schwab 2016). This is supported by Coleman (2016), who argued that the 4IR seriously impacts both society and business as it is grounded on digital transformation.

The 4IR is shown to stimulate important advances in science and technology through its use of big data analytics, artificial intelligence (AI), Internet of Things (IoT), and other supporting technologies to develop customized customer relationships, optimize production chains and improve efficiencies (Liao et al. 2017; Schwab 2016). These advances are not confined to a business or a particular industry; they involve how we work, live and interact with one another. It is synonymous with agility, intelligence, and networking.

As such, efforts are ongoing to ensure that South Africa and other countries worldwide are ready to seize the opportunities associated with this revolution while mitigating the inherent challenges. These opportunities include increased business opportunities and shaping the new and emerging landscape of flexible working to rebalance the gender divide (World Economic Forum 2016). Other inherent opportunities include products-as-services as well as exports where small-scale manufacturers on the African continent can become more efficient and competitive because of the use of the 4IR technologies (Naude 2017). This author explained that for countries to seize these opportunities, they need to develop and enhance new business models of bringing goods and services to consumers.

It is clear that South Africa and Africa cannot let these opportunities slip past as they offer an opportunity to re-industrialise, considering the country’s vast wealth of natural resources and its population profile as one of the youngest in the world, among other factors. This is particularly important as African markets are underdeveloped but have substantial potential within Africa’s geography, demography, and ongoing urbanization (Naude 2017).

As a point of departure, it is critical to understand citizens’ knowledge of the 4IR. This is important as knowledge is understanding, and understanding leads to participation and action. As such, the study’s primary research question is ‘What is the level of the South African citizens’ self-assessed knowledge about the fourth industrial revolution?’

Advent of the fourth industrial revolution

The 4IR follows the previous three revolutions (Baldassari and Roux 2017; Naude 2017; Schwab 2016). The first industrial revolution was synonymous with mechanized production from the conversion of water to steam. During this time, the world saw the advent of the steam engine, which powered mainly agriculture and textile manufacturing. Flowing from the first industrial revolution was the second industrial revolution, which saw the advent of science and mass production. This revolution resulted in inventions such as chemical fertilizers, diesel and petrol engines, and aeroplanes, which were key for economic development and growth. This revolution also saw the use of electricity as a source of power. In the third industrial revolution, inventions were associated with the digital revolution and the information age (Blinder 2006). During this time, the world moved from analogue, electronic and mechanical devices to digital technology. There was also the start of automation of production and a global supply chain utilizing electronics and information technology.

In the past few years, the world has seen the emergence of the 4IR. The revolution is shown to stimulate critical advances in science and technology, to customized products for customers, optimized production chains and improved efficiencies (Liao et al. 2017). These include products-as-services, the sharing economy and digital services and exports. This creates opportunities for markets that are currently underdeveloped in Africa but have substantial potential, within the context of Africa’s geography, demography and ongoing urbanization (Naude 2017). This offers a further opportunity for re-industrialisation in Africa, which can assist in making the continent more competitive.

The opportunities offered by the 4IR are possible if countries around the world have an integrated and comprehensive response (Schwab 2016). This can be achieved by involving global polity stakeholders, such as civil society, academia, and the private and public sectors. This is because all these stakeholders have a significant role to play in shaping and driving success during the 4IR.

Relevance of knowledge about the 4IR

The speed of change associated with the 4IR cannot be ignored as it brings a shift in power, wealth and knowledge (Xu, David, and Kim 2018). Knowledge is particularly important as it allows for understanding as well as awareness of a body of ideas that are gained through learning or experience (Agbedia 2013).

Knowledge about the 4IR is critical and relevant, as most knowledge enables understanding of the issues, i.e., mediating personal implications through analysis and aiding the development of possible solutions (Scaratti et al. 2017). This is against the backdrop of the 4IR, which is expected to have transformative effects on education, the workplace, healthcare and other socio-economic areas in society. With knowledge, there is a migration of developed knowledge to impactful knowledge, which is then actionable, as the incumbent of that knowledge takes action to deal with that knowledge’s realities (Sannino, Engeström, and Lemos 2016).

This knowledge is vital as it makes for sensitivity and consciousness, while practice is the actual action in performance, based on knowledge and attitude (Agbedia 2013). As such, knowledge, attitude, and practice influence each other and can be regarded as a chain of action (Agbedia 2013). Increased knowledge may then be the catalyst and motivator for engaging with the 4IR. Citizens must be prepared for this. Citizens and businesses alike need to leverage opportunities that are brought by such increased productivity, which would result in new learning, increased markets and reach for products and services (Schwab 2016) and to mitigate the threats that are associated with it, such as job loss (Wanberg, Kanfer, Hamann, and Zhang 2016).

4IR in African countries

The World Economic Forum conducted a study in 100 countries, which represent 96% of the global gross domestic product (GDP) and manufacturing value added (MVA) and rated them on their state of readiness for future production assessment (World Economic Forum 2017a). Of the 25 countries classified as leading in structure and drivers of production, with the highest readiness for future production, there was not a single African country, with most countries concentrated in Europe, North America and East Asia. Together with Argentina and Brazil, South Africa was regarded as one of the least ready countries within the G20 nations (World Economic Forum 2017). This underpins the work ahead for South Africa and the other African countries to effectively leverage the opportunities associated with the 4IR while working against time to ensure that the adverse effects are kept under control.

Despite this, there are technologies of the 4IR which have been effectively used in Africa to save lives and improve people’s well-being. Noticeably, Rwanda and Tanzania employ unmanned aerial vehicles (UAVs), known as drones, to deliver medication in remote country areas (Hotz 2017). In Zanzibar, drones are used to map out the habitats of the mosquitoes that carry the malaria parasite, and this initiative had resulted in malaria prevalence levels dropping from 40% to less than 1% (The Conversation 2017). In addition, drones are used in the mining and agricultural sectors. In the Northern Cape, South Africa, Kumba Iron Ore is using drones to optimize the surveying process to increase coverage and improve reach, including small, constricted areas (Mining Weekly 2021). The major spin-off of this is improved safety for this type of operation in the mine. These drones have also seen increased inspection of hazardous areas, such as cell towers and wind turbines. PwC argues that drones have revolutionized the agricultural sector, focusing on soil and field analysis, planting, crop spraying, crop monitoring, irrigation and health assessment (PwC 2016).

David Meads believes that the 4IR will impact the economy and has the potential to drive Africa forward, with the hope that the delivery period will be shortened and improved (World Economic Forum 2017b). Mawasha (2017) is hopeful that this will enable innovation within the economy and propel new business models. The development of digital skills can achieve this through public-private partnerships as powerful levers for development, since industries being disrupted digitally require an effective response (Mawasha 2017).

Methodology

This study on citizens’ understanding and knowledge about the 4IR is part of a more extensive study on the factors and influences of an economically active citizen during the time of the 4IR. Ethical approval for the study was obtained from the University of Johannesburg Education Department Ethics Committee, South Africa (ref. no. 2018/157).

The study was conducted using a quantitative method by means of a survey questionnaire. This study is based on individual participant’s judgement and evaluation grounded on their self-knowledge. The survey quantified descriptions of the level of understanding, attitude, and extent of the uptake or use of disruptive technologies associated with the 4IR.

The population for the study were citizens of South Africa who were sampled using multi-stage sampling. In stage one, the nine provinces were proportionally sampled, according to population, with a minimum of 5% for the province with the lowest population. In stage two, the districts within provinces were sampled, of which 20 out of the 44 districts were sampled. In each district, a minimum of 30 participants was surveyed. A total sample size of 1111 was targeted in line with the guidelines for populations larger than 1 00 000 at ±3% precision level at 95% confidence (Tejada & Punzalan 2012). The final responses numbered 1105, Gauteng comprised 21.0%, followed by KwaZulu-Natal (14.5%), Limpopo with 12.4%, Western Cape and North West (both with 10.8%), Free State and Eastern Cape provinces (both 8.4%), Mpumalanga (8.1%), then Northern Cape (5.3%) as the least populated province. The collected data were analyzed using IBM Statistical Package for Social Sciences (SPSS) version 25.

The socio-demographic and socio-economic characteristics of the participants were tabulated with numbers (n) and percentages (%). Descriptive statistics such as mean, standard deviation, median, and quantiles were used to analyze the Likert scale and continuous variables. For the differences between the groups, independent t-test and analysis of variance (ANOVA) with post hoc test, Bonferroni (equal variance with unequal sample size) and Games-Howell (unequal variance and unequal sample size) (Shingala and Rajyaguru 2015) were used. Equality of variance was checked using Levene’s test, and this assumption, in some situations, was violated (p < .05), but this was tolerated as all the cells had over 30 responses (Pallant 2005).

The elements of the 4IR were Likert-type items (Boone and Boone 2012), and thus Mann Whitney U and Kruskal Wallis H were used to investigate the differences. A post hoc test with Bonferroni correction was used for pairwise evaluation of the significant groups (Armstrong 2014). For all significant analysis at 95%, confidence interval (p <.05), the effect size was calculated for the t-test using Cohen’s d while partial eta squared (η_p²) was used for ANOVA guidelines to assign the magnitude of the differences (and r for Mann Whitney U and Kruskal Wallis tests) (Onwuegbuzie and Leech 2007).

Results and discussion

Results

Socio-demographic and socio-economic profile of the participants

The socio-demographic factors were gender, age, and occupation group, while the socio-economic factors were the highest education level, employment status, household income, and access to internet data (Table 1).

Table 1: Socio-demographic and socio-economic profile of the participants.

Factors			Number of participants (n)	Percentage (%)
Socio-demographic factors	Gender	Male	612	55.4
		Female	493	44.6
		Total	1105	100
	Age	25 years or less (≤25)	274	24.8
		26–35 years (26–35)	471	42.6
		36–45 years (36–45)	237	21.4
		46–55 years (46–55)	83	7.5
		Older than 55 years (>55)	40	3.6
		Total	1105	100
Socio-economic factors	Occupational groups	Public sector (GVN)	203	18.4
		Private sector and funders (PVT)	220	19.9
		Entrepreneurs (ENT)	160	14.5
		Academics (ACD)	234	21.2
		Connectors of change (COC)	64	5.8
		Unemployed (UN)	158	14.3
		Other (OT)	66	6.0
		Total	1105	100
	Educational level	No formal education or Primary education (NPE)	89	8.1
		Secondary education (SE)	139	12.6
		Grade 12 (Matric) (MT)	368	33.3
		Diploma (DP)	232	21.0
		Degree (DG)	183	16.6
		Post degree (PD)	94	8.5
	Household* income	R550 pm or less (≤550)	148	13.4
		R551 pm–R1150 pm (551-1150)	107	9.7
		R1151 pm–R3500 pm (1.1K–3.5K)	237	21.5
		R3501 pm–R6500 pm (3.5K–6.5K)	216	19.6
		R6501 pm–R14000 pm (6.5K–14K)	203	18.4
		More than R14000 per month (>14K)	193	17.5
	Accessibility of data*	I don’t have access to data (I cannot afford it) (CAD)	230	20.8
		I spend R150 or less a month on data (150D)	392	35.5
		I spend more than R150 per month on data (>150D)	323	29.2
		I have access to uncapped data (UD)	160	14.5
	Employment status	Full-time employed (FTE)	587	53.1
		Part-time employed (PTE)	114	10.3
		Self-employed (SET)	135	12.2
		Unemployed (UE)	269	24.3

* R (ZAR) 1.00 = US$0.07

Gender and age

Of the total of 1105 particpants, the gender distribution profile comprised 55.4% males and 44.6% females. The highest number of participants were between 26–35 years with 42.6%, followed by the age category of 25 years or less with 24.8%. Of the participants in this study, 21.5% were in the 36–45 years age category. Participants older than 55 years recorded the lowest number of participants, with 3.6%.

Sectors

These participants were from the public and private sectors, academia, entrepreneurs and connectors of change, such as non-governmental organizations, public benefit organizations and funders. The highest groups represented were academia, and the private and public sectors with 21.2%, 19.9% and 18.4%, respectively. The least represented were connectors of change which comprised 5.8%.

Education levels

The leading group of participants’ highest level of education was grade 12 (matric) (33.3%), followed by participants with a diploma (21.0%). Of the participants, 16.6% had a degree as their highest level of education. The smallest group, only 8.1%, had primary education as their highest level of education.

Employment status, income, and spend

Concerning the participants’ employment status, 53.1% were employed full-time, followed by 24.3% unemployed. Of these participants, 10.3% indicated part-time employment, while 12.2% were self-employed. The largest group of participants in this study fall under the R1151pm–R3500pm (US$82.21–US$250.00) category with 21.5%, followed by the R3501 pm–R6500 pm (US$250.07–US$464.29) category with 19.6%. Of the participants, 18.4% earn between R6501 pm–R14000 pm (US$464.36–US$1000.00). In addition, 9.7% of the participants earned between R551 pm–R1150 pm (US$39.36–US$82.14), followed by 13.4% who earn R550 pm or less (US$39.29 or less).

The level of accessibility to data

Of the participants, 35.5% indicated that their monthly spend on data never exceeded R150, while 29.2% reported exceeding R150 per month. Furthermore, 20.8% of the participants reported no access to data or indicated that they could not afford it. Only 14.5% of the participants reported access to uncapped data.

Overall knowledge about the 4IR

Table 2 presents the overall knowledge of the 4IR. Only 1% of the participants were highly knowledgeable about the 4IR. A further 5% of participants ranked themselves at a score of 9 out of 10 in overall knowledge of the 4IR. A quarter of the participants rated themselves with 70% knowledge about the 4IR. In comparison, half of the participants rated themselves at 5 out of 10 in overall knowledge of the 4IR.

Table 2: The score distribution of self-knowledge assessment about the 4IR.

Mean and median		Quantile
N	1057
Mean	5,01	25th	3.00
Median	5,00	50th	5.00
Std. Deviation	2,496	75th	7.00

Further analysis was conducted to ascertain if there were differences in the knowledge of the 4IR within the socio-demographic and socio-economic factors groups.

Analysis of the differences between males and females is presented in Table 3. All test assumptions were met, with Leven’s test showing the equality of variance (F = 2.71, p = 0.100). An independent t-test was performed to assess a difference in the knowledge about the 4IR between the male and female participants. The results show that there was a significant difference in the knowledge scores with males having higher knowledge scores (M = 5.20, SD = 2.547) than females (M = 4.74, SD = 2.412); t (1071) = 3.02, p = 0.003, d = 0.187. Since the difference was significant, Cohen’s d, the standardized measure of the effect size, was utilized to understand the magnitude of mean differences between the knowledge of males and females. Based on the results, even though the t-test has shown a difference, these differences are small (d = 0.19).

Table 3: Independent t-test with effect sizes on the differences between genders on knowledge.

Levene’s Test for Equality of Variances		t-test for Equality of Means							Cohen’s d (d)
F	Sig.	t	df	Sig. (2-tailed)	Mean Difference	Std. Error Difference	95% Confidence Interval of the Difference
							Lower	Upper
2,712	0,100	3,017	1071	0,003	0,460	0,153	0,161	0,759	0,19

Cohen’s d: small = 0.2 medium = 0.5 large ≥ 0.8

Analysis of variance showed the main effects of level of knowledge about the 4IR F (4, 1052) = 4.30, p = .002, η_p² = .016. Post hoc analysis using Games-Howell indicates that the knowledge for respondents who were 25 years and younger was less than those aged between 36–45 years. In addition, the knowledge for respondents older than 55 years was lower than both the respondents who were aged 36–45 years and those aged 26–35 years (Table 4). Individuals most knowledgeable about the 4IR were between the age of 36–45 years. This is not surprising as 70% of them reported being employed full time, 76% have either matric or a tertiary qualification, and 86% reported having access to internet data.

Table 4: One-way ANOVA and Post-hocs with effect size on the differences between socio-demographic and socio-economic factors.

Source	Type III Sum of Squares	df	df (error)	Mean Square	F	Sig.	Partial Eta Squared (ηp^2)	Post hoc
Age	106	4	1052	26,5	4,30	0,002	0,016	(≤25) < (36-45); (>55) < (36-45); (>55) < (26-35) **
Occupational group	148	6		24,6	4,02	0,001	0,022	UN < GVN; UN < ACD*
Education level	249	5		49,9	8,28	0,000	0,038	NPE < PD; SE < DP; SE < DG; SE < PD; MT < DP; MT < PG*
Employment status	53,7	3		17,9	2,89	0,035	0,008	SET < FTE
Household income	277	5		55,4	9,24	0,000	0,042	1.1K-3.5K < 3.5K-6.5K; 1.1K-3.5K < 6.5K-14K; 1.1K-3.5K < (>14K) **
Level of access of internet data	483	3		161	27,8	0,000	0,073	CAD < 150D; CAD < (>150D); CAD < UD; 150D < (>150D) **

* - Bonferroni ** – Games-Howell η_p² – small = 0.01 medium = 0.06 large = 0.14

Knowledge of the elements of the 4IR

The participants were also asked about their knowledge of the 4IR’s five core elements: the Internet of Things (IoT), multilevel customer interaction, big data analytics and advanced algorithms, advanced human-machine interfaces and location detection technologies. Of the participants, 19.3% indicated that they are highly knowledgeable, while 41.1% indicated that they are knowledgeable about the IoT, 14.2% indicated they do not know about IoT, and 25.4% indicated that they have heard of it.

On multilevel customer interaction and customer profiling (MCICP), 35.4% of participants indicated that they are knowledgeable about multilevel customer interaction and customer profiling, followed by 33.6% who have only heard about it. There were only 6.4% highly knowledgeable participants, while 24.5% reported not knowing about MCICP.

On big data analytics and advanced algorithms (BDAAA), of the 1105 responses, there were 1090 completed responses to this question. Of these, 24.0% indicated that they are knowledgeable, while 3.9% indicated that they are highly knowledgeable about big data analytics and advanced algorithms. A total of 72.1% had either heard of it or did not know about BDAAA.

For advanced human-machine interfaces (AHMI), a total of 1099 out of 1105 responses were completed for this question.Of these, 43.5% of the participants indicated that they are knowledgeable about the advanced human-machine interfaces, followed by 27.1% who indicated that they have heard of these. Only 7.6% were highly knowledgeable, while 21.8% indicated they do not know about AHMI.

For location detection technologies (LDT), a total of 1083 out of 1105 participants completed this question. Figure 1 shows that a total of 47.6%, which is almost half of the participants, are knowledgeable about the location detection technologies, with a further 11.5% indicating that they are highly knowledgeable. Of those who completed this question,17.3% indicated they do not know about LDT, and 23.6% indicated they have heard of these.

Figure 1: Self-rating on knowledge of elements of 4IR.

Further analysis on the difference between males and females was evaluated for the five elements of the 4IR. The results show significant differences between the males and females for all five elements of the 4IR. The knowledge score about the IoT was greater for males (mean rank = 584) than females (mean rank = 518), U = 133522, z = −3.62, p < .001. The same pattern was found for MCICP (U = 131095, z = −3.93, p < .001), BDAAA (U = 132636, z = −3.00, p = .003), AHMI (U = 131051, z = −3.78, p < .001) and LDT (U = 134494, z = −2.29, p = .022), with knowledge of males higher than females. With all the results being significant, the effect size was evaluated to determine the magnitude of the difference, and all the differences between males and females were small (r = 0.07–0.12) (Table 5).

Table 5: Mann Whitney U and Kruskal Wallis H tests with effect sizes on the differences between gender on knowledge on elements of the 4IR.

Gender		N	Mean Rank	Sum of Ranks	Mann-Whitney U	Z	Asymp. Sig. (2-tailed)	Effect size (r)
IoT	Male	607	584	354507	133522	−3,62	0,000	0,11
	Female	500	518	258772
	Total	1107
MCICP	Male	603	585	352512	131095	−3,93	0,000	0,12
	Female	500	513	256345
	Total	1103
BDAAA	Male	594	570	338703	132636	−3,00	0,003	0,09
	Female	496	516	255892
	Total	1090
AHMI	Male	600	581	348649	131051	−3,78	0,000	0,11
	Female	499	513	255801
	Total	1099
LDT	Male	590	561	330722	134494	−2,29	0,022	0,07
	Female	493	520	256265
	Total	1083

A Kruskal Wallis H test shows a statistical difference in the knowledge level about the elements of the 4IR between the different age groups (Table 6). The results show a significant difference between IoT knowledge across the different age groups, χ²(4) = 17.3, p = 0.002, r = 0.016. Overall, younger participants had better knowledge than older ones. There was also significant difference on the knowledge of MCICP, χ²(4) = 30.2, p < 0.001, r = 0.029 and LDT, χ²(4) = 22.2, p < 0.001, r = 0.021. Based on the Bonferroni correction results, the most knowledgeable groups were aged 36–45 years and 26–35 years, with the least knowledgeable group being older than 55 years and 46–55 years. There were no significant differences between the level of knowledge for BDAAA, and AHMI.

Table 6: Kruskal Wallis H tests with effect sizes on the differences between genders on knowledge on elements of the 4IR.

		IoT	MCI	BDA	AHMI	LDT	Post hoc (Bonferroni correction)
	N	1091	1087	1074	1083	1067
Age	Median	3,00	2,00	2,00	3,00	3,00	IoT: (>55) < (≤25); (46–55) < (≤25) MCICP: (>55) < (≤25); (>55) < 36–45; (>55) < 26–35; 46–55 < 36–45; 46–55 < 26–35; (≤25) <26–35 LDT: (>55) < 36–45; (>55) < 26–35; 46–55 < 26–35; (≤25) < 26–35
	Chi-Square	17.3	30.2	9.03	7.77	22.2
	df	4	4	4	4	4
	Asymp. Sig.	0,002	0,000	0,060	0,100	0,000
	Effect size (r)	0.016	0.029			0.021

Focusing on the socio-economic factor of academics, which included tertiary students, early childhood development (ECD) educators, basic school educators and tertiary school-level educators, these participants had higher knowledge about IoT, with the entrepreneurs having the lowest knowledge. The same pattern of lowest knowledge of the occupation group is found for entrepreneurs in BDAAA, and LDT (Table 7).

Table 7: Kruskal Wallis H test on the differences between factors of elements of the 4IR.

Socio-economic factor		IoT	MCI	BDA	AHMI	LDT	Post hoc (Bonferroni correction)
Occupation group	N	1091	1087	1074	1083	1067	IoT: OT < PVT; OT < GVN; OT < ACD MCI: UN < ENT; UN < ACD; UN < GVN; UN < PVT; UN < COC; OT < ENT; OT < ACD; OT < GVN; OT < PVT; OT < COC BDA: OT < ACD; OT < GVN; UN < PVT; UN  < ACD; UN < COC; UN < GVN; ENT < GVN
	Median	3,00	2,00	2,00	3,00	3,00
	Chi-Square	13.406	46.864	27.053	7.378	5.842
	df	6	6	6	6	6
	Asymp. Sig.	0,037	0,000	0,000	0,287	0,441
	Effect size (r)	0.012	0.043	0.025	-	-
Education	N	1091	1087	1074	1083	1067	IoT: NPE < SE; NPE < MT; NPE < DG; NPE < DP; NPE < PD; MT < PD. MCI: SE < MT; SE < DP; SE < DG; SE < PD; NPE < MT; NPE < DP; NPE < DG; NPE < PD; MT < DP; MT < DG; MT < PD. BDA: NPE < MT; NPE < DG; NPE < DP; NPE < PD; SE < DG; SE < DP; SE < PD; MT < DG; MT < DP; MT < PD. AHMI: NPE < MT; NPE < DG; NPE < DP; NPE < PD; SE < DP; SE < PD. LDT: NPE < DG; NPE < DP; NPE < MT; NPE < PD; SE < DG; SE < DP; SE < MT; SE < PD.
	Median	3,00	2,00	2,00	3,00	3,00
	Chi-Square	21.897	72.877	106.075	15.597	21.254
	df	5	5	5	5	5
	Asymp. Sig.	0,001	0,000	0,000	0,008	0,001
	Effect size (r)	0.020	0.067	0.099	0.014	0.020
Employment	N	1091	1087	1074	1083	1067	IoT: SET < FTE MCI: UN < SET; UN < FTE; PTE < FTE BDA: PTE < FTE; UN < FTE. AHMI: UN < FTE; PTE < FTE. LDT: PTE < SET; PTE < UN; PTE < FTE.
	Median	3,00	2,00	2,00	3,00	3,00
	Chi-Square	10.368	57.867	24.398	10.563	10.562
	df	3	3	3	3	3
	Asymp. Sig.	0,016	0,000	0,000	0,014	0,014
	Effect size (r)	0.010	0.053	0.023	0.010	0.010
Household income	N	1090	1086	1073	1082	1066	IoT: (551–1150) < (1.1K-3.5K); (551–1150) < (3.5K-6.5K); (551–1150) < (>14K); (551–1150) < (6.5K-14K). MCI: (551–1150) < (6.5K-14K); (551–1150) < (>14K); (≤550) < (6.5K-14K); (≤550) < (>14K); (1.1K-3.5K) < (6.5K-14K); (1.1K-3.5K) < (>14K); (3.5K-6.5K) < (6.5K-14K); (3.5K-6.5K) < (>14K). BDA: (1.1K-3.5K) < (3.5K-6.5K); (1.1K-3.5K) < (6.5K-14K); (1.1K-3.5K) < (>14K); (551–1550) < (>14K); (≤550) < (6.5K-14K); (≤550) < (>14K); (3.5K-6.5K) < (>14K). AHMI: (551–1150) < (>14K); (551–1150) < (3.5K-6.5K); (551–1150) < (6.5K-14K); (1.1K-3.5K) < (>14K); (1.1K-3.5K) < (3.5K-6.5K); (1.1K-3.5K) < (6.5K-14K); (≤550) < (>14K); (≤550) < (3.5K-6.5K); (≤550) < (6.5K-14K) LDT: (1.1K-3.5K) < (3.5K-6.5K); (1.1K-3.5K) < (6.5K-14K); (1.1K-3.5K) < (>14K); (≤550) < (6.5K-14K); (≤550) < (6.5K-14K)
	Median	3,00	2,00	2,00	3,00	3,00
	Chi-Square	11.531	88.692	86.137	11.272	12.311
	df	5	5	5	5	5
	Asymp. Sig.	0,042	0,000	0,000	0,046	0,031
	Effect size (r)	0.011	0.082	0.080	0.010	0.012
Level of internet access	N	1091	1087	1074	1083	1067	IoT: CAD < 150D; CAD < (>150D); CAD < UD; 150D < UD; (>150D) < UD; MCI: CAD < 150D; CAD < (>150D); CAD < UD; 150D < (>150D); 150D < UD; CAD < 150D; CAD < (>150D); CAD < UD; 150D < UD; AHMI: CAD < 150D; CAD < (>150D); CAD < UD; 150D < (>150D). LDT: CAD < (>150D); 150D < (>150D); UD < (>150D)
	Median	3,00	2,00	2,00	3,00	3,00
	Chi-Square	75.273	126.711	64.719	23.567	33.098
	df	3	3	3	3	3
	Asymp. Sig.	0,000	0,000	0,000	0,000	0,000
	Effect size (r)	0.069	0.117	0.060	0.022	0.031

With education, there was an increase in knowledge with an increase in education level. The post-graduates had more knowledge that is significantly different to those with primary or no formal education. The knowledge for the different income groups was significant for the level of internet data access for all the five elements of the 4IR (IoT, MCICP, BDAAA, AHMI, LDT). The same pattern was also found with the citizens with better income and those who have more than R150 (US$10.72) (1 GB) data access or uncapped internet data having higher knowledge than those in lower economic positions. Across all socio-economic factors, the effective size was small, with r = (0.031–0.069). This means that the magnitude of the difference is low across the different socio-economic groups.

Discussion

The advent of the 4IR marks a new important journey for the world, especially developing countries in Africa, and specifically for South Africa in the context of factors considered in the current study, which explored the understanding and knowledge of South African citizens about the 4IR. This was important as there is a general agreement that the emergence of 4IR is a reality, and it will increasingly influence the way we work, live and interact with one another.

The findings reveal that the average knowledge score is 5 out of 10, with only a quarter of the participants assessing themselves to have 7 out of 10 knowledge about the 4IR. This means that only a quarter of the participants had adequate knowledge as the literature indicates that 70%–80% is the threshold for acceptable knowledge during evaluation (Naidoo et al. 2014; Ogunrin, Daniel, and Ansa 2016; Sodano et al. 2015).

Within the 4IR elements, location detection technologies and the IoT were best known, while the BDAAA and MCICP were the least known by participants. Knowledge of these elements will assist citizens in seizing the associated opportunities to buffer the adverse effects of large-scale job losses anticipated during the 4IR (Shanks and Parr 2016). For example, there was a market expectation increase in the IOT that is expected to be worth US$7 trillion by 2020 (Bloomberg 2016). The number of connected devices was expected to double to an estimated 50 billion in 2020 from 2015, with 25 billion customers (Hatzakis 2016).

The higher knowledge levels were biased towards the participants with higher socio-economic status: education, employment, household income, and access to internet data (Figure 2). The educated citizens had more knowledge of the 4IR, including the 4IR elements. Participants with the most knowledge about the 4IR have post-degree qualifications, followed by those with diplomas and degrees. Those who have primary education or no formal education have a lower level of understanding and knowledge than those who are more educated.

Figure 2: Means plots of the socio-economic factor group on understanding of the 4IR.

This confirms the views of Badsha and Cloete (2011) that education levels are critical to the design and production of new technologies. This help with a country’s innovative capacity and societal development. This is not surprising as the level of education for individuals in society is aligned to socio-economic factors and overall well-being (Economic and Social Research Council: UK 2014). These factors such as social status, employment and income tend to increase as an individual’s level of education increases, which is known as the Education Effect. International Organisation of Employers (2017) posit that skilled workers generate knowledge that can be used to create and implement innovations. Educated workers also have a better start for the acquisition of additional skills.

Employment status showed differences within the level of knowledge. Participants employed full-time were the most knowledgeable about the 4IR, in contrast to those unemployed. The unemployment knowledge is driven by tertiary institution students who are unemployed graduates. The participants who were employed part-time indicated that they do not have much knowledge, while the slightest knowledge and understanding about the 4IR was reported by the self-employed. The overall low knowledge, especially by entrepreneurs and the unemployed (non-students), who are targeted for economic emancipation, is alarming. The 4IR is supposed to promote entrepreneurship, considering that the 4IR holds opportunities for skills and innovation associated with entrepreneurship (Naude 2017).

The income levels and access to internet data based on affordability show that higher knowledge of the 4IR was biased towards citizens with higher economic status. Participants with higher knowledge of the 4IR either had access to uncapped data or spent more than R150 per month on internet data. Participants who had the least knowledge were those who could not access data or could not afford it. The results indicated that those whose monthly data spend which does not exceed R150 also had a lower understanding of the 4IR than those who have uncapped internet data or whose monthly data spend exceeds R150. This finding resonates with the most commonly used measure of inequality, the Gini coefficient, which measures household income distribution within economies (World Bank n.d.). Knowledge of 4IR is highly segregated, with people with a high income having more knowledge.

In the final analysis, there is an urgent requirement to develop adequate knowledge about the 4IR so that this knowledge can then be the catalyst and motivator to prepare the citizens for it and for them businesses to leverage opportunities. This is possible as Agbedia (2013) explains that knowledge, as well as attitude and practice, tend to be a chain of action that influence each other.

Research strength and limitations

This is the first study in South Africa that evaluated the understanding and knowledge of citizens, on a large scale, about the 4IR. The strength of this study is that it covered all the provinces of South Africa and obtained balanced perspectives across the different socio-demographic and socio-economic clusters of society. Despite this, the study included only about half of the districts in South Africa and thus will have limited generalizability.

It was also worth noting that the study was conducted using a self-assessment of knowledge and, as such, answering in a socially desirable way cannot be wholly ruled out despite several studies, such as those by Sundström (2005), showing that this is only possible when the respondent had a higher stake, such as promotion or selection. This was not the case in this study. The use of self-assessment as a measurement tool has gained popularity and has been found to be an effective approach to assessment (Butler and Lee 2010; Gholami et al. 2011; Nelms 2015). Findings show that the self-assessment compared well with the external instrument assessment for knowledge (Sundström 2005).

This study worked on the basis that the participants could understand and know their current levels of knowledge about the 4IR. This approach is supported by the studies of Clauss and Geedey (2010), who found a correlation between self-assessment and examination scores, with no evidence of variation between these two assessment methods. This was also confirmed by a study by Hofer (2010), which found a statistically significant positive correlation between self-perceived and actual knowledge.

Conclusion

These findings revealed that knowledge about the 4IR is inadequate and requires enhancement. Of concern in these findings, is the inequality of knowledge with higher knowledge levels found among participants with higher socio-economic status. This further entrenches the challenges of inequality in South Africa, which is classified as the most unequal country, with a Gini coefficient of 0.63 (Statistics South Africa 2019). The 4IR is critical and should be on all nations’ national agenda. Central to this agenda should be how all the citizens must participate in it without perpetuating inequality within society.

The implication of the study is that there is a need for all stakeholders to collaborate to develop citizen’s knowledge about 4IR, assist them in identifying the associated opportunities and prepare them to mitigate the challenges associated with it. Stakeholder collaboration in this context would include initiatives such as hosting conversations on varied sectors, with different categories of population groups or different demographics, toward promoting an inclusive 4IR.

Thus, there is also a need to customize knowledge sharing by different levels of population groups to increase degrees of understanding and uptake. Increasing knowledge levels of 4IR and related disruptive technologies will facilitate the understanding of and interest in grasping opportunities brought about by this revolution and, most importantly, guide choices for future skills (careers) and jobs.

Declaration

The contents are based on reflections from research commissioned and published by Kagiso Trust. As such, the authors of this article are publishing on behalf of Kagiso Trust.

Disclosure statement

No potential conflict of interest was reported by the authors.