Sharing economy platforms as mainstream: balancing pro-social and economic tensions

Figures & data

Figure 1. Theoretical framework.

Table 1. Measurement items of the constructs.

Item (original items adapted from)	N	Mean	SD	Skewness	Kurtosis	Loading
Sharing orientation (Akbar et al., 2016)
I would try to share rides even if the CG* did not exist.	524	5.39	1.804	−2.734	−0.170	0.442
I think that it is important to share.	634	5.92	1.273	−1.262	1.563	0.925
I plan to buy less and to share with others instead.	564	5.13	1.743	−0.740	−0.371	0.827
Trend orientation (Moeller & Wittkowski, 2010)
Carpooling is a new and unique phenomenon for me.	603	3.63	2.129	−2.931	0.024	0.589
It is important that a carpooling platform looks modern.	583	4.84	1.834	−0.568	−.0646	0.686
It is important to follow the latest lifestyle trends.	607	3.00	1.886	0.667	−0.633	0.886
Platform authenticity (Napoli et al., 2014)
The CG represents an authentic sharing culture.	500	5.51	1.430	−0.808	0.207	0.903
The CG is loyal to its original values.	249	5.62	1.333	−0.827	0.062	0.953
Grassroots engagement (own measures)
It is important to me that the CG is a participatory movement.	573	5.40	1.717	−0.991	0.145	0.728
It is important to me as a participant to get involved in the CG movement.	563	3.21	1.877	0.490	−0.849	0.906
It is important to me as a participant to get involved in the non-profit association the CG.	563	2.79	1.807	0.762	−0.472	0.883
Attitude towards the CG platform (Hawlitschek et al., 2018b)
Participating in the CG is a good idea.	606	6.16	1.207	−1.649	2.765	0.831
I like to participate in the CG.	479	5.36	1.597	−0.641	−0.554	0.872
Participating in the CG is nice.	425	5.34	1.540	−0.565	−0.539	0.899
Norms regarding the CG platform (Hawlitschek et al., 2018b)
Those who care about me think I should join the CG.	253	3.70	2.154	0.161	−1.297	0.982
Those who care about me prefer me to participate in the CG.	240	3.45	2.166	0.354	−1.258	0.980
I feel an obligation to share rides.	593	3.41	2.025	0.357	−1.085	0.625
Perceived behavioural control of using the CG platform (Hawlitschek et al., 2018b)
I think it is easy to participate in the CG.	474	4.03	1.869	0.063	−1.067	0.848
Participating in the CG is within my control.	482	4.74	1.892	−0.499	−0.856	0.891
I have the resources, knowledge and abilities to participate in the CG.	535	5.20	1.783	−0.823	−0.292	0.817
Behavioural intention to use the CG platform (Hawlitschek et al., 2018b)
I intend to participate in the CG in the future.	522	5.44	1.492	−0.868	0.265	0.903
I will always try to participate in the CG.	458	4.76	1.875	−0.472	−0.868	0.890
I plan to participate in the CG as often as I can.	519	4.58	1.895	−0.334	−0.932	0.862

Note: Items were measured on a seven-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree).

*CG = Carpool Group

Table 2. Sample demographics.

Gender	N	%	Age (average):		42.16
Man	260	39.7
Woman	366	55.9	Education
Other	18	2.7	(average no. of years):		16.13
Not available	11	1.7
Urban population (no. of inhabitants)	N	%	Occupation	N	%
			Employee	388	59.2
			Self-employed	59	9.0
1–500	47	7.2	Student	78	11.9
501–3000	49	7.5	Housewife/-husband	1	0.2
3001–8000	33	5.0	On parental leave	8	1.2
8001–25,000	61	9.3	Retired	46	7.0
25,001–100,000	129	19.7	Jobless	36	5.5
More than 100,000	316	48.2	Sick	39	5.9
Not available	20	3.1	Not available	–	–
Household composition	N	%	Household income	N	%
Alone, without children	232	35.4	Less than 2400€	150	22.9
Alone, with 1 child	28	4.3	2400€–3400€	114	17.4
Alone, with 2 children	19	2.9	3500€–4400€	107	16.3
Alone, with 3 + children	5	0.8	4500€–5400€	50	7.6
With someone, without children	223	34.0	5500€–6400€	46	7.0
With someone, with 1 child	63	9.6	6500€–7400€	50	7.6
With someone, with 2 children	50	7.6	7500€–8400€	30	4.6
With someone, with 3 + children	35	5.3	More than 8400€	38	6.5
Not available	–	–	Not available	70	10.7

Table 3. Measurement model: reliability and validity.

Construct	Indicators	Valid N	Mean	SD	CA	ρC	ρA	AVE
Sharing orientation	3	465	5.457	1.013	0.575	0.781	0.696	0.563
Platform authenticity	2	247	5.638	0.997	0.858	0.929	0.888	0.867
Trend orientation	3	533	3.878	1.004	0.557	0.766	0.636	0.530
Grassroots engagement	3	504	3.831	1.001	0.786	0.876	0.894	0.710
Attitude towards the CG	3	408	5.700	0.996	0.844	0.904	0.850	0.759
Norms regarding the CG	3	226	3.578	1.010	0.813	0.900	0.928	0.756
Perceived behavioural control of using the CG	3	431	4.688	0.995	0.825	0.892	0.832	0.734
Behavioural intention to use the CG	3	419	5.031	0.993	0.878	0.920	0.880	0.794

Note: Constructs assessed using: CA = standardized Cronbach’s Alpha (reliability measure limited by the assumption that all indicators are equally reliable); ρC = composite coefficient (reliability measure prioritising indicators according to their individual reliability and most-suited to PLS-SEM); ρA = approximately exact (or consistent) coefficient (reliability measure considered as good compromise as it lies between CA and ρC); and AVE = Average Variance Extracted (convergent validity). In addition, discriminant validity was established the heterotrait-monotrait (HTMT) ratios which are all below 0.61 and the confidence intervals (estimated using the bootstrapping procedure) exclude 1 (see Appendix 1). There are no multicollinearity issues to report (see Appendix 2).

Figure 2. Path estimates and statistical significance.

Note: *** p < 0.001; ** p < 0.01; * p < 0.05 (two-tailed); dotted paths illustrate non-statistically significant relationships. The percentage of explained variance (i.e. R²) can be considered substantial for behavioural intention to use the CG (51.91%), moderate for attitude towards the CG (35.85%) and norms (48%), and weak for perceived behavioural control (23.39%). The Q²_predict values (obtained from the PLSpredict procedure) indicate that the model has high predictive relevance (see also Appendix 3).

Table 4. Structural model (path analysis).

Direct effects	β	f^2	t	p	CI [2.5%−97.5%]
Sharing Orientation → Attitude	0.238	0.070	3.000	0.003	0.072	0.381
Sharing Orientation → Norms	0.192	0.055	2.193	0.028	0.012	0.355
Sharing Orientation → Perceived Behavioural Control	0.000	0.000	0.001	0.999	-0.185	0.157
Sharing Orientation → Behavioural Intention	-0.088	0.012	1.080	0.280	-0.251	0.056
Platform Authenticity → Attitude	0.336	0.120	3.583	0.000	0.158	0.528
Platform Authenticity → Norms	0.023	0.001	0.219	0.826	-0.184	0.236
Platform Authenticity → Perceived Behavioural Control	0.352	0.111	3.396	0.001	0.156	0.557
Platform Authenticity → Behavioural Intention	0.105	0.013	0.838	0.402	-0.148	0.339
Trend Orientation → Attitude	-0.022	0.001	0.419	0.675	-0.129	0.078
Trend Orientation → Norms	0.301	0.146	4.395	0.000	0.158	0.427
Trend Orientation → Perceived Behavioural Control	0.081	0.007	1.403	0.161	-0.037	0.190
Trend Orientation → Behavioural Intention	-0.071	0.008	0.800	0.424	-0.248	0.082
Grassroots engagement → Attitude	0.211	0.059	4.339	0.000	0.115	0.306
Grassroots engagement → Norms	0.368	0.213	5.255	0.000	0.231	0.506
Grassroots engagement → Perceived Behavioural Control	0.165	0.030	3.014	0.003	0.058	0.271
Grassroots engagement → Behavioural Intention	0.012	0.000	0.141	0.888	-0.153	0.148
Attitude → Behavioural Intention	0.228	0.060	2.436	0.015	0.040	0.406
Subjective norms → Behavioural Intention	0.455	0.206	2.725	0.006	0.186	0.767
Perceived Behavioural Control → Behavioural Intention	0.237	0.081	2.975	0.003	0.081	0.388
Control: Income → Perceived Behavioural Control	0.137	0.021	2.526	0.012	0.030	0.242
Control: Age → Attitude	-0.121	0.020	2.417	0.016	-0.223	-0.024
Control: Gender → Perceived Behavioural Control	-0.111	0.014	2.094	0.036	-0.208	0.002

Note: β = standardized path coefficient estimates. f² = effect sizes; considered small above 0.02, medium above 0.15 and large above 0.35. CI = 95% confidence intervals (bias corrected) of path coefficients (estimated using the bootstrapping procedure).

Table

	01	02	03	04	05	06	07	08
01 Sharing Orientation		0.437–0.747	0.154–0.343	0.217–0.421	0.457–0.690	0.284–0.588	0.150–0.399	0.232–0.491
02 Platform Authenticity	0.590		0.193–0.447	0.190–0.400	0.417–0.633	0.139–0.402	0.269–0.505	0.219–0.476
03 Trend Orientation	0.214	0.305		0.275–0.537	0.111–0.336	0.398–0.692	0.160–0.410	0.182–0.415
04 Grassroots Engagement	0.298	0.294	0.404		0.363–0.535	0.391–0.637	0.256–0.460	0.350–0.528
05 Attitude	0.573	0.526	0.215	0.450		0.381–0.604	0.440–0.624	0.534–0.696
06 Norms	0.422	0.266	0.543	0.513	0.494		0.293–0.554	0.501–0.729
07 P. Behav. Control	0.254	0.389	0.274	0.360	0.533	0.421		0.463–0.641
08 Behav. Intention	0.361	0.350	0.290	0.440	0.616	0.617	0.554

Note: HTMT values are given below the diagonal. Bias-corrected confidence intervals [2.5%–97.5%] are given above the diagonal – estimated using the bootstrapping procedure (10,000 samples).

Table

	01	02	03	04	05	06	07	08
01 Sharing Orientation					1.308	1.308	1.308	1.447
02 Platform Authenticity					1.465	1.465	1.465	1.747
03 Trend Orientation					1.183	1.183	1.183	1.383
04 Grassroots Engagement					1.197	1.197	1.197	1.469
05 Attitude								1.826
06 Norms								2.163
07 P. Behav. Control								1.456
08 Behav. Intention

Note: The VIF values, obtained from the PLS algorithm, are all well below the conservative threshold of 3, thus there are no multicollinearity issues to consider.

Table

	Q²predict	PLS RMSE	LM RMSE	PLS MAE	LM MAE
I intend to participate in the CG in the future.	0.072	1.284	1.287	0.971	0.975
I will always try to participate in the CG.	0.120	1.472	1.477	1.121	1.134
I plan to participate in the CG as often as I can.	0.120	1.584	1.586	1.243	1.249

Note: Compared to the linear regression model (LM) scores, PLS-SEM scores yield lower prediction errors in terms of Root Mean Squared Error (RMSE) – and but also the Mean Absolute Error (MAE) considered a more relaxed measure – for all indicators, such as the model can be said to have high predictive power.

Table

	01	02	03	04	05	06	07	08
01 Sharing Orientation					0.072–0.381	0.012–0.355	-0.185–0.157	-0.124–0.210
02 Platform Authenticity					0.158–0.528	-0.184–0.236	0.156–0.557	0.042–0.501
03 Trend Orientation					-0.129–0.078	0.158–0.427	-0.037–0.190	-0.035–0.182
04 Grassroots Engagement					0.115–0.306	0.231–0.506	0.058–0.271	0.161–0.364
05 Attitude	0.238	0.336	-0.022	0.211				0.040–0.406
06 Norms	0.192	0.023	0.301	0.368				0.186–0.767
07 P. Behav. Control	0.000	0.352	0.081	0.165				0.081–0.388
08 Behav. Intention	0.054	0.276	0.080	0.266	0.228	0.455	0.237

Note: The total effects (given below the diagonal) on Behavioural intention to use the CG platform, obtained from the PLS algorithm, are considered medium for Attitude towards the CG platform and Perceived behavioural control of using the CG platform, and large for Norms. Bias-corrected confidence intervals [2.5%–97.5%] are given above the diagonal, obtained using the bootstrapping procedure (10,000 samples).

Akbar, P., Mai, R., & Hoffmann, S. (2016). When do materialistic consumers join commercial sharing systems. Journal of Business Research, 69(10), 4215–4224. https://doi.org/10.1016/j.jbusres.2016.03.003

 Web of Science ®Google Scholar

Moeller, S., & Wittkowski, K. (2010). The burdens of ownership: Reasons for preferring renting. Managing Service Quality: An International Journal, 20(2), 176–191. https://doi.org/10.1108/09604521011027598

 Google Scholar

Napoli, J., Dickinson, S. J., Beverland, M. B., & Farrelly, F. (2014). Measuring consumer-based brand authenticity. Journal of Business Research, 67(6), 1090–1098. https://doi.org/10.1016/j.jbusres.2013.06.001

 Web of Science ®Google Scholar

Hawlitschek, F., Teubner, T., & Gimpel, H. (2018b). Consumer motives for peer-to-peer sharing. Journal of Cleaner Production, 204, 144–157. https://doi.org/10.1016/j.jclepro.2018.08.326

 Web of Science ®Google Scholar