The Use of Neutralisation Techniques in the Context of Responsible Online Shopping : A Latent Profile Analysis

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY 4.0 https://creativecommons.org/licenses/by/4.0/ The Use of Neutralisation Techniques in the Context of Responsible Online Shopping : A Latent Profile Analysis © 2024 Unverza v Mariboru, Univerzitetna založba Published version Makkonen, Markus; Frank, Lauri; Holkkola, Matilda; Paananen, Tiina Makkonen, M., Frank, L., Holkkola, M., & Paananen, T. (2024). The Use of Neutralisation Techniques in the Context of Responsible Online Shopping : A Latent Profile Analysis. In A. Pucihar, M. Kljajić Borštnar, S. Blatnik, R. W. H. Bons, K. Smit, & M. Heikkilä (Eds.), 37th Bled eConference : Resilience Through Digital Innovation : Enabling the Twin Transition (pp. 693710). University of Maribor Press. https://doi.org/10.18690/um.fov.4.2024.41 2024 DOI https://doi.org/10.18690/um.fov.4.2024.41 ISBN 978-961-286-871-0 THE USE OF NEUTRALISATION TECHNIQUES IN THE CONTEXT OF RESPONSIBLE ONLINE SHOPPING: A LATENT PROFILE ANALYSIS Keywords: responsible consumption, responsible online shopping, neutralisation techniques, anticipated guilt, latent profile analysis M ARKUS M AKKONEN ,1, 2 L AURI F RANK ,2 MATILDA HOLKKOLA,2 TIINA PAANANEN2 1 Tampere University, Faculty of Management and Business, Tampere, Finland markus.m[email protected] 2 University of Jyvaskyla, Faculty of Information Technology, Jyvaskyla, Finland [email protected], [email protected], [email protected], tiina.e.[email protected] Although many consumers use various neutralisation techniques to eliminate the anticipated guilt that results from not engaging in responsible consumption, the use of such techniques in the context of responsible online shopping has attracted little attention in prior research. In this study, we aim to address this gap by examining (1) whether it is possible to segment consumers in terms of their use of neutralisation techniques to eliminate the anticipated guilt that results from not engaging in responsible online shopping and (2) how these segments potentially differ from each other in terms of demographics (e.g., gender, age, and income), online shopping frequency, and anticipated guilt. The examination is based on 478 responses from Finnish consumers that were collected in spring 2023 and are analysed with latent profile analysis. Our findings suggest the existence of four distinct consumer segments with several differences between them in terms of demographics and anticipated guilt. 694 37 TH B LED E C ONFERENCE : R ESILIENCE T HROUGH D IGITAL I NNOVATION : E NABLING THE T WIN T RANSITION 1 Introduction Today, more and more consumers are engaging in responsible consumption, which refers to consumption that has a less negative or more positive impact on the environment, society, self, and others (Ulusoy, 2016). Because of this, it is not surprising that responsible consumption has attracted more and more attention also in academic research (Nangia et al., 2024) and has been predicted to remain a prominent research topic also concerning the consumption environments of tomorrow, such as the novel metaverse marketplaces (Pellegrino et al., 2023) that can be seen as digitally mediated spaces that immerse users in shared, real-time experiences (Hadi et al., 2024). According to prior studies (e.g., Onwezen et al. 2013, 2014a, 2014b; Antonetti & Maklan, 2014a, 2014b; Theotokis & Manganari, 2015; Lindenmeier et al., 2017), one main driver for consumers to engage in responsible consumption is their anticipated guilt, which refers to the feelings of guilt that arise from contemplating a potential deviation from one’s standards (Rawlings, 1970), such as engaging in consumer behaviour that cannot be considered responsible. However, despite this driver, there are still many consumers who do not commonly engage in responsible consumption, for which one explanation may be the various neutralisation techniques suggested in the neutralisation theory by Sykes and Matza (1957) that consumers may use to eliminate their anticipated guilt. Prior studies (e.g., Strutton et al., 1994; Chatzidakis et al., 2007; McGregor, 2008; Antonetti & Maklan, 2014b; Gruber & Schlegelmilch, 2014) have shown the use of such techniques among consumers to be relatively common. However, their use in the specific context of online shopping has attracted little attention in prior information systems (IS) and marketing research. In this study, we aim to address the aforementioned gap in prior research. More specifically, in order to differentiate the study from prior studies on the topic, our objective is to focus less on the potential effects of the use of neutralisation techniques on other constructs, such as anticipated guilt (cf. Makkonen et al., 2023), and more on the precise use patterns of neutralisation techniques among consumers. As such, we examine (1) whether it is possible to segment consumers in terms of their use of neutralisation techniques to eliminate the anticipated guilt that results from not engaging in responsible online shopping and (2) how these segments potentially differ from each other in terms of demographics (e.g., gender, age, and income), online shopping frequency, and anticipated guilt. The examination is based on 478 responses from Finnish consumers that were M. Makkonen et al.: The Use of Neutralisation Techniques in the Context of Responsible Online Shopping: A Latent Profile Analysis 695 collected in spring 2023 and are analysed by using latent profile analysis (cf. Ferguson et al., 2020) as the main analysis method. 2 Theoretical Foundation Table 1: Neutralisation techniques examined in the study Name Reference Description Denial of responsibility (DOR) Sykes & Matza (1957) Claiming not to be responsible for the deviant behaviour Denial of injury (DOI) Sykes & Matza (1957) Claiming that the deviant behaviour caused no injury Condemnation of the condemners (COC) Sykes & Matza (1957) Claiming that those who condemn the deviant behaviour engage themselves in similar behaviour Appeal to higher loyalties (AHL) Sykes & Matza (1957) Claiming that the deviant behaviour was due to actualising a higher-order ideal or value Metaphor of the ledger (MOL) Klockars (1974) Claiming that the previous good behaviour counterbalances the present bad behaviour Defence of necessity (DON) Minor (1981) Claiming that the deviant behaviour was necessary Claim of relative acceptability (CRA) Henry & Eaton (1999) Claiming that the deviant behaviours of others are even worse than my deviant behaviour Claim of individuality (COI) Henry & Eaton (1999) Claiming not to care about what others think of me or my behaviour Justification by comparison (JBC) Cromwell & Thurman (2003) Claiming that the deviant behaviour is still better in comparison to some other behaviours Claim of entitlement (COE) Coleman (2005) Claiming to have the right to engage in the deviant behaviour and to gain the benefits from it The theoretical foundation of the study is based on the neutralisation theory by Sykes and Matza (1957), which suggests that when individuals engage in deviant behaviour, they may try to eliminate the resulting feelings of guilt or shame by using various justifications for the deviant behaviour that are referred to as neutralisation 696 37 TH B LED E C ONFERENCE : R ESILIENCE T HROUGH D IGITAL I NNOVATION : E NABLING THE T WIN T RANSITION techniques. Although originally developed for the context of juvenile delinquency, the neutralisation theory has later been applied to also other contexts, such as inappropriate consumer behaviour (Strutton et al., 1994), fair trade (Chatzidakis et al., 2007), immoral and unethical consumption (McGregor, 2008), employee IS security policy violations (Siponen & Vance, 2010), software piracy (Siponen et al., 2012), music piracy (Riekkinen & Frank, 2014), sustainable consumption (Antonetti & Maklan, 2014b; Gruber & Schlegelmilch, 2014), shadow IT use (Silic et al., 2017), employee unauthorised computer access (Lin et al., 2018), and responsible online shopping (Makkonen et al., 2023). Of the various neutralisation techniques proposed in prior literature, this study focuses specifically on the ten neutralisation techniques in Table 1. These have all been found to be used by consumers in the context of sustainable consumption by Gruber and Schlegelmilch (2014), which is why we assume them to be relevant for consumers also in the closely connected context of responsible online shopping. 3 Methodology The data for the study was collected from Finnish consumers between February and March 2023 with an online survey conducted by using the LimeSurvey service. The survey respondents were recruited by promoting the survey on social media and via the communication channels of Finnish universities and student associations. As an incentive for responding, all the respondents who completed the survey were able to take part in a prize drawing of ten gift boxes worth about 25 € each. In the survey questionnaire, the use of the ten neutralisation techniques was measured with two items each. These were developed for the study based on the studies by Siponen and Vance (2010) as well as Gruber and Schlegelmilch (2014). In turn, anticipated guilt was measured with three items. These were adapted from the guilt inventory by Kugler and Jones (1992) as exemplified by Onwezen et al. (2013, 2014a, 2014b). The wordings of these 23 items are reported in Appendix A, and before presenting them to the respondents, we also provided a brief definition of responsible online shopping as “making consumption choices that take various ecological and ethical values (e.g., sustainable development and fair trade) into account while shopping online”. The measurement scale of all the aforementioned items was the traditional five-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree). In contrast, gender, age, income, and online shopping frequency were measured with only one item each, with age being M. Makkonen et al.: The Use of Neutralisation Techniques in the Context of Responsible Online Shopping: A Latent Profile Analysis 697 measured on a continuous scale and the other variables on a categorical scale. To avoid forced responses, the respondents also had the option to skip any item in the survey. The collected data was analysed in three phases. First, we calculated a composite score for each neutralisation technique construct and the anticipated guilt construct by averaging the scores of the individual items that were measuring them as well as assessed their reliability in terms of internal consistency by using Cronbach’s alphas and their validity in terms of discriminant validity by using disattenuated correlations as suggested by Rönkkö and Cho (2022). Second, we used the Mplus 8.8. statistical software (Muthén & Muthén, 2024) to conduct a latent profile analysis for the neutralisation technique constructs by estimating multiple models with a varying number of profiles and assessing their goodness of fit with the data. To estimate the models, we used the robust maximum likelihood (MLR) estimator, with the full information maximum likelihood (FIML) estimator used for handling the potential missing values. In turn, to assess model fit, we used four information criteria and two likelihood ratio tests recommended in recent methodological literature (e.g., Nylund-Gibson & Choi, 2018; Ferguson et al., 2020; Weller et al., 2020). The four information criteria were the consistent Akaike information criterion (CAIC) by Bozdogan (1987), the Bayesian information criterion (BIC) by Schwarz (1978), the sample-size adjusted Bayesian information criterion (SABIC) by Sclove (1987), and the approximate weight of evidence (AWE) by Banfield and Raftery (1993). In the case of these all, a lower value suggests a better fitting model, thus typically resulting in the selection of the model with the lowest value. Or, if the values continue to decrease while increasing the number of profiles, then the model after which the improvements in model fit become only marginal may also be selected (NylundGibson & Choi, 2018). In turn, the two likelihood ratio tests were the Vuong-LoMendell-Rubin adjusted likelihood ratio test (VLMR-LRT) by Vuong (1989) and Lo et al. (2001) as well as the bootstrapped likelihood ratio test (BLRT) by McLachlan and Peel (2000). These are used to compare a model with k profiles against a model with k – 1 profiles to see whether the additional profile provides a statistically significant improvement in model fit or whether the model with fewer profiles is sufficient. In addition, we also diagnosed the quality of the estimated models by examining their entropy (Celeux & Soromenho, 1996), in the case of which a value that is greater than 0.8 is commonly considered to suggest sufficient differentiation between the profiles (Nylund-Gibson & Choi, 2018). Third, we used the IBM SPSS 698 37 TH B LED E C ONFERENCE : R ESILIENCE T HROUGH D IGITAL I NNOVATION : E NABLING THE T WIN T RANSITION Statistics 28 software to conduct post-hoc analyses of the potential differences between the members of each profile in terms of their gender, age, income, online shopping frequency, and anticipated guilt based on the most likely profile membership. In the case of gender, income, and online shopping frequency, this was done by using cross-tabulation analysis, whereas in the case of age and anticipated guilt, this was done by using one-way analysis of variance. 4 Results In total, we received 478 valid responses to the conducted online survey. The descriptive statistics of this sample in terms of the gender, age, yearly personal taxable income, socioeconomic status, and average online shopping frequency of the respondents are reported in Table 2. As can be seen, most of the respondents were women and students as well as had a relatively low income, which was not surprising when considering how they were recruited. The age of the respondents ranged from 19 to 75 years, with a mean of 28.3 years and a standard deviation of 9.0 years. Most of the respondents (68.8%) were also relatively active online shoppers who shopped online at least monthly on average. Table 2: Sample statistics (N = 478) N % N % Gender Socioeconomic status Man 88 18.4 Student 341 71.3 Woman 364 76.2 Employee or selfemployed 132 27.6 Other 26 5.4 Unemployed or unable to work 10 2.1 Age Pensioner 4 0.8 Under 25 years 206 43.1 Other 4 0.8 25–49 years 253 52.9 Online shopping frequency 50 years or over 19 4.0 At least weekly 31 6.5 Yearly personal taxable income At least monthly 298 62.3 Under 15,000 € 286 59.8 At least yearly 139 29.1 15,000–29,999 € 71 14.9 Less frequently than yearly 8 1.7 30,000 € or over 98 20.5 Has never shopped online 1 0.2 No response 23 4.8 No response 1 0.2 M. Makkonen et al.: The Use of Neutralisation Techniques in the Context of Responsible Online Shopping: A Latent Profile Analysis 699 4.1 Construct Reliability and Validity Table 3 reports for each neutralisation technique construct and the anticipated guilt construct the mean (M) and standard deviation (SD) of its composite score as well as its Cronbach’s alpha (on-diagonal) and disattenuated correlations (off-diagonal). Of them, Cronbach’s alphas of at least 0.7 are commonly considered to suggest sufficient construct reliability in terms of internal consistency (Nunally & Bernstein, 1994). This criterion was met by all the constructs except for the claim of individuality, which was also so close to meeting the criterion that we decided not to drop it. In turn, disattenuated correlations of less than 0.85 are commonly considered to suggest sufficient construct validity in terms of discriminant validity (Rönkkö & Cho, 2022). This was met by all the constructs. Table 3: Construct statistics N M SD Cronbach’s alphas and disattenuated correlations DOR DOI COC AHL MOL DON CRA COI JBC COE AG DOR 478 2.134 1.118 0.932 DOI 476 1.913 0.943 0.476 0.889 COC 460 2.548 1.164 0.554 0.455 0.870 AHL 478 4.271 0.794 0.274 0.228 0.311 0.882 MOL 477 1.932 0.901 0.418 0.486 0.543 0.285 0.786 DON 475 4.024 0.931 0.149 0.028 0.056 0.355 0.138 0.847 CRA 469 2.457 1.090 0.553 0.444 0.754 0.414 0.657 0.196 0.720 COI 478 2.690 1.058 0.399 0.627 0.509 0.332 0.434 0.031 0.428 0.695 JBC 478 2.522 1.088 0.607 0.638 0.663 0.315 0.522 0.088 0.713 0.611 0.833 COE 476 2.532 1.159 0.350 0.572 0.514 0.365 0.368 0.005 0.457 0.716 0.591 0.881 AG 470 3.310 1.057 -0.322 -0.446 -0.245 -0.279 -0.196 0.012 -0.210 -0.520 -0.366 -0.428 0.838 4.2 Latent Profile Analysis Table 4 reports the log-likelihood (LL) value, the values of the four information criteria (i.e., CAIC, BIC, SABIC, and AWE), the p-values of the two likelihood ratio tests (i.e., VLMR-LRT and BLRT), and the entropy value of the estimated models in which the number of profiles (k) ranged from one to seven. The values of the four information criteria are also plotted graphically in Appendix B. The four information criteria all suggested the selection of the four-profile model because both CAIC and AWE reached their lowest value in the case of this model and also 700 37 TH B LED E C ONFERENCE : R ESILIENCE T HROUGH D IGITAL I NNOVATION : E NABLING THE T WIN T RANSITION the values of BIC and SABIC showed only marginal decreases when increasing the number of profiles beyond four. This suggestion was also supported by VLMRLRT, which showed that increasing the number of profiles from four to five would not result in a statistically significant improvement in model fit (p = 0.232). Despite the lack of support from BLRT, we thus decided to proceed with the four-profile model. This model also had a very high entropy value of 0.927, which suggests good differentiation between the profiles. Table 4: Fit and entropy of the estimated models k LL CAIC BIC SABIC AWE VLMRLRT BLRT Entropy 1 -5,994.406 12,447.667 12,383.667 12,180.539 12,479.667 < 0.001 < 0.001 – 2 -5,890.131 12,317.983 12,242.983 12,004.942 12,355.483 0.002 < 0.001 0.933 3 -5,813.364 12,243.315 12,157.315 11,884.361 12,286.315 0.046 < 0.001 0.929 4 -5,764.002 12,223.456 12,126.456 11,818.590 12,271.956 0.232 < 0.001 0.927 5 -5,727.538 12,229.394 12,121.394 11,778.615 12,283.394 0.354 < 0.001 0.928 6 -5,686.565 12,226.314 12,107.314 11,729.622 12,285.814 0.601 < 0.001 0.925 7 -5,662.880 12,257.809 12,127.809 11,715.205 12,322.809 0.508 < 0.001 0.918 Table 5: Estimation results of the four-profile model Mean score (from 1 to 5) Result of the Wald test (p-value) LP1 (62.8%) LP2 (23.4%) LP3 (7.5%) LP4 (6.3%) LP1 vs. LP2 LP1 vs. LP3 LP1 vs. LP4 LP2 vs. LP3 LP2 vs. LP4 LP3 vs. LP4 DOR 1.634 3.917 1.616 1.389 < 0.001 0.902 0.046 < 0.001 < 0.001 0.193 DOI 1.794 2.522 1.665 1.314 < 0.001 0.440 < 0.001 < 0.001 < 0.001 0.071 COC 2.342 3.384 2.552 1.637 < 0.001 0.424 < 0.001 0.002 < 0.001 0.001 AHL 4.387 4.583 4.229 2.157 0.007 0.354 < 0.001 0.050 < 0.001 < 0.001 MOL 1.816 2.400 1.879 1.511 < 0.001 0.709 0.038 0.006 < 0.001 0.082 DON 4.253 4.257 1.922 3.416 0.973 < 0.001 0.004 < 0.001 0.005 < 0.001 CRA 2.325 3.171 2.364 1.609 < 0.001 0.862 < 0.001 < 0.001 < 0.001 0.003 COI 2.590 3.145 2.858 1.899 < 0.001 0.239 < 0.001 0.244 < 0.001 < 0.001 JBC 2.315 3.395 2.204 1.873 < 0.001 0.621 0.007 < 0.001 < 0.001 0.200 COE 2.428 2.989 2.791 1.669 < 0.001 0.211 < 0.001 0.499 < 0.001 < 0.001 Table 5 reports the estimation results of the four-profile model in terms of the mean scores of the neutralisation technique constructs in each of the four latent profiles (LP1–LP4) and the p-values of the Wald test for the pairwise comparisons of the differences in the mean scores between the profiles. The mean scores are also plotted M. Makkonen et al.: The Use of Neutralisation Techniques in the Context of Responsible Online Shopping: A Latent Profile Analysis 707 Kemppainen, T., Frank, L., Makkonen, M., & Hyvönen, O.-I. (2021). Barriers to responsible consumption in e-commerce: Evidence from fashion shoppers. In A. Pucihar, M. Kljajić Borštnar, R. Bons, H. Cripps, A. Sheombar & D. Vidmar (Eds.), Proceedings of the 34th Bled eConference (pp. 323–335). Maribor, Slovenia: University of Maribor Press. Klockars, C. B. (1974). The Professional Fence. New York, NY: Free Press. Kugler, K., & Jones, W. H. (1992). On conceptualizing and assessing guilt. Journal of Personality and Social Psychology, 62(2), 318–327. Levene, H. (1960). Robust tests for equality of variances. In I. Olkin (Ed.), Contributions to Probability and Statistics (pp. 278–292). Palo Alto, CA: Stanford University Press. Lin, T.-C., Hsu, J. S.-C., Wang, Y.-C., & Wu, S. (2018). Examining the antecedents of employee unauthorized computer access. Journal of Statistics & Management Systems, 21(3), 493–517. Lindenmeier, J., Lwin, M., Andersch, H., Phau, I., & Seemann, A.-K. (2017). Anticipated consumer guilt: An investigation into its antecedents and consequences for fair-trade consumption. Journal of Macromarketing, 37(4), 444–459. Lo, Y., Mendell, N. R., & Rubin, D. B. (2001). Testing the number of components in a normal mixture. Biometrika, 88(3), 767–778. Makkonen, M., Frank, L., & Kemppainen, T. (2019a). The effects of individual values on online shopping spending. In A. Pucihar, M. Kljajić Borštnar, R. Bons, J. Seitz, H. Cripps & D. Vidmar (Eds.), Proceedings of the 32nd Bled eConference (pp. 969–994). Maribor, Slovenia: University of Maribor Press. Makkonen, M., Frank, L., & Paananen, T. (2023). The role of anticipated guilt and its neutralisation in explaining responsible online shopping. In A. Pucihar, M. Kljajić Borštnar, R. Bons, G. Ongena, M. Heikkilä & D. Vidmar (Eds.), Proceedings of the 36th Bled eConference (pp. 595–613). Maribor, Slovenia: University of Maribor Press. Makkonen, M., Nyrhinen, J., Frank, L., & Karjaluoto, H. (2022). The effects of general and mobile online shopping skilfulness and multichannel self-efficacy on consumer showrooming behaviour. In A. Pucihar, M. Kljajić Borštnar, R. Bons, A. Sheombar, G. Ongena & D. Vidmar (Eds.), Proceedings of the 35th Bled eConference (pp. 479–494). Maribor, Slovenia: University of Maribor Press. Makkonen, M., Riekkinen, J., Frank, L., & Jussila, J. (2019b). The effects of positive and negative emotions during online shopping episodes on consumer satisfaction, repurchase intention, and recommendation intention. In A. Pucihar, M. Kljajić Borštnar, R. Bons, J. Seitz, H. Cripps & D. Vidmar (Eds.), Proceedings of the 32nd Bled eConference (pp. 931–953). Maribor, Slovenia: University of Maribor Press. McGregor, S. L. T. (2008). Conceptualizing immoral and unethical consumption using neutralization theory. Family and Consumer Sciences Research Journal, 36(3), 261–276. McLachlan, G., & Peel, D. (2000). Finite Mixture Modeling. New York, NY: Wiley. Minor, W. W. (1981). Techniques of neutralization: A reconceptualization and empirical examination. Journal of Research in Crime and Delinquency, 18(2), 295–318. Muthén, L. K., & Muthén, B. O. (2024). Mplus Home Page. Available at https://www.statmodel.com Nangia, P., Bansal, S., & Thaichon, P. (2024). Doing more with less: An integrative literature review on responsible consumption behaviour. Journal of Consumer Behaviour, 23(1), 141–155. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric Theory (3rd ed.). New York, NY: McGraw-Hill. Nylund-Gibson, K., & Choi, A. Y. (2018). Ten frequently asked questions about latent class analysis. Translational Issues in Psychological Science, 4(4), 440–461. Onwezen, M. C., Antonides, G., & Bartels, J. (2013). The Norm Activation Model: An exploration of the functions of anticipated pride and guilt in pro-environmental behaviour. Journal of Economic Psychology, 39, 141–153. Onwezen, M. C., Bartels, J., & Antonides, G. (2014a). Environmentally friendly consumer choices: Cultural differences in the self-regulatory function of anticipated pride and guilt. Journal of Environmental Psychology, 40, 239–248. 708 37 TH B LED E C ONFERENCE : R ESILIENCE T HROUGH D IGITAL I NNOVATION : E NABLING THE T WIN T RANSITION Onwezen, M. C., Bartels, J., & Antonides, G. (2014b). The self-regulatory function of anticipated pride and guilt in a sustainable and healthy consumption context. European Journal of Social Psychology, 44(1), 53–68. Pellegrino, A., Wang, R., & Stasi, A. (2023). Exploring the intersection of sustainable consumption and the Metaverse: A review of current literature and future research directions. Heliyon, 9(9), e19190. Rawlings, E. I. (1970). Reactive guilt and anticipatory guilt in altruistic behavior. In J. Macaulay & L. Berkowitz (Eds.), Altruism and Helping Behavior (pp. 163–177). New York, NY: Academic Press. Riekkinen, J., & Frank, L. (2014). Music piracy neutralization and the youth of the 2010’s. In A. Pucihar, C. Carlsson, R. Bons, R. Clarke & M. Kljajić Borstnar (Eds.), Proceedings of the 27th Bled eConference (pp. 44–54). Kranj, Slovenia: Moderna organizacija. Rönkkö, M., & Cho, E. (2022). An updated guideline for assessing discriminant validity. Organizational Research Methods, 25(1), 6–47. Schwarz, G. (1978). Estimating the dimension of a model. Annals of Statistics, 6(2), 461–464. Sclove, S. L. (1987). Application of model-selection criteria to some problems in multivariate analysis. Psychometrika, 52(3), 333–343. Silic, M., Barlow, J. B., & Back, A. (2017). A new perspective on neutralization and deterrence: Predicting shadow IT usage. Information & Management, 54(8), 1023–1037. Siponen, M., & Vance, A. (2010). Neutralization: New insights into the problem of employee information systems security policy violations. MIS Quarterly, 34(3), 487–502. Siponen, M., Vance, A., & Willison, R. (2012). New insights into the problem of software piracy: The effects of neutralization, shame, and moral beliefs. Information & Management, 49(7–8), 334–341. Strutton, D., Vitell, S. J., & Pelton, L. E. (1994). How consumers may justify inappropriate behavior in market settings: An application on the techniques of neutralization. Journal of Business Research, 30(3), 253–260. Sykes, G. M., & Matza, D. (1957). Techniques of neutralization: A theory of delinquency. American Sociological Review, 22(6), 664–670. Theotokis, A., & Manganari, E. (2015). The impact of choice architecture on sustainable consumer behavior: The role of guilt. Journal of Business Ethics, 131(2), 423–437. Ulusoy, E. (2016). Experiential responsible consumption. Journal of Business Research, 69(1), 284–297. Vuong, Q. H. (1989). Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica, 57(2), 307–333. Welch, B. L. (1951). On the comparison of several mean values: An alternative approach. Biometrika, 38(3/4), 330–336. Weller, B. E., Bowen, N. K., & Faubert, S. J. (2020). Latent class analysis: A guide to best practice. Journal of Black Psychology, 46(4), 287–311 M. Makkonen et al.: The Use of Neutralisation Techniques in the Context of Responsible Online Shopping: A Latent Profile Analysis 709 Appendix A: Item Wordings Item Wording I find that is OK for me not to make responsible consumption choices when shopping online because… DOR1 … one person cannot really trigger any change with his or her choices. DOR2 … one person cannot really change anything with his or her choices. DOI1 … it causes no actual harm to anybody. DOI2 … it caused no actual damage to anybody. COC1 … people who call for responsibility from others sometimes do the same. COC2 … people who call for responsibility from others do not always themselves make responsible choices. AHL1 … I have to consider also other values or criteria (e.g., price) when making my choices. AHL2 … I have to take into account also other values or criteria (e.g., price) when making my choices. MOL1 … I have already made enough responsible choices earlier in my life. MOL2 … the responsible choices that I have made earlier in my life compensate for it. DON1 … the lack of responsible alternatives sometimes makes it necessary. DON2 … responsible alternatives are not always available. CRA1 … many other people fail to make them even more often than me. CRA2 … I still fail to make them less often than many other people. COI1 … I do not care what other people think about my choices. COI2 … my choices do not belong to other people. JBC1 … there are far worse things in the world. JBC2 … it is not a very bad thing compared to many other things. COE1 … I am entitled to do so if I want to. COE2 … I have the right to do so if I wish. If I do not make responsible consumption choices when shopping online, I feel… AG1 … guilty. AG2 … remorseful. AG3 … bad. 710 37 TH B LED E C ONFERENCE : R ESILIENCE T HROUGH D IGITAL I NNOVATION : E NABLING THE T WIN T RANSITION Appendix B: Information Criteria of the Estimated Models Appendix C: Estimation Results of the Four-Profile Model 11.700 11.800 11.900 12.000 12.100 12.200 12.300 12.400 12.500 1234567 CAIC BIC SABIC AWE 1,000 2,000 3,000 4,000 5,000 DOR DOI COC AHL MOL DON CRA COI JBC COE LP1 (62.8%) LP2 (23.4%) LP3 (7.5%) LP4 (6.3%)