Author(s) This pre-registration is currently anonymous to enable blind peer-review. It has 2 authors.
Pre-registered on 2019/12/11 - 03:03 PM (PT)
1) Have any data been collected for this study already? No, no data have been collected for this study yet.
2) What's the main question being asked or hypothesis being tested in this study? CENTRAL QUESTION: Do our previously observed relations between SELF-REPORTED (i) trait Compassion and inclusiveness, (ii) trait Openness and inclusiveness, and (iii) trait Enlightened Compassion (EC) and inclusiveness generalise to a BEHAVIOURAL measure of inclusiveness?
CENTRAL HYPOTHESIS: We predict that Compassion, Openness, and EC will each be positively related to a behavioural measure of inclusiveness.
3) Describe the key dependent variable(s) specifying how they will be measured. ‘Inclusiveness’ will be operationalised using the ‘Parochial Charitable Donation (PCD) Task’. The PCD Task is a novel researcher-designed measure, modelled off an existing measure called the ‘Moral Expansiveness Scale’ (MES; Crimston et al., 2016). Like the MES, the PCD Task intends to measure the narrowness versus breadth of people’s concern for a range of human, animal, environmental, and inanimate targets. Unlike the MES, which is a self-report measure, the PCD Task measures inclusiveness behaviourally (i.e., there are real consequences for participants’ decisions in the task).
Specifically, in the PCD Task, the participant is allocated $1, and is then presented with a list of charitable causes. They are told that ONE of these causes will be randomly selected by the researchers and revealed at the end of the study. In the meantime, the participant must respond to each charitable cause by indicating how much of their $1 they would be willing to donate to that cause (versus keep for themselves) if it were to be randomly selected at the end of the study. At the end of the study, the randomly selected charitable cause is revealed, and the researchers donate the specified portion of the participant’s $1 to this cause (and pay the participant the remainder).
An overall inclusiveness score for the PCD Task will be computed, and sub-scores will also be separately computed by combining specific charitable causes. Conceptually, the charitable causes can be combined as follows: ‘the afflicted’ (cancer patients, foster children, disaster victims, people with disabilities), ‘the battlers’ (military personnel, emergency service workers, farmers, seniors), ‘the stigmatized’ (people with mental illness, homeless people, refugees, people with substance dependence), ‘the marginalized’ (LGBT+ people, African Americans, women, Muslims), ‘the non-human’ (companion animals, farmed animals, wild animals, captive animals), ‘the non-sentient’ (the planet, oceans, forests, natural landmarks), ‘the non-natural’ (sacred artifacts, heritage icons, artistic creations, scientific works), and ‘the foreign’ (people living in poverty). Our final groupings of the charitable causes, and consequent computation of sub-scores, may shift based on analyses of their intercorrelations.
4) How many and which conditions will participants be assigned to? This is a correlational design, and thus there are no experimental conditions. However, we conceptualise Compassion, Openness, and EC as our independent variables (NB: ‘EC’ is conceptualised as the covariance among Compassion and Openness).
The Big Five Aspects Scales (BFAS; DeYoung, Quilty, & Peterson, 2007) will be used to measure Compassion and Openness. Several methods may be used to measure EC: (1) the average of BFAS Compassion and BFAS Openness scores, (2) a latent factor extracted from BFAS Compassion and BFAS Openness items, and/or (3) a novel researcher-designed scale explicitly designed to target the theorised content of EC. In relation to the latter, an initial item pool of 72 items was developed by the current research team and administered to our previous samples (see the ‘Trait Compassion/Openness/Enthusiasm Relations, Melbourne, July 2017’ pre-registration for details). This item pool will be refined pending further analyses of these previous samples. Because this scale refinement exercise is still ongoing, we will administer the 40 most conceptually relevant items from our item pool in the current study—in addition to 24 supplementary items described at the bottom of this pre-registration—but only the (yet to be determined) refined set of items will be analysed.
5) Specify exactly which analyses you will conduct to examine the main question/hypothesis. Support for each hypothesis will be determined by examining whether each independent variable is correlated with inclusiveness, as measured by overall PCD scores:
For Compassion, we expect a positive and statistically significant bivariate correlation with inclusiveness. We also expect a positive and statistically significant partial correlation with inclusiveness when controlling for Openness. For Openness, we expect a positive and statistically significant bivariate correlation with inclusiveness. We also expect a positive and statistically significant partial correlation with inclusiveness when controlling for Compassion.
For EC, we expect a positive and statistically significant bivariate correlation with inclusiveness. We also expect a positive and statistically significant partial correlation with inclusiveness when controlling for Compassion, and a positive and statistically significant partial correlation with inclusiveness when controlling for Openness.
We will supplement our main analysis of overall PCD scores with analyses at the sub-score level. Though we have not yet determined how these sub-scores will be computed, for the purposes of this pre-registration, we use our aforementioned conceptual groupings of charitable causes (‘the afflicted’, ‘the battlers’, ‘the stigmatized’, ‘the marginalized’, ‘the non-human’, ‘the non-sentient’, ‘the non-natural’, and ‘the foreign’):
For Compassion, we expect positive and statistically significant bivariate correlations with all sub-scores except ‘the non-natural’. We also expect a positive and statistically significant partial correlation with ‘the afflicted’, ‘the battlers’, ‘the stigmatized’, ‘the marginalized’, and the ‘foreign’ sub-scores when controlling for Openness. For Openness, we expect positive and statistically significant bivariate correlations with all sub-scores except ‘the afflicted’ and ‘the battlers’. We also expect a positive and statistically significant partial correlation with ‘the non-human’, ‘the non-sentient’, and ‘the non-natural’ sub-scores when controlling for Compassion.
For EC, we expect positive and statistically significant bivariate correlations with all sub-scores. We also expect a positive and statistically significant partial correlation with all sub-scores when controlling for Compassion, and a positive and statistically significant partial correlation with all sub-scores when controlling for Openness.
6) Describe exactly how outliers will be defined and handled, and your precise rule(s) for excluding observations. Our survey includes three attention checks. If a participant fails the first of these, they will receive a warning, and will not be able to proceed until they answer the attention check correctly. One of the remaining checks is embedded in a personality questionnaire, and the other remaining check is embedded in the PCD Task. Participants will be excluded from analysis if they fail one or both of the latter two checks.
Our survey also includes a comprehension check, embedded in the PCD Task. Participants will not be excluded from analysis for failing this comprehension check. Rather, participants will complete this comprehension check prior to making their donation decisions, and will not be able to proceed until they answer it correctly.
We anticipate that one source of non-comprehension in the PCD Task could be the failure to realise that each donation decision is independent from the others (i.e., that the $1 does not have to be ‘split’ across the charitable causes, but rather the $1 amount ‘resets’ for each charitable cause). To additionally protect against such non-comprehension, participants whose donation decisions sum to exactly $1 will receive a warning message. These participants will then be given the option to revise their responses. These revised responses (rather than their original responses) would then be used in the analysis.
7) How many observations will be collected or what will determine sample size? No need to justify decision, but be precise about exactly how the number will be determined. A correlation of r = .20 represents a ‘typical’ effect in personality psychology (Gignac & Szodorai, 2016). Furthermore, the sample-weighted average correlations between our focal traits and the conceptually relevant facets of the MES across our previous studies are typically r > .20.
To detect a correlation of r = .20 at 80% power, the current study would require a sample-size of N = 194. Based on previous experience, we expect ~10% of recruited MTurk participants to fail our attention checks. Thus, we aim to recruit a sample of N = 216 to ensure adequate power following the exclusion of inattentive participants.
8) Anything else you would like to pre-register? (e.g., secondary analyses, variables collected for exploratory purposes, unusual analyses planned?) In addition to the 40 items from our EC item pool, we will administer 24 supplementary items that we have not administered to any previous samples. These items were extracted from a large pool of public domain ‘International Personality Item Pool’ (IPIP) items (https://ipip.ori.org/; Goldberg, 1999). Based on analyses of the Eugene Springfield Community Sample, which the authors of DeYoung et al. (2007) conducted and recently shared with us, we selected these items because they are each strongly correlated (r > .30) with BOTH Compassion and Openness (with the difference between Compassion and Openness’ correlations with each item being r < .20), suggesting they are reasonable candidate EC items. We will therefore explore the convergence of these IPIP items with our own EC items, and possibly incorporate them into our final EC scale.
In addition to our core independent and dependent variables, we additionally administer the aforementioned Moral Expansiveness Scale (MES; Crimston et al., 2016), as well as the Self-Transcendence scale of the Portrait Values Questionnaire 5X (PVQ-5X; Schwartz et al., 2012), which both measure self-reported inclusiveness. We include these measures so that we can i) explore the convergent validity of our PCD Task, and ii) explore how strongly EC converges with the PCD Task relative to how strongly the MES and PVQ-5X converge with the PCD Task.
For similar reasons to the above, we additionally administer the Identification with all Humanity Scale (McFarland et al., 2012), Solidarity with Animals Scale (Amiot & Bastian, 2017), and Connectedness to Nature Scale (Mayer & Frantz, 2004), which measure how strongly one feels connected to all of humanity, animals, and nature, respectively. However, unlike the MES and PVQ-5X, we may remove these three scales following initial pilot testing if we find that the typical survey completion time exceeds our advertised time of 45minutes.
