• Keine Ergebnisse gefunden

Jarecki & Wilke (in preparation)

A.3.1 Instructions of Study 1

The instructions for the aspect listing in study 1 read as follows:

Under which conditions would you personally be more[less]likely to engage in the described behavior and under which conditions would you personally be less[more]likely to engage in the described behavior?

Please list one situational aspect at a time. Enter the first aspect in the box below and, as soon as you are done, hit the "Enter" key to submit.

If the situation was ... I would be ... (upon clicking this text disappeared)

After the first aspect was entered, the following text appeared:

Please continue to list the circumstances that would make you personally more[less]likely or less[more]

likely to engage in the described behavior. Please keep going until you cannot think of any more. If you need to pause to consider more aspects of this situation, feel free to do so.

Enter the next situational aspect in the box below and, as soon as you are done, hit the "Enter" key to submit.

A.3.2 Methods: Details of the Qualitative Data Analysis

We generated the coding manuals from a subset of statements. These statements were chosen identically for each domain. We obtained the statements from the following four participants: two men and two women, one above and one below median age; specifically the person with most responses among the women below, the women above, the men below, and the men above median age. The assistant and JJ coded domains separately. The rason for this is that we wanted a heterogeneous sample, because evidence suggests that women are more risk-averse than men, and older more risk-averse than younger people.

We took four steps in the manual preprocessing of the qualitative data, broadly resolving discrepancies and applying manual changes.

(1) The raters discussed allcoder discrepancies(statements for which coder 1 and coder 2 disagreed); and resolved them settling on one category.

(2) Thereafter we splitmultiple codings(statements for with both coders assigned multiple codes) into separate statement with separate codes. For example "I would be more likely (...) if their ideas were bad or I had an especially good one (...)" was separated into two codes. Splitting increased the total number of cues for the participant by+1. We preserved the original order of cues. Importantly, we

Appendix A Supplementary materials to the studies 165 decided also to split compound statements (e.g., the text "If there were few fans of my team there,and my support would make a difference (...)", emphasis added). The reason was that with study 1 we were really only interested in collecting frequencies of individual cue categories, and compound cues were very infrequent (33 of 1598 statements); further we investigated the usage of compound cues in study 2.

(3) We removederrors(e.g., two people pressed the enter key before their text, and they wrote the corrected version as the next statement). Removal decremented the total number of a writer’s cues by 1.

(4) We discussedmisunderstandings(statements for which we assumed the writer had misunderstood the question), and removed them. Again, this decremented the writer’s total number of cues.

A.3.3 Results: Comparison to Study 2 by Wilke et al.

TableA.3.1shows the replication results in terms of regression coefficients between the present study and study 2 by Wilke et al.,2014.

166 Appendix A Supplementary materials to the studies

TableA.3.1 Comparisonoftheregressionresults(risktakinglikelihooddomaingender)betweenthepresentstudyandstudy2by Wilkeetal.(2014). VariablePresentStudyStudy2byWilkeetal.(2014) Similar Effect95%CIspEffect95%CIsp

Domain

Within-groupCompetition0.6590.302,1.016<.0010.111-0.096,0.318.292Dir Status/Power-1.607-1.989,-1.226<.001-2.480-2.701,-2.260<.001Yes EnvironmentalExploration-0.618-0.989,-0.247.002-0.242-0.454,-0.030.026Yes FoodAcquisition-0.813-1.187,-0.439<.001-0.854-1.065,-0.644<.001Yes FoodSelection0.5850.219,0.952.002-0.847-1.059,-0.636<.001 ParentalInvestment-0.297-0.667,0.074.117-0.904-1.116,-0.692<.001Dir Kinship1.8391.460,2.217<.0011.5451.325,1.765<.001Yes MateAttraction0.258-0.110,0.625.171-1.137-1.352,-0.921<.001 MateRetention-1.056-1.428,-0.684<.001-1.997-2.212,-1.783<.001Yes Gen.

Female-0.700-1.148,-0.253.003-0.807-1.032,-0.583<.001Yes

Domain Gender

Within-groupCompetitionFemale0.124-0.381,0.630.6300.6510.378,0.923<.001Dir Status/PowerFemale-0.321-0.889,0.246.268-0.245-0.546,0.057.112Yes EnvironmentalExplorationFemale0.421-0.108,0.950.1200.110-0.172,0.392.445Yes FoodAcquisitionFemale0.181-0.351,0.713.5050.3680.091,0.644.010Dir FoodSelectionFemale1.5020.976,2.028<.0011.7151.435,1.994<.001Yes ParentalInvestmentFemale0.8860.355,1.417.0020.9720.693,1.251<.001Yes KinshipFemale1.2730.740,1.806<.0011.1710.884,1.459<.001Yes MateAttractionFemale-0.460-0.987,0.066.0870.063-0.222,0.349.663Yes MateRetentionFemale-0.097-0.630,0.437.7240.112-0.173,0.397.442Yes Note:Femaledenotesagenderdummywith1=female.Effectsareregressioncoefficients,Similarsummarizesthereplication:Yes =replication(effectshavethesamedirectionandaresignificant,orbotharenotsignificantat=.05),Dir=direction(effectshave equaldirectionbutonlyoneissignificant).Thebaselinedomain(intercept)isbetween-groupcompetition.

Appendix A Supplementary materials to the studies 167 Further, the sociodemographic variables differ in the same manner as outlined in the main text between our sample and study 2 from Wilke et al., 2014; see TableA.3.2. There are significant differences between the present and the past sample regarding age (mean 33 vs. 19 years, respectively), ; relationship status, t(123) =13.987,p<.001, Cohen’sd =1.785; number of offspring per person (mean 0.60 vs. 0.057, respectively),t(205)<.001,p=.486, Cohen’sd=0.078; and the proportion of married participants (54% vs. 22%, respectively), 2(1) =38.576,p<.001, Cohen’sh=0.676.

168 Appendix A Supplementary materials to the studies Table A.3.2

Demographics and life history variables in the present study and study 2 by Wilke et al.,2014 Source Statistic FemaleAge Married Children Siblings Life Ex-pectancy

Present sudy Mean 51.7% 33.37 54.20% 0.61 1.56 66.98

Range 18 – 65 0 – 5 0 – 6 35.5 – 100

Study 2 by Wilke et al.Mean 58.1% 19.36 42.35% 0.05 1.62 82.83

Range 17 – 49 0 – 4 0 – 11 34 – 100

Appendix A Supplementary materials to the studies 169 A.3.4 Results: Inter rater reliability

TableA.3.3shows the interrater reliability for categorizing responses in study 1. Two raters coded participants self reports about which aspects of a situation made them more or less likely to engage in risky behaviors across the ten domains. The interrater reliability coefficients are suitable for our data (two raters, multiple nominal categories, and unequal marginal category distributions); for details see Feng,2014. The measures are Gwet’sAC1(Gwet,2008), which performs particularly well when categories are unequally frequent (which is the case in our data).

Table A.3.3

Inter rater reliability for categorizing the situational aspects across the ten domains

Domain Cohen’s Krippendorf’s Gwet’sAC1 Brennan-Prediger’s n

Between-group Competition 0.61 0.61 0.63 0.62

TableA.3.4shows the interrater reliability with respect to the direction of the answer (whether the statement indicated more or less risk-taking) between two raters. For the interrater measures, see Feng, 2014.

Table A.3.4

Inter rater reliability for categorizing the direction across the ten domains

Domain Cohen’s Krippendorf’s Gwet’sAC1 Brennan-Prediger’s n

Between-group Competition 1.00 1.00 1.00 1.00

170 Appendix B Erklärung über den Eigenanteil

A PPENDIX B

E RKLÄRUNG ÜBER DEN E IGENANTEIL

Erklärung über den Eigenanteil an den zur Veröffentlichung vorgesehenen eingereichten wissenschaft-lichen Schriften innerhalb meiner Dissertationsschrift gemäß § 6 Abs. 2 Satz 7 der Promotionsordnung der Mathematisch-Naturwissenschaftliche Fakultät II.

Hiermit erkläre ich, dass ich die vorliegende Arbeit selbständig verfasst und keine anderen als die angegebenen Hilfsmittel benutzt und alle wörtlich oder inhaltlich übernommenen Stellen als solche gekennzeichnet habe.

Hiermit erkläre ich, dass nur die unten genannten Personen an den Studien mitgewirkt haben.

Arbeit

Name: Jana Bianca Jarecki

Institut: Mathematisch-Naturwissenschaftliche Fakultät II Promotionsfach: Psychologie

Thema: Modeling the Decision Making Mind: Does Form follow Function?

Eingereichte Schriften

1. Jarecki, J. B., Tan J. H., & Jenny, M. A. (under review) What is a cognitive process model? A disambiguation.

2. Jarecki, J., Meder, B., & Nelson, J. D. (2013). The Assumption of Class-Conditional Independence in Category Learning. In M. Knauff, M. Pauen, N. Sebanz, & I. Wachsmuth (Eds.),Cooperative Minds: Social Interaction and Group Dynamics. Proceedings of the 35th Annual Conference of the Cognitive Science Society(pp. 2650–2655). Austin, TX: Cognitive Science Society.

3. Jarecki, J. B., Meder, B., & Nelson, J. D. (submitted). The Assumption of Class-Conditional Independence in Category Learning.meanwhile accepted for publication in Cognitive Science, Dec.

2016

4. Jarecki, J. B., & Wilke A. (in preparation)