Table 1.
Descriptive statistics for Studies 1 and 2.
Table 2.
Pearson correlations for variables in Studies 1 and 2, by language
Fig 1.
Study 1 mediation analysis, with conditional indirect effects of language on temporal discounting via future tense use.
Credibility intervals are plotted at 90% (thick bars) and 95% (thinner whiskers). For two-tailed hypotheses, thin bars represent cut-off criteria; for one-tailed hypotheses, thick bars do (providing sign is as predicted). Counterfactual parameters are in grey. b) Marginal effects of participant-level loge(k) values over mean future tense use conditioned on language (i.e. for each language) with
credibility intervals. Lower values indicate less discounting. c) Conceptual model diagram; interaction term is represented by paths intersecting. Full parameters are given in Table 3, keyed by path name. To ease interpretation, figures throughout use raw predictor values, not the mean-centred z-scaled variables used in regressions.
*95% CI does not contain zero.
Table 3.
Regression coefficients for mediation analyses Studies 1 and 2.
Fig 2.
Study 2 mediation analysis, with conditional indirect effects of language on temporal discounting via future tense use, moderated by language.
(a) Estimates of conditional indirect effects; see Fig 1 for notes on interpretation. (b) Marginal effects of participant-level loge(k) over mean future tense use conditioned on language ( credibility intervals). Lower values indicate less discounting, so higher estimations of delayed future rewards. (c) Conceptual model diagram; full parameters are given in Table 3, keyed by path name. Study 1 replicated: The indirect effect in English was negative and did not contain zero (a, bottom right). The indirect effect in Dutch was non-significant (a, top left).
*95% CI does not contain zero.