X-Message-Number: 30448 From: David Stodolsky <> Subject: What makes people take action? Date: Sun, 10 Feb 2008 17:56:58 +0100 After log transforming all attitude and disposition items plus the action item from the Badger (1998) data in order to achieve better distributional characteristics, I again attempt to identify the critical items. First a jackknife outlier analysis is used to remove 94 persons from the data set. Then we perform an ordinal logistic analysis and identify the following significant predictors of the action item, "I believe that Cryonics is an exciting idea and intend on looking into it further.": young again * will not work * too costly too selfish * love life under no circumstances * * p<.01 We conclude that any of the above items can predict whether a person will indicate intention to act. A stepwise regression selects the following as predictors (in entry order, which tends to indicate strength of influence): under no circumstances * young again * will not work * dog revived * no revival * too costly too young (the below are not significant at p<.05, together they raise RSquare from about .47 to about .49, that is not much) million signups too selfish love life optimistic future about death human revived From this we conclude that when the effects of three of the most significant items from the first analysis are taken into account the effect of "dog revived" becomes highly significant, as does the effect of "no revival". Also, "too young" becomes significant, when earlier selected items are taken into account (see the earlier post for discussion of these effects). Since our results are again dominated by "under no circumstances", that is, outright rejection of cryonics (that one item accounts for a majority of the the effect - RSquare=.3), we try to predict what causes this response with an ordinal logistic fit (effects Wald tests) (+ indicates a positive relationship): + will not work * + too complicated + too selfish * - human revived * - cheaper * - only chance - no loss due to ageing - young again Those who agree more that their attitude toward suspension would change "under no circumstances" agree more that cryonics "will not work". The relationship is dominated by those who respond neutral and disagree to "under no circumstances" and correspondingly neutral and disagree to "will not work". The same pattern appears with "too selfish". We see this pattern reflected in the case of "human revived", but the responses are dominated by those indicating disagree to "under no circumstances" and agree to "human revived". The pattern for "cheaper" is in the same direction but with those indicating neutral to "under no circumstances" also responding neutral to "cheaper". In all cases, the responses cluster around neutral and disagree to "under no circumstances". The response neutral appears to be overused. Those indicating disagree, that is, that they would change their attitude toward suspension, are more likely to say it would change, if a human were revived and if suspension was cheaper. On a less significant level they tended to agree it would be nice to avoid losses due to ageing of family/friends, to be young again, and that cryonics is their only chance. Those indicating disagree, that is, that they would change their attitude toward suspension, are less likely to say that cryonics "will not work", that it is "too complicated", and that it is "too selfish". Keep in mind that this last test concerns only the possibility of changing ones attitude, not the intention to actively seek information that could be useful in changing ones attitude. If we select all those indicating disagree to "under no circumstances", that is, those that would change their attitude toward suspension, an ordinal logistic fit predicting action, "I believe that Cryonics is an exciting idea and intend on looking into it further." shows significant: young again * will not work * love life * People willing to change their attitude indicated that the were more likely to seek information, if they agreed more that "I'm excited about the prospect of waking up in a body made young again through bio- technological advances.", "I love being alive and I want to remain alive and healthy for as long as I can.", and agreed less with, "Cryonics doesn't interest me because I just don't think it will work." A stepwise regression selects the following, in order of entry: young again * will not work * human revived * too weird * love life optimistic future * (those below are not individually significant and only raise the RSquare from .33 to .34) about death thousand signups no revival without friends "I'm excited about the prospect of waking up in a body made young again through bio-technological advances." is the runaway best seller accounting for half of the effect (RSquare = .17). "Cryonics doesn't interest me because I just don't think it will work." boosts the RSquare to .23 and from there on the effect of items goes down fast. Previously, I suggested that more religious persons would be difficult to reach, even if they were willing to change their attitude. Here we have support for this conclusion. Even when we include only those who claim they are willing to change their attitude, the strongest negative effect is due to, "Cryonics doesn't interest me because I just don't think it will work." In other words, these people will not seek information about cryonics, because they don't think it will work. The problem is that they are willing to change their attitude, but they never will seek information that could make them change their attitude. So, getting these people to signup would require an educational effort directed toward the entire population or at least the most likely demographic, which still would mean a large segment of the population. The alternative would be to repackage bio-stasis so that it comes with other benefits, that would promote information seeking or reeducation. This appears to be the most economical and self-sufficient approach given the resources available to the cryonics movement at this time. The transformation of the variables has improved the stability of the results, however, we can still see differences between, for example, what items are significant with an ordinal logistic, which is not dependent upon the distribution of the action item and the stepwise regression that is (However, this alone could be due to the increased power of a parametric test, etc.). So, a statistics package that can overcome the limitations of the data would be an improvement. Therefore, we will still have to treat all but the major findings above as provisional. dss David Stodolsky Skype: davidstodolsky Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=30448