Framing effect

YESP

Hi all,

we need your help for our university project.

The research topic is framing which describes the phenomenon that changes in the wording (with the same expected value) might influence individual's preferences respectively the decision-making behavior.

We have the following two null and alternative hypotheses:

H0a: In the life-death domain under time pressure conditions, framing has the same effect on female and male students or a stronger effect on male students.
H1a: In the life-death domain under time pressure conditions, framing has a stronger effect on female students compared to the male counterparts.

H0b: In the monetary domain under time pressure conditions, framing has the same effect on female and male students or a stronger effect on female students.
H1b: In the monetary domain under time pressure conditions, framing has a stronger effect on male students compared to the female counterparts.

Additional information to the experiment:

Research design, procedure and measures:

• we investigated the decision-making process of female and male students in two domains, namely life-death and money
• every participant had to solve two tasks (first task was life-death, second task money) by selecting the preferred option (safe or risky)
• the two tasks were either positively (i.e. lives safed) or negatively (i.e., lives lost) framed depending on the situation
• we have a sample size of 64 students
• 4 situations (treatment group: situation 1 and situation 2, control group: situations 3 and situation 4)
• treatment groups = with time pressure, control groups = without time pressure
• variables: independent: framing valence (nominal: 1 = positive, 2 = negative), gender (nominal: 1 = female, 2 = male), time pressure (situation 1/situation 2)
• dependent: choice preference (nominal: 1 = safe option, 2 = risky option)

An exemplary task sheet (for situation 1 respectively situation 3), our dataset and the research study are attached.

Regarding statistical tests: There is a past study using logistic regression (with a Wald test). This study was also investigating gender differences in framing effects but without analyzing the effect of time pressure. I assume that we also have to apply a logistic regression. We are not sure which statistical method we have to use to answer our research questions respectively to test our hypotheses.

Thank you very much in advance for your answers and help .

eduard

Hi,

I think there are several options how to approach the problem. Things you need to keep in mind are (1) You have a binary DV (safe vs risky), so logistic regression would be one approach unless you only want to look at accuracy averages for each cell of your design. In this case, I think, an ANOVA would also work. (2) You have three between-subject factors (framing x time pressure x gender) for two different tasks, or is the task itself also a factor? In that case, task would be a within--subject factor. So when specifying your model, you need to make sure that you take this into account.
I recommend Jasp (https://jasp-stats.org/), for this analysis, then you can easily also check Bayesian analysis.

Hope this helps! (and no warranties for the accuracy...)

Eduard

YESP

Thanks for the reply. Yes, the task itself is a factor. I tried the logistic regression with JASP but had some problems with it. I have to deepen my knowledge using this statistical software.