Quizlet Methods Know How to Interpret the Influence of a Dichotomous Iv on a Continuous Dv
Hello,
I need to run a correlation in SPSS between two variables.
One is a dichotomous variable (A). People have either answered the question correctly or incorrectly (coded as '1' for correct or '0' for incorrect).
The other is a continuous variable (B), ranging between 6-36. (This number was calculated by asking six questions, each with a score from 1-6. Responses across all six questions were added together to give an overall score between 6 and 36).
I need to see whether answers to the dichotomous variable (A) are correlated with answers to the score (B).
Can I simply run a correlation in SPSS (bivariate-Pearson-two-tailed)? This seems to tell me that my variables are correlated and significant, but given that one was dichotomous and one was continuous, I wonder if this can be correct?
The correlation is .117** How can I interpret this?
Sorry if this is a very simple question - any help much appreciated! Thank you! 
You should run polychoric correlations. Pearson R assumes both variables will be continuous and the results won't be accurate with a dichotomous variable. Unfortunately, SPSS does not do this in its normal statistics. You have to use its R studio (essentially this appears to use R code inside SPSS) to get these.
You could intepret this as statistically signficant (that is what the double star means in SPSS, this should be in the table that ran). Using Cohen's rule (of thumb) that is not a very large effect I believe (you should look up Cohen's rule as it has been a while since I used it).
Thank you very much for replying.
Unfortunately I am unfamiliar with R Studio or how to add anything to the version of SPSS that I am using.
Is there any other way of me being able to look at the association between my two variables in SPSS? I have just been told to look at how they are correlated, but it seems I can't do this?
So is there another way to explore associations?
Thank you again
Thank you noetsi.
I had found that article before when googling this issue, but found it a little difficult to understand whether a 'point-biserial' correlation essentially means I can just perform the normal/standard type of bivariate correlation, and be confident in the answer.
So to confirm that I have understood: I can just go to analyse - correlate - bivariate, and then look at the Pearson correlation co-efficient, for a two-tailed test?
If so, that is what I had done and it appears as though the two are significantly correlated. It just seemed a little too easy...?
Thanks for taking the time to help me with this, it is much appreciated!!
It has been too long in using SPSS for me to be sure about the code although that sounds right.
Correlation is easy. Wait until you get to logistic regression or multilevel analysis (or the crazy stuffy dason or jake run)
The one thing that is a bit more difficult is knowing is what the .1... (the effect size) really means in practice. It is common just to see if the correlation is signficant and ignore interpretation of the effect size - but it is probably more useful to go further and try to determine whether that relations is meaningful. You can be statistically significant with a large sample and have very weak effect size.
I reccomend looking up something like Cohen's rule.
N.B. I also did an independent t-test (with scores on A as my IV, and scores on B as my DV).
It showed that people who answered A incorrectly had a significantly lower mean score for B.
So I suppose this is another way of showing that the two scores are related...?
Yes. You could run linear regression and see what the slope is. But with one predictor you probably would get what you got on the correlation.
Source: https://www.talkstats.com/threads/correlations-between-dichotomous-and-continuous-variables.50987/
0 Response to "Quizlet Methods Know How to Interpret the Influence of a Dichotomous Iv on a Continuous Dv"
Post a Comment