Question 776:

Asked on July 8, 2009

1

Answer:

No answer provided yet.Part A:
Question 1: Independent variables are those that are manipulated in the experiment. In this case that would make the independent variable learning program (either the in-class one or the online one). The dependent variable is the thing we measure and cannot control. In this case it's the scores on the multiple choice test.

Question 2: There were two independent sets of participants who provided scores. We would use the 2-sample t-test to compare means. The t-test assumptions are that the data are roughly normally distributed, the variances between the groups are roughly equal, the data are independent in each group, and some would argue the data need to be interval scaled (which I disagree with but that's what's in many stats materials unfortunately).

The t-test is what's called robust to violations of its assumptions. With that said, to test the assumption of normality the data can be graphed in a histogram or ideally in a normal probability plot to look for any marked deviations from normality. For equal variances, we would look at the ratio of the variances and as long as the ratio of standard deviation is less than about 2 we're fine. One could also conduct an F-test on the ratio's but these tend to be too sensitive and give too many false readings of different variances.

Part B
Question 1:  The strongest predictor of caregiver psychological distress is   Caregiving * * tasks per week as the absolute value of the correlation coefficient is .70, the highest among all other variables.

Question 2: The correlation coefficient of -.10 can be interpreted as a weak and negative correlation. Squaring -.10 provides us with an R-squared value of .01 or 1% of the variation between variables is accounted for in the relationship--which is really low.