tutorial 007 Compariing group means anova

welcome to the SPSS tutorial video for comparing group means with ANOVA in this video we'll be covering one way ANOVA with two levels one way ANOVA with three or more levels including post-hoc tests and two-way anova well begin by looking at the data set we're using the therapy day that we've used in previous videos and the variable will be using today are depression expand this label depression at time 2 will be using the therapy whether they are in therapy yes or no 0 and 1 as well as which therapy group people are in whether they're in the control or no therapy CBT therapy or humanistic therapy groups the first analysis will do is a one-way anova with two levels in this case we'll be looking at the effect of the therapy variable with two levels yes or no whether they're in therapy on the depressions are under analyze will go to compare means and one-way anova will include our depression at time 2 as our dependent variable then we'll include our therapy variables under our factors we also want to go to the options menu and ask it to report our descriptive statistics and we click OK we see that we get our descriptive statistics the mean depression at time 2 for those who are not in therapy as well as our mean depression at time 2 for those who are in therapy and we get an anova or an f-test with significant value to show us that there is a statistically significant difference between those who are in therapy and not in therapy as you'll notice from the previous video this s value is the exact same as the t-test with independent samples this is only the case when we use a one-way anova with only two looks so what happens when we have more than two levels look at this question will come to our one-way anova instead of having our two-level therapy will ask for our therapy type which includes cbt control and human ethics again we want to make sure that our descriptive statistics are included and unlike with two levels will be going into the post hoc command asking for our LSD or our bonferroni's whichever method your instructor requires tosok test when we scroll down in the menu we get the means of depression at time 2 for our control those are not in therapy for those who are in CBT therapy as well as for those who are in humanistic therapy our F test tells us whether there is a difference between any of the three and down below because we requested our post hoc test we see the pairwise comparisons between all groups that is the comparison between control and cbt is statistically significant the comparison between control and humanistic is statistically significant as well as the comparison between CBT and humanistic is statistically significant in this case we have a main effect for therapy and all pairwise comparisons are different now that we've seen how to look at inova or analysis of variance for a one variable with two levels and one variable foot three levels let's look at how to use a no vote with more than one variable and both will be between groups factors I'm going to analyze instead of going to compare means we'll go to the general linear model with univariate in this case you see more potential boxes to fill with variables our dependent variable will remain depression at time to our fixed factors is our independent variable scope here let's use their therapy type again our three levels of therapy as well as participants X we also see several more options to get post-hoc tests we can either go the box that we did before in this case I won't choose that we can also go under options and in addition to including our descriptive statistics we can select over each of our variables including the interaction term and compare the main effects this also generates the same post hoc tests that we would have gotten before let's also include our estimates of effect size and continue so I run this model we see first that we get our sample size for those in the control group c bt group and humanistic we also see sample size for participants x how many males and females we have in addition we get descriptive statistics or mean depression scores at time 2 for males in the control group for females in the control group males in the cbt group females in the cbt group as well as males and females in the community group and our overall averages by gender accordingly for our tests of between subjects effects we can see the main effect of therapy type which is statistically significant and has a strong effect size we can see the main effect for participants X in this case called female which is not statistically significant we can also see the interaction between therapy type and sex which is also not statistically significant we can scroll down below under the section called estimated marginal mean to examine our post hoc test we see the estimates or the mean for the therapy types and the specific comparison between them this is very similar original notice to our one-way anova except that a controls full participant gender we also get participants x means and the direct comparisons between them in this case there is no overall difference between males and females lastly we see the interaction between therapy type and participants X we see the mean for males and females within each therapy type which are not statistically different is indicated by the test in this video now we've seen one way ANOVA with two levels one way ANOVA with three or more levels as well as a two-way between subjects ANOVA

Channel: Dustin Thoman

Views: 11

Rating: 0 out of 5

Keywords:

Id: YRcBq15GBDQ

Length: 6min 13sec (373 seconds)

Published: Thu Feb 09 2017