I have two groups of people that have comenced two trials. Both trials recorded their heart rates 4 times. I'm just not sure now on how to analys this data to get the best information from it. Would I used a two way anova or a mixed anova?
The more I try to work it out the more confused I am getting.
Thank you
I don't have the whole picture, but to me it seems like time is your first independent variable (IV), within-subjects and with four levels and there is a between subjects IV with two levels (two groups, I imagine this division is based on a differential trait).
The answer would be two way (mixed) ANOVA with one between-subjects variable (two levels: groups) and one within-subjects variable (four levels: times).
Thank you for helping me.
I'll give you more information. The two groups are runners and rowers, they both did treadmill running on 4 levels for each two trials.
Where I am getting confused is when I am inputting the data into spss, would the within subject factor be the two trials and then also the 4 intensity levels and then the between subject be the two groups?
SPSS doesn't change the nature of the variable, so the two trials would still be between-subject (each subject "performed" one and not the other).
Is it possible your confusion comes from how data is entered in SPSS. The between-subject variable has a unique column in which you would enter "group 1" or "group 2". The within-subject should be entered as one level per column. This is particular to SPSS. Does this help to solve the problem?
Dear Gracie,
To answer your question, I think a concrete example is necessary, including how to enter or ‘organize’ the datafile. The following lines create a hypothetical dataset with 2 runners and 2 rowers, that each have performed 4 tests, presumably at increasing ‘levels’. The dependent variable is the resulting heart rate at each level. In that case, ‘Group’ is the between subjects factor, and ‘Level’ the within subjects factor, and ‘Heartrate’ is the dependent variable. The syntax to enter and analyze the hypothetical 4 cases is:
---------------------------------------------------------------------
data list free /Case, Group,
Heartrate_at_Level_1, Heartrate_at_Level_2,
Heartrate_at_Level_3, Heartrate_at_Level_4.
begin data.
1 1 20 25 50 75
2 1 15 45 65 85
3 2 35 36 35 36
4 2 38 39 38 39
end data.
value labels Group 1 'Runners' 2 'Rowers'.
glm Heartrate_at_Level_1, Heartrate_at_Level_2,
Heartrate_at_Level_3, Heartrate_at_Level_4 by Group
/wsfactor=Level 4 polynomial
/measure=Heartrate
/plot=profile(Level*Group)
/print=descriptive
/wsdesign=Level
/design=Group.
---------------------------------------------------------------------
Amongst others, the output results in the plot (attached):
📷
Additionally, the output will say, in this hypothetical case, that the two groups do not differ in their heart rate on average (all levels taken together; main effect of ‘Group’ is n.s., p=0,182). The linear increase in heart rate among the Runners, however, is clearly and significantly higher than the linear increase in heart rate among the Rowers (interaction Level * Group, linear component, is significant, p=0,009).
Note that this is not the only way to ‘organize’ the data. Here we chose to enter the heart rates of each case ‘horizontally’, creating a so called ‘wide’ file. It is becoming increasingly popular, however, to enter the heart rates ‘vertically’, creating a so called ‘long’ file (i.e., four rows per case). The syntax given above only applies to the ‘wide’ setup. When you enter the heart rates ‘vertically’ and create a ‘long’ file, the analysis will be ‘multilevel’, with ‘subject’ nested under ‘level’ (in this case). Both the ‘wide’ and the ‘long’ setup can be analyzed by all major statistical software packages such as SAS, Stata, R, and SPSS. The results, however, are nearly similar, as we cannot change the world by choosing a slightly different type of statistical test, nor by using another software package.
- Badges
- Science topic