I'm not fully clear on the design, perhaps you could elaborate a bit on your initial description?
1. The two time points: Is this a repeated measures (within-subjects) factor, or were cases randomly assigned to be measured at Time A vs. Time B?
2. The three temperature levels: Is this a between-subjects factor (cases assigned to one temperature only), or a within-subjects/repeated measures factor?
3. What does "untreated" mean: No temperature exposure?
If the design is a one between-subjects (4 levels: temp A, temp B, temp C, control) and one within-subjects (2 levels: time A, time B) arrangement,
Then in spss, go with (from the main menu bar): Analyze / General Linear Model / Repeated Measures...
In the subsequent dialog box, give the within-subjects factor a name (e.g., Time), and indicate the number of levels (2). Click "Add," then click on the "Define" button at the bottom.
In the next dialog box, pull the two time point measures into the top box where the question mark/blanks are, then add Group as the between subjects factor (where I'm presuming there are 4: 3 different temperature treatments plus the comparison. Note that this is a single variable having 4 levels, so be sure you have such an indicator ready in your data set.)
-The control for my study was uncooked sample (so no temperature or time value can be associated with it)
-Temperature has 3 levels (20, 40, 60 degrees) and time has 2 levels (10, 50min), giving 6 treatment combinations (3 temperature× 2 times).
- however because I also want to determine whether at the same temperature (e.g.- 20 degrees), whether there was any change due to increased time (10 to 50min), can I use repeated measure for this analysis ?
- the control in this study sits outside the experimental block as it does not have either time or temperature value, can it be included directly in analysis or shall I use Dunnett's test to determine whether it has any significant difference when compared to my samples.
1. This description ("Temperature has 3 levels....) makes the design sound like a between-subjects design, not a "mixed" or within-subjects (repeated measures) design. So, ordinary factorial anova will permit comparisons by temperature, time, or their interaction/combination to be evaluated.
2. Change due to increased time would ideally be evaluated only by repeated measures (e.g., once at 10 min, again at 50 min). Difference due to time can be evaluated via the ordinary factorial design (as in #1, above).
3. You could, if the spirit moved you to do so, run additional t-tests (perhaps with Holm's adjustment, or Benjamani-Hochberg adjustment) comparing any given time/temperature combination results with the control/uncooked sample results.
Thank you so much again. But I am still not a 100% sure on how to perform the statistical analysis. Is it possible for you to please have a look at the file attached and just explain it to me.
First, I would like to correct on your notation. 2^3 factorial design implies three (3:A,B,C) each at two levels. You wrote two time points at three temperature levels instead of Three Temperature (Factors, and a control) at two levels (time points each), that is 2^3x2 or 2^4 Factorial Design.
#NOW two consideration, that may lead to your research questions being answered:
1. If it include repeated measure, NO! From your attached docx file; the sample was allocated for the treatment and control, i.e., three temperatures (factors) of 20, 30 and 50(units?) and a control temperature (lab temp)
2. Using a 2 Way ANOVA rather than 3-way ANOVA in SPSS you can compare the mean effects of the temperature & time {see https://statistics.laerd.com/spss-tutorials/two-way-anova-using-spss-statistics.php}
You need to rearrange your table to include a fourth column for the control temperature (no temperature treatment)