Hello Lara,

As you indicate in your query, there are all sorts of examples (no two seemingly the same!) of reporting available in the literature.

Minimally, it would make sense to report on: (a) the second (BAS, FFFS) addition result at stage two; and (b) overall (I presume stage 2 is the end) model result, with discussion about individual IVs that were/were not significant. As to emphasis and sequence, I suggest you be guided by your research question(s).

It wouldn't be surprising to me that the stage 2 addition was the most salient point for your work. So, you'd obviously want to discuss that first. And so on.

Specific Qs:

1. Yes, the "full" model p value would come from the anova table with all IVs (4, apparently) included.

2. Yes, the stage 2 test of whether delta R-square (change in R^2) was significant would/should be included at the spot you indicate.

Finally, though you didn't ask, your summary statistics should include mean, SD, and correlations among all var's (IVs and DV).

Good luck with your work.

I am confused about the two reports you give on the probability: p (Sig. F Change) = .036 and p (Sig.) = .065.

In general, the information you are reporting on the change in R-square is what is expected, and there is no need for an ANOVA table.

Thank you for your reply David,

I currently have:

After the addition of BAS and FFFS in Step 2 the total variance explained was 20%, R2adjusted = .11, F (4, 40) = 2.415, p = .036 (sig. f change). The two measurements explained an additional 15% of the variance after controlling for gender and years of experience, F change (4, 40) = 3.620, p = .065 (sig.)

From your answer I am assuming the different sig. values are currently in the wrong places in my paragraph?

For example, correction:

First point--> (R2adjusted = .11, F (4, 40) = 2.415), the significance value should be p = .065. Therefore, the model as a whole is non-significant?

Second point--> ...after controlling for gender and years of experience, F change (4, 40) = 3.620, p = .036, therefore the variables added in that step did not significantly improve the prediction?

Hi David LM (latest reply),

In the model summary there is 'Sig. F Change' and in the ANOVA table there is 'Sig.'

I am confused about what both of them mean and how to report them. I have used the SPSS manual's chapter on hierarchical multiple regression, however their numbers are the same for these values, therefore I don't understand whether to report 'Sig.' with the whole variance explained by model/step 2, or vice versa.

From the reply above, I am assuming that when reporting the total variance explained by my two variables I need to report the Sig. value (reported in the ANOVA table on SPSS), and when detailing the additional variance explained by the variables (R2 change), I need to report the Sig. F change value..?

I have attached the Model Summary and ANOVA tables from my hierarchical multiple regression analysis for reference.

Hello Lara,

The F for testing R^2 change in stage 2 should have 2, 40 df associated with it. (assuming your data set had 45 total cases). The F for the final, "full" model would have 4, 40 df (assuming two IVs in stage 1, and adding two more in stage 2). This latter result should be identical to what you'd get if you ran a single model with all 4 IVs included.

Good luck finishing your report.

David Morse

Thank you, I understand the df values, however I am asking about the Sig and Sig. F change values and where they should be.

The first table is for the change in R-sq across four regressions. In the first regression, your predictors are not significant. In the second and third regressions, each of the sets of variables that you add makes a significant further contribution to the R-sq. The variable(s) in the fourth equation do not add a significant amount of explained variance.

The second table is for each regressions as a whole. The R-sq for the fourth equation is significant because it includes the significant variables that were added at steps 2 and 3.

David L Morgan

Thank you,

I understand this, but when reporting these values, does the sig. f change statistic need to be in the sentence with the F change (...) = ..., and therefore the sig. statistic would be with the F (.....) = ...?

Regardless of your statistical values, the reporting standard in APA Style is first describing the analysis result in words and then writing the decision statement. Examples are illustrated by Cooper (2020) and Kahn (2020). Full citations are below.

Cooper, H. M. (2020). Reporting quantitative research in psychology: How to meet APA Style journal article reporting standards (2nd ed.). American Psychological Association. https://www.apa.org/pubs/books/reporting-quantitative-research-psychology-second-edition-revised

Kahn, J. (2020, September 3). Reporting statistics in APA Style. My Illinois State. https://my.ilstu.edu/%7Ejhkahn/apastats.html

Good luck,

Hi Lara, my recommendation is that in a table, you report your summary statistics (b, SE(b), and B, Listing in the first column your predictors and in the columns, your results by Step, as you estimated (i.e., Col1 to Colx corresponding to Step 1 to Step x respectively.).

Then, you will discuss your models,. In your write-up, that is, using words, you will explain your models results, the way you stated in your posting. That is, write up your results for Step 1. Then you write up your Step 2: After the addition of BAS and FFFS in Step 2 the total variance explained was 20%, R2adjusted = .11, F (4, 40) = 2.415, p = (?), and lastly, write-up your Step 3. The notation needed for Step 1 and 3 are similar to the one you summarized in Step 2.

The two measurements explained an additional 15% of the variance after controlling for gender and years of experience, F change (4, 40) = 3.620, p = (?)

Thus, note that the Table, is where you report your betas, and SE(b). In words, you provide R2, F, df, and p-value.

I hope this is helpful. Please feel free to let me know if I can be of further help. Good Luck!

Here is an example of a table that illustrates how to report a set of hierarchical regressions.

Patricia Rodriguez de Gil

Thank you, this is very helpful!

David L Morgan

Thank you!