Statistics help...

jbellni

Hi there,

I was wondering if anyone could help me with deciding on which statistical test to choose in my psychology project. My project is on bullying and whether people in different social acceptance categories get bullied more or less. My independent variable has 4 levels of categories (accepted, neglected, rejected and controversial) and the dependent variable is a bullying score ranging from 9 to 45. I'm using SPSS version 15. Any help would be MUCH appreciated.

Josh.

jbellni

It's ok. I figured it out. Still trying to understand parts of the output though!

2Scoops

I can help with your interpretation, if you like. I'm pretty good with SPSS.:cool:

2Scoops

Sent you a PM:)

Scigaithris

jbellni wrote:

Hi there,

I was wondering if anyone could help me with deciding on which statistical test to choose in my psychology project. My project is on bullying and whether people in different social acceptance categories get bullied more or less. My independent variable has 4 levels of categories (accepted, neglected, rejected and controversial) and the dependent variable is a bullying score ranging from 9 to 45. I'm using SPSS version 15. Any help would be MUCH appreciated.

Josh.

Josh... need more details on your research design. Sorry if I am late in responding, but been busy.

Details for stat formula selection needed:

(1) If you have a representative sample of a population that was randomly drawn, or a nonprobability sample?

(2) Size of sample (n)?

(3) Regarding the items in your questionnaire, were these yes/no nominal measures, or ordinal level opinion scales, or answers that produced interval or ratio level data (e.g., number of times bullied), or mixed measures between nominal-ordinal-interval-ratio level data?

(4) Were you testing for significant differences (e.g., differences between groups), correlations (i.e., associations between independent and dependent variables), regressions (i.e., predictions), or what?

These are just the bare bones for formula decision making...

jbellni

scigaithris, sent you a pm.

jbellni

Another quick question about the writeup this time. After doing an ANOVA out of

-Descriptives
-Test of Homogeneity of Variances
-ANOVA
-Post Hoc Multiple Comparisons
-Homogeneous Subsets
-Means Plots
-Any other graphs

What should I include in my results section?

2Scoops

jbellni wrote:

Another quick question about the writeup this time. After doing an ANOVA out of

-Descriptives
-Test of Homogeneity of Variances
-ANOVA
-Post Hoc Multiple Comparisons
-Homogeneous Subsets
-Means Plots
-Any other graphs

What should I include in my results section?

Descriptives, ANOVA, post hocs, graphs.

BUT don't just copy and paste the tables because they include lots of non-essential information. Descriptives are only really necessary to mention briefly in text - especially if you have drawn the graphs, which provide the exact same information.

You'll want to state the significance of the ANOVA F-test (and maybe the f-value itself but it's non-essential). Post hocs are a must once you have a significant ANOVA - it's where the interesting info lies - but, again, the information can be included in the graphs with stars/asterix notation.

The usual way would be to put all the salient information in a table and a couple of graphs then use the text to expand on it briefly. Don't add any critical analysis or interpretation of the results yet, though - that's for the discussion.

jbellni

2Scoops, you continue to be a legend. One final question (I promise, only one more!), what do my results actually mean in a real world situation? How do they apply to and benefit the world of psychology?

2Scoops

That's a question that can't be answered by the simple rules of maths:D Maybe the first thing to do is find out if the significant differences are meaningful e.g. a significant difference of 0.1 on the scale you used may not translate into any kind of practical difference in the real world. Find out how much of a difference there is between groups, then figure out if it's big enough to mean anything based on the results of others that have used the scale (assuming it has been used before).

The benefit to the world of psychology?! Sounds like the usual canned BS question so probably deserves a BS answer. If your results are novel (no one else has seen the differences before) then it adds to the existing knowledge. Or, if your sample is sufficient novel and no one has seen the results in a similar group before, it can support the fact that they exhibit similar patterns to other groups or that they appear different to other groups. As long as there is something novel in there, you can talk it up as adding to existing knowledge. How important the new information is to the world of psychology is anyone's guess:)