How can I find tendencies in specific cases between two groups?

So, I'm doing this project where we are investigating whether social networks can influence peoples distinctiveness. Briefly summarized, we have primed a group with a social media and another is the control group. Afterwards, they were given a questionnarie where they have 1. Written 12 identity elements 2. Rated these elements on how central they are to their identity on a likert-scale from 1-7 where seven is very central to their identity (this scale is used throughout) 3. Rated these elements on how a) different they make them b) how it constitutes a special social role and c) how seperated these elements make them. We are not using the words further, just their ratings.

So far so good.

So, we have alot of variables in our database

1: Age
2: Sex
3: Priming or control-group
4: Centrality of the elements (is the mean of 2 questions about the centrality asked in different ways) (12 variables)
5: How different it makes them (12 variables)
6. How "roly" it makes them (12 variables)
7: How seperated it makes them (12 variables)

So, the way i have tackles this so far:
Found the mean of centrality for each element, multiplied it by variable 5-6-7 (look up) casewise and squarerooted the number (gives a 1-7 number). Then the mean of these three new variables (CentralityXDifference, CentralityXRole, CentralityXSeperateness) gives us a number of "Total distinctiveness". Unfortunately these means doesnt differ depending on priming or controlgroup, but that's okay. We can probably argue why.

Here comes some questions

1. What do i do if I want to check for tendencies that the priming or controlgroup accordingly have high scores in variable 5-6-7 when they have high in variable 4; Do the priming group tend to rate elements of high centrality as highly distinctive? I cannot figure out a way to do this in a smart way, so if you have an idea i'll be grateful.
2. Have I done something wrong the way i've done it already, or are there room for improvement in my method?

[Maguin, Eugene]

I may not understand completely what you want to do but have you looked at scatter plots (with the regression line plotted also) or correlations between relevant variables by group (primed or control)? You could categorize the variables into three or more groups and crosstab them by group to see what the cell distributions look like. You could build interaction terms (including group by variable interaction terms) and use them in regressions to check magnitudes and significance. It seems like you are looking for distribution and relationship differences between the two groups. So you have to work through comparisons of the within group distributions and relationships. For this effort, sample size is going to matter; I assume you have it.

Gene Maguin



-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of [hidden email]
Sent: Thursday, May 02, 2013 9:23 AM
To: [hidden email]
Subject: How can I find tendencies in specific cases between two groups?

So, I'm doing this project where we are investigating whether social networks can influence peoples distinctiveness. Briefly summarized, we have primed a group with a social media and another is the control group. Afterwards, they were given a questionnarie where they have 1. Written 12 identity elements 2. Rated these elements on how central they are to their identity on a likert-scale from 1-7 where seven is very central to their identity (this scale is used throughout) 3. Rated these elements on how a) different they make them b) how it constitutes a special social role and c) how seperated these elements make them. We are not using the words further, just their ratings.

So far so good.

So, we have alot of variables in our database

1: Age
2: Sex
3: Priming or control-group
4: Centrality of the elements (is the mean of 2 questions about the centrality asked in different ways) (12 variables)
5: How different it makes them (12 variables) 6. How "roly" it makes them (12 variables)
7: How seperated it makes them (12 variables)

So, the way i have tackles this so far:
Found the mean of centrality for each element, multiplied it by variable
5-6-7 (look up) casewise and squarerooted the number (gives a 1-7 number).
Then the mean of these three new variables (CentralityXDifference, CentralityXRole, CentralityXSeperateness) gives us a number of "Total distinctiveness". Unfortunately these means doesnt differ depending on priming or controlgroup, but that's okay. We can probably argue why.

Here comes some questions

1. What do i do if I want to check for tendencies that the priming or controlgroup accordingly have high scores in variable 5-6-7 when they have high in variable 4; Do the priming group tend to rate elements of high centrality as highly distinctive? I cannot figure out a way to do this in a smart way, so if you have an idea i'll be grateful.
2. Have I done something wrong the way i've done it already, or are there room for improvement in my method?



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/How-can-I-find-tendencies-in-specific-cases-between-two-groups-tp5719881.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[David Marso]

With a handle like MEGAMASTER4000 I would think you would be completely self sufficient.
Maybe you might upgrade to the MEGAMASTER5000 which has a real stats real easy cranial implant which would make this exercise moot.

[e-mail] wrote

So, I'm doing this project where we are investigating whether social networks can influence peoples distinctiveness. Briefly summarized, we have primed a group with a social media and another is the control group. Afterwards, they were given a questionnarie where they have 1. Written 12 identity elements 2. Rated these elements on how central they are to their identity on a likert-scale from 1-7 where seven is very central to their identity (this scale is used throughout) 3. Rated these elements on how a) different they make them b) how it constitutes a special social role and c) how seperated these elements make them. We are not using the words further, just their ratings.

So far so good.

So, we have alot of variables in our database

1: Age
2: Sex
3: Priming or control-group
4: Centrality of the elements (is the mean of 2 questions about the centrality asked in different ways) (12 variables)
5: How different it makes them (12 variables)
6. How "roly" it makes them (12 variables)
7: How seperated it makes them (12 variables)

So, the way i have tackles this so far:
Found the mean of centrality for each element, multiplied it by variable 5-6-7 (look up) casewise and squarerooted the number (gives a 1-7 number). Then the mean of these three new variables (CentralityXDifference, CentralityXRole, CentralityXSeperateness) gives us a number of "Total distinctiveness". Unfortunately these means doesnt differ depending on priming or controlgroup, but that's okay. We can probably argue why.

Here comes some questions

1. What do i do if I want to check for tendencies that the priming or controlgroup accordingly have high scores in variable 5-6-7 when they have high in variable 4; Do the priming group tend to rate elements of high centrality as highly distinctive? I cannot figure out a way to do this in a smart way, so if you have an idea i'll be grateful.
2. Have I done something wrong the way i've done it already, or are there room for improvement in my method?