Can you suggest a graphical presentation for a Categorical DV and 4 Continuous IVs.
=====================
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
Thank you. Lema |
Dear Lema Questions like yours are best answered with the rationale and case scenario(s) already fully explained. Best wishes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: SPSSX(r) Discussion <[hidden email]>
On Behalf Of 3J LEMA This email was sent to you by someone outside the University.
You should only click on links or attachments if you are certain that the email is genuine and the content is safe. Can you suggest a graphical presentation for a Categorical DV and 4 Continuous IVs.
Thank you. Lema ===================== 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 elaborate on the previous good advice:
What are the variable names? For the categorical variable, what are the value labels? What questions do the visualizations address? For whom are you producing the visualization? Just checking for potential data entry error? Formal presentation in court of law, etc. "categorical" can refer to coarsened continuous variables, nominal level variables, ranks, etc. What constitutes a case? How many cases do you have? *And so forth.* ----- Art Kendall Social Research Consultants -- Sent from: http://spssx-discussion.1045642.n5.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
Art Kendall
Social Research Consultants |
In reply to this post by 3J LEMA
Speaking abstractly -- meaning, perhaps, special software -- you
may represent continuous numbers with light/dark and rainbow
colors, for two dimensions. Are you plotting points? or functions?
Assuming you are plotting data points rather than a function, you
may use length and width of an object for two more, though these
are often combined as "size". The complex shape of the object is
very often used for categories (circle, square, star, ...). Of course,
this is less usable for hundreds of points than for dozens.
X and Y are the usual axis labels for two continuous variables, and
Z is sometimes represented effectively, one way or another.
--
Rich Ulrich
From: SPSSX(r) Discussion <[hidden email]> on behalf of 3J LEMA <[hidden email]>
Sent: Sunday, June 13, 2021 8:30 PM To: [hidden email] <[hidden email]> Subject: Visualization for Categorical DV and Continuous IVs Can you suggest a graphical presentation for a Categorical DV and 4 Continuous IVs.
===================== 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
Thank you.
Lema
|
In reply to this post by 3J LEMA
It really depends on what you want to show, but one possibility would be to fit a multinomial regression and then plot the category probabiliities against the regressors or something like emm's, On Sun, Jun 13, 2021 at 6:30 PM 3J LEMA <[hidden email]> wrote:
|
I haven't done the multinomial regression that Jon suggests, but
what I have used when the categories are relevant is the
discriminant function for multiple groups. It has nice ancillary
statistics and displays.
The plot of category centroids on the first two discriminant factors
shows the group separation, and the factors can be interpreted by
their loadings like a factor analysis. The "within-group" correlations
are also available.
--
Rich Ulrich
From: SPSSX(r) Discussion <[hidden email]> on behalf of Jon Peck <[hidden email]>
Sent: Monday, June 14, 2021 11:32 AM To: [hidden email] <[hidden email]> Subject: Re: Visualization for Categorical DV and Continuous IVs It really depends on what you want to show, but one possibility would be to fit a multinomial regression and then plot the category probabiliities against the regressors or something like emm's,
On Sun, Jun 13, 2021 at 6:30 PM 3J LEMA <[hidden email]> wrote:
|
Free forum by Nabble | Edit this page |