goodness of fit
Locked
 
|
Hi folks,
I have the following data in SPSS…the four columns are: length bin, total catch, observed probability of capture, and predicted probability of capture. I have calculated the probability data using logistic model in Excel (along with solver) to get an s-shaped selectivity curve. Now I have the data in SPSS and I want to test how well does the predicted model fit the observed data. I thought this would have been a simple procedure but I am stumped. Any help would be appreciated.
57.00 3.00 0.00 0.00144 62.00 5.00 0.20 0.00337 67.00 10.00 0.10 0.00782 72.00 12.00 0.08 0.01807 77.00 20.00 0.05 0.04118 82.00 13.00 0.46 0.09110 87.00 23.00 0.61 0.18960 92.00 41.00 0.61 0.35320 97.00 48.00 0.81 0.56036 102.00 91.00 0.82 0.74843 107.00 87.00 0.84 0.87412 112.00 91.00 0.92 0.94189 117.00 59.00 0.83 0.97425 122.00 19.00 0.89 0.98880 127.00 9.00 0.89 0.99517 132.00 4.00 0.25 0.99793 137.00 2.00 1.00 0.99911 142.00 3.00 1.00 0.99962 147.00 1.00 1.00 0.99984
Wade
Hiscock Tel.: (709) 778-0366
|
|
|
Hi folks…still looking for some advice on problem below….
From: Wade Hiscock
Hi folks,
I have the following data in SPSS…the four columns are: length bin, total catch, observed probability of capture, and predicted probability of capture. I have calculated the probability data using logistic model in Excel (along with solver) to get an s-shaped selectivity curve. Now I have the data in SPSS and I want to test how well does the predicted model fit the observed data. I thought this would have been a simple procedure but I am stumped. Any help would be appreciated.
57.00 3.00 0.00 0.00144 62.00 5.00 0.20 0.00337 67.00 10.00 0.10 0.00782 72.00 12.00 0.08 0.01807 77.00 20.00 0.05 0.04118 82.00 13.00 0.46 0.09110 87.00 23.00 0.61 0.18960 92.00 41.00 0.61 0.35320 97.00 48.00 0.81 0.56036 102.00 91.00 0.82 0.74843 107.00 87.00 0.84 0.87412 112.00 91.00 0.92 0.94189 117.00 59.00 0.83 0.97425 122.00 19.00 0.89 0.98880 127.00 9.00 0.89 0.99517 132.00 4.00 0.25 0.99793 137.00 2.00 1.00 0.99911 142.00 3.00 1.00 0.99962 147.00 1.00 1.00 0.99984
Wade
Hiscock Tel.: (709) 778-0366
|
|
|
In reply to this post by Wade Hiscock
Wade,
I am not familiar with this area of research, which in part might be contributing to some of my confusion regarding your post. Perhaps someone else can provide an adequate response based on the information you've provided thus far. However, in order for me to help you, I would need to know exactly how you calculated the observed probabilities. Also, would you mind operationalizing your dependent variable? Typically, when I perform a logistic regression analysis, the dependent variable is either binary (e.g., 0=failure/1=success) or binomial (e.g., # of successes / # of trials). In addition to clarifying the dependent variable, could you please describe the predictors in your model, if there are any?
I would prefer to gather some more information before diving into a discussion about goodness of fit for your particular model.
Ryan
On Thu, Oct 28, 2010 at 12:11 PM, Wade Hiscock <[hidden email]> wrote:
|
|
|
|
|
In reply to this post by Wade Hiscock
Hi
Wade,
Contrary to
friends who need more details of your analysis, there is no need to know
anything more of how you did it. All needed are the data provided. If only
your prob.obs and prob.pred are provided then there are several SPSS procedures
to see goodness of your fit. I also added a quick logistic SPSS fit, normal fit,
....etc You can find more on SPSS capabilities.
Max.
Attached:
Wade.Hiscock.logistic.pdf
|
|
|
In reply to this post by Wade Hiscock
Hi Luciano,
Attached pdf. |-----Original Message----- |From: luciano basso [mailto:[hidden email]] |Sent: Friday, October 29, 2010 18:10 |To: [hidden email] |Subject: Re: goodness of fit | | |Hi Max, | |How can I access your attachment? | |thanks | | |Luciano | | |2010/10/29 MaxJasper <[hidden email]>: |> Hi Wade, |> |> Contrary to friends who need more details of your analysis, there is |> no need to know anything more of how you did it. All needed are the |> data provided. If only your prob.obs and prob.pred are provided then |> there are several SPSS procedures to see goodness of your |fit. I also |> added a quick logistic SPSS fit, normal fit, ....etc You can |find more |> on SPSS capabilities. |> |> Max. |> |> Attached: Wade.Hiscock.logistic.pdf |> |> -----Original Message----- |> From: SPSSX(r) Discussion [mailto:[hidden email]] |On Behalf |> Of Wade Hiscock |> Sent: Friday, October 29, 2010 5:41 |> To: [hidden email] |> Subject: FW: goodness of fit |> |> Hi folks…still looking for some advice on problem below…. ===================== 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 |
Administrator
|
Those of us reading via Nabble don't get attached files. Do you have a URL for it? Thanks.
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
|
At 09:30 PM 10/29/2010, Bruce Weaver wrote:
>MaxJasper wrote: > >>Hi Luciano, >> >> Attached pdf. > >Those of us reading via Nabble don't get attached files. I don't think any recipients of SPSSX-L see attached files. I think they're removed by ListServ at UGA before the postings are distributed anywhere. A URL is good, or simply include the text in your posting. ===================== 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 |
| Free forum by Nabble | Edit this page |