Sawtooth Software Forum - Recent questions and answers

Any updates to SSI Web Data Generator planned?

Wed, 16 May 2012 19:59:04 +0000

I've recently started using the Data Generator tool with SSI version 7 (probably v8 soon). Wondering if there's any plans for future updates / enhancements? It's a very useful tool so I really hope so. However I think there has not been a new release since early 2010.

Something I'd really like to see is better handling of multiple response questions, namely in having valid randomized responses to questions with exclusive choices, and following limts on min/max number of answers.

Also, minor thing, but for some reason on my install the "Simultaneous Respondents" and "Drop Out Rate" text boxes are really tiny so you can 't see what's in them.

Answered: Is zero treated as missing data in user defined variables (CBC/HB)?

Wed, 16 May 2012 12:52:06 +0000

Zero is a missing value.

Answered: How do I generate different links for each segment of my target group?

Tue, 15 May 2012 12:08:34 +0000

You can use pass-in fields, as Bahadir mentioned, or you can use merged fields. Pass-in fields are used to append a value for a variable as part of the linke, i.e. gender=1, so it's kind of pre-answering a question when the person starts the survey.

Merged fields are similar, but they are tied to a password. For example, if you knew I was going to take your survey using the password 1234, you could set up some merged fields to populate things like FirstName=Brian, Gender=2, and so forth. When I type in 1234 to start the survey, those variables are automatically populated according to the merged fields you have set up.

For short video on creating links that have passwords and pass-in fields in them, please see http://www.sawtoothsoftware.com/support/issues/ssiweb/help_movies/pass_in_info.htm.

To read a bit more about passwords versus pass-in fields versus merged fields, you can read the help files in SSI Web or online at www.sawtoothsoftware.com/support/issues/ssiweb/online_help/hid_web_passwords_intro.htm

Answered: Sliders For Grids

Tue, 15 May 2012 11:17:29 +0000

Hi Nate,

I don't know if it this is helpful to you, but I got to put a jQuery slider to work in a free format question in version 7.0.26. Here you can see the details: http://www.sawtoothsoftware.com/forum/1214/using-jquery-inside-ssi-web

Hope it can help you, regards,
João

Answered: Data Coding in CBC/HB

Mon, 14 May 2012 21:45:24 +0000

The choice of "Task Weight" comes down to this: to what degree do respondents tend to push all their "chips" to just their favored alternative (task weight should be much closer to 1), vs. distributing their chips over many favored options (task weight should be larger than 1, and perhaps as high as 10).

Answered: How do I best include different languages in a CBC Survey?

Mon, 14 May 2012 20:09:14 +0000

First of all, all the non open-end questions are coded with values (at least v6 and v7). so you can basically accumulate all the data to one place (let's say english one) and when you export, your tables will be in english.

as for translation, make several copies of your ssi study folder, then either copy/paste the whole text in each question or use the export for translation tool and a freeware tool such as virtaal to translate it and then re-import. this process overrides the original file without much chance to undo, so make sure to backup!

Answered: Different results in Hierarchical bayes and OLS with CVA

Mon, 14 May 2012 18:02:14 +0000

I don't like focusing ones attention on importance scores, which are not as useful or meaningful as part-worth utilities or especially market simulations.

Importances are kind of a strange calculation based on the maximum difference observed between best and worst levels for an attribute, at the individual level. But, if there is an attribute with little impact for a respondent, then just random noise for the utilities of levels for that attribute will drive a positive importance score. "Reversals" (levels out of order from expected rational preferences) still are counted as a positive weight toward the importance score.

HB tends to reduce reversals at the individual-level, and should give a cleaner view of attribute impact, in my opinion.

But, you could also estimate your utilities by imposing utility constraints (monotonicity constraints) for either OLS or HB. And, that should also sharpen the importance calculation.

Again, I'm not a fan of importances.

And, the differences you are noting seem relatively small, to my eyeball.

Attributes 1 and 3 seem about tied, and attributes 2, 4, and 5 seem about tied. So, there could be some trading around of rank orders due to minor changes in the absolute importance scores.

Answered: Radio button auto selected when text entered in textbox/textarea

Mon, 14 May 2012 12:00:41 +0000

What you want to do should be possible with something like:

var myTxtBox = document.getElementById("txtBox"); // your question name here
myTxtBox.addEventListener("onkeydown", textclick, false);

function textclick()
{
// SSI code to set radio button. 
}

Answered: Exclude estimation of "None" in aggregate logit model in SMRT

Thu, 10 May 2012 14:18:14 +0000

It's not available as an option in SMRT's logit. But, it is available (to exclude tasks with None answers) in our Latent Class standalone software (and a 1-group LC estimation is the same as an aggregate logit run).

The other option is to modify your .cho file or .csv file that holds respondent answers, and change all answers of "None" to answers of zero. Answers of zero tell our utility estimation software programs to skip that task.

Answered: How do I export ACBC data to .cho file from SSI web version 8.0.2?

Thu, 10 May 2012 14:09:05 +0000

File + Data Management.

Go to Export Data tab.

Edit or Add a job.

To that job (on the Export Job dialog) click Add and you will see an option to add an ACBC .cho export.

Then, export that job, and a .cho file will be exported to a subfolder in your project folder (you are told what the subfolder name is in the export job report).

Answered: How to get field statistics when using several CAPI packages?

Thu, 10 May 2012 05:18:28 +0000

Nice work u started in CAPI package. In ur work u want to do store the Database in the package and study about the performance of interviews. In Ms-Access there is a option for store the database and also create the key which is useful for to protect ur databse. And also some statistical packages like SPSS, STATISTICA, MINTAB these packages are useful for to study the performance for attributes and give the numeric result which is more scientific accuracy value. if u r interested u can try this.

Answered: Add text next to radio buttons in CBC

Wed, 09 May 2012 14:21:35 +0000

Add a function to replace innerhtml of your radio buttons on page onload event, something like
Cel.innerHTML = Cel.innerHTML + "blah"

Answered: How do you code log-linear pricing?

Wed, 09 May 2012 13:44:03 +0000

Yes, replace price levels with natural log of price. None needs no special treatment.

Answered: Slider question

Wed, 09 May 2012 05:15:02 +0000

I can't check the code now, but you can use javascript verification to make sure your variable is not -1 to continue.

As for -100 100 range, can't you just change -100 to 0?

Are statistical results of part worth utilities transferable to WTP results based on part worth utilities?

Tue, 08 May 2012 10:24:35 +0000

Hi there,

I calculated the willingness to pay for attribute levels based on the part worth utilities and using SMRT. Now, I am wondering if the statistical results regarding the part worth utilities (e.g. correlations) are transferable to the WTP of the attribute levels as they are calculated based on the utitlities. Can somebody tell me if this assumption is correct or mistaken, please?

Best,
Thomas

Ci3 data saving wierdness

Mon, 07 May 2012 17:09:58 +0000

Never seen this happen before:

Two projects called Mary and George (not their real names).

Finished Mary last month. Started George and finished it this month.

Took a look at the data in George and a bunch was missing. Couldn't find the data until looked at the Mary project and saw that her data got bigger and the date/time stamp changed.

Sure enough, data from George tried to save into Mary (but the surveys were different so all that saved was some administrative variables which were common to both).

Running on a network using Ci3 v2.6.16. SSIfield and .Ci3 files looked fine (pointed to the correct projects). The server was changed just prior to beginning George and there is some suspicion that that had something to do with it (e.g., permissions or some-such).

Question is has anyone ever encountered this or a similar issue and what was the resolution?

Answered: 2 issues exporting lists do data set

Fri, 04 May 2012 10:41:49 +0000

Just guessing, but the constructed list may not be populated if it is never used...
You can export the predefined lists to text by clicking the print button on the lists userform

Answered: Sawtooth latent class gives T ratio and standard errors, what to do with it? how check for significance?

Thu, 03 May 2012 20:26:54 +0000

What typically one would do would be to look at the fit statistics overall for the Latent Class run. You will get a Log-Likelihood for the Latent Class model. You can compare the LL for the latent class models with main effects only vs. the LL for the result when you include interactions. (Make sure to keep the number of classes constant between the models).

I may be wrong here, but the statistical test would involve counting the number of parameters added to the model times the number of segments in your LC solultion (somebody please correct me if I'm wrong).

So, if you were adding a 2-way interaction between an attribute with 5 levels and another with 4, there would be an additional (5-1)(4-1)=12 parameters to the model. But, since we have three groups, I think it means there are actually 12*3 = 36 degrees of freedom for our statistical test (the 2-LL test).

So, if you saw a difference in LL of 10 points between the model with and without interactions, then the Chi-square test involves a critical value of 10x2 = 20 with 36 degrees of freedom. Use the =ChiDist(20,36) formula in Excel to compute the p-value. A p-value of 0.05 or less would indicate that the interaction effect was adding significant fit.

Answered: Trying to use pass-in field in quota - any ideas?

Wed, 02 May 2012 20:28:19 +0000

Found that logic was skipping me over the quota rather than through it. The pass-in information was/is utilized in the same manor as an answered question, but a participant must pass through a quota question for it to tally. If skip logic skips a participant over the quota page, then it will not tally.

Thought I would put it out there in case someone stumbles onto this thread in search of an answer.

Answered: Correlation and MaxDiff

Wed, 02 May 2012 03:36:28 +0000

Yes, I think you can...though you are taking estimates from one model and feeding those into a second model. Purists don't like model results that feed into other models.

And, you'll need to worry about normalizing the respondent's scores, so that each respondent has similar across-item range of scores. With logit-based estimation (for example, HB uses logit at the lower level of the hierarchy), a respondent with low response error has much larger magnitude (and range) of scores than a respondent with more response error. Typically something is done to normalize the scale factor across respondents before using the MaxDiff scores in a next-stage model, such as regression.

Answered: Stat testing weighted share of preference

Wed, 02 May 2012 00:48:50 +0000

We have simply counted the number of respondents for n rather than compute "effective base."

Answered: How can I randomize 2 sets of question by keeping the order of questions within each set?

Tue, 01 May 2012 18:13:29 +0000

SSI Web version 7.0.0 and later can do this. It is a feature called "Randomized Blocks."

Answered: Mark up capability

Tue, 01 May 2012 16:24:36 +0000

I think we did something like this for someone once. It was one of those questions where you could click to highlight a phrase as good or bad (or turn off the highlighting). What was clicked was then stored to a hidden variable as part of the respondent data record. I would imagine something similar could be done for images in a heat-map type of thing.

I personally am not skilled enough in JavaScript to do this, but there are a few people here who probably could, depending on your time frame and schedules and such. They might be more inclined to take it on as a small consulting project in order to test the waters for adding it as a question type to SSI Web (no promises).

If you would like to pursue that avenue, feel free to send an e-mail into support at sawtoothsoftware dot com with what exactly you need and when you'd need it and we can route it to the right people.

If anyone else wants to add a comment, are these type of heat map or highlighting questions pretty common requests?

Answered: Please help: Error messages

Tue, 01 May 2012 12:03:45 +0000

If you wanted to investigate it yourself, you could download the data and look at what browser someone was using who had problems and their respondent number. Install the same browser on your computer, log into your admin module and go into test mode. Test mode will allow you to take the survey with their respondent number to recreate their experience of taking the survey.

If the error message is "Please answer question 1" it sounds like perhaps you have either custom error messages or custom JavaScript validation. Hopefully this would make it easier to figure out the problem, as the default missing answer error message is "A response is required."

You can also send a copy of your survey in to support at sawtoothsoftware dot com and any additional information (such as the data file, a screenshot of the error message, etc.).

Question about how the information inside each .dat file generated by SSI Web is dsiplayed

Mon, 30 Apr 2012 12:01:06 +0000

Hi all,

I searched the documentation and I found nothing, so I guess I already know the answer, I just want a confirmation of that :] Does someone knows if it is possible to change the layout of the information inside each .dat file generated by SSI Web (i. e., how it is organized the information)?

I'll give a pratical example:

Currently, the generated .dat file goes something like this:

(...)
Items      2       5
(...)

The thing is, these are the IDs of the items in a general list of items, but in fact it contains different levels, containing different factors. So, for instance, item 1 and 2 are the level 1 and level 2 respectively of the same factor (let's call it factor "A", for simplicity's sake), and item 3, 4, 5 are 3 different levels of factor "B".

So the ideal thing would be something like this (applied to the previous example):

(...)
Factor     A      B
Items      2      3
(...)

I don't know if I'm being sufficiently explicit. Of course I can manipulate the lists, separating each set of different levels for the same factor in their own list, but that will be too complex, and if there is a more simple method it would be much easier for me.

I guess that another option would be doing an excel / libreoffice macro to rearrange the file as I wish, although I don't have experience with that.

Anybody knows if there is a third, more simple approach? :]

Answered: See if a constructed list contains an item in free format question

Mon, 30 Apr 2012 10:42:42 +0000

So, thank you all for your kind answers.

To summarize, in my case I have 2 predefined lists: one containing the items and another containing the possible combinations. My list of items is something like:

Item 1
Item 2
Item 3
Item 4

My combinations list has the combinations of items possible. It goes like this:

So, I created a general "RandomQuestions" constructed list. If I want to display 4 questions in my questionnaire, it would be like this:

Begin Unverified Perl
    my $number_questions = 4;

    ADD("Combinations");
    RANDOMIZE();
    LISTMAX($number_questions);
End Unverified

Then, for each question I created another constructed list. For instance, for question 1, it is:

Begin Unverified Perl
    my $tmpString = "";
    my @tmpItems = ();

    $tmpString = LISTLABEL("RandomQuestions", 1); # in the 2nd question it would be LISTLABEL("RandomQuestions", 2), etc
    @tmpItems = split(" ", $tmpString);
    ADD("listItems",$tmpItems[0]);
    ADD("listItems",$tmpItems[1]);
    RANDOMIZE();
End Unverified

Just a final question: although I think this is correctly working, does anybody know a way of debugging this, other than running the questionnaire several times and empirically see if the combinations are being repeated?

Thank you all again!

Answered: Best simulation method for an Alternative Specific Design CBC

Sat, 28 Apr 2012 02:50:32 +0000

SMRT's RFC routine indeed cannot handle alternative-specific designs. We figured out how to make it work in our newest Online Market Simulator.

That said, if you have individual-level utilities (via HB, for example), the Share of Preference (logit) simulation method should work fine, and I'd recommend you use it if you need to stick with SMRT for this project.

The tuning of the scale factor indeed works very similarly as with RFC. Adjust the Exponent to make the shares steeper, relatively, or flatter.

Complete ignore ommited levels of constructed lists in ACBC part worth estimation?

Fri, 27 Apr 2012 13:29:38 +0000

I´m fokussing an issue relating to a recent study. We conducted an ACBC task, the sample is splitted in "active customers" and "unactive customers" without an annual fee-contract.

This information was used to create a different price level, the 1st group got something like "+ 10$ per month on top" and the 2nd group "110$ per month" - with 100 = annual fee and 10/XX/YY $/month the acual levels of price increase.

We set up a parent list with 5 "active customers" levels and 5 "unactive" levels - brought into ACBC with constructed list.

As you can see, there is no need of estimating hypothetic part worths for items which are comepletely irrelevant for the segment. But in HB estimation the price attribute is handled as 10-level attribute, with estimating part worths for all levels.

No I´m considering:
1. Is it possible to seperate the ommited levels before starting the utility runs - rescaling the price level to the 5-level attribute as it was intended? Even with "unavailable" selection in HB settings i´m getting some hypothetic part-worths, making every remaining price point "look" completely different as any other attribute´s zero centered diffs.

2. If 1. not: Is there a way to rescale the part-worths back to the "normal" look an feel of zero centered diffs?

3. If both is not possible: I´d do the workaround to not present price part-worths (the usefullness is questionable anyway) and only report price as sensitivity curve out of SMRT, asuming that my relative Information in part-worths of chosen levels aren´t polluted due to this issue?

thanks in advance,

Sven

Answered: Anyone know about the "evolutionary algorithm" Affinova is talking about?

Fri, 27 Apr 2012 04:11:41 +0000

Fortunately the Affinova guys actually list the patent numbers, with links, for the methods they have developed, so in theory you could find out exactly what it is they do. [So many companies seem to claim to have patented something, then try to keep it a big secret, when the whole idea of patents is to allow others to find out the detail, so that you don't run the risk of infringing.]

Have a look at the Excel add-in called Genehunter. it's a rather cool piece of software that is fun to play with, and also rather useful (e.g. I use it for TUR analysis).

http://www.wardsystems.com/genehunter.asp

Answered: Can I require a screen to display for some minimum amount of time?

Tue, 24 Apr 2012 21:06:41 +0000

http://sawtoothsoftware.com/forum/135/delay-navigation-button

Answered: How to test if parth worth utilities are significant?

Tue, 24 Apr 2012 15:42:42 +0000

This might seem like an easy question, but it's actually quite involved.

First, what do you mean by "significant"? That word packs a lot of potential meaning, and you need to be more specific. For example, are you wondering about what the confidence intervals are around each of the utility estimates, given a 95% confidence?

Sometimes, researchers want to know if there are significant differences between Shares of Preference for different products, as simulated throught the What-If simulator.

An article is available on our website regarding statistical tests for conjoint analysis results:

http://www.sawtoothsoftware.com/education/ss/ss20.shtml

The article is entitled: Testing for Significant Differences in Conjoint Analysis

Answered: What does this code mean in my script?

Tue, 24 Apr 2012 14:41:12 +0000

It looks like script from a Ci3 program.

Q: shows the question name

T: indicates what line on the screen to begin the text on

next is the question text followed by the answer choice text

I: is followed by the actual question logic

LOC 7 8 1 Natural indicates to locate the answer choices starting on line 7 with 8 answer choices and 1 line per choice

SEL 8 1 7 1 ON OFF indicates use a 'Select' type question with 8 total choices with a minimum of 1, a maximum of 7 and 1 exit code

The question logic indicates 8 answer choices, but the text only shows 6 so there is a logic conflict here.

Answered: How to have user create a hierarchy or order in answers

Tue, 24 Apr 2012 13:12:58 +0000

One way to do this would be to implement a drag and drop ranking question. Respondents would see 5 items on the screen and would move them from top to bottom based the rank order.

Answered: Share of preference sig testing

Mon, 23 Apr 2012 15:36:59 +0000

Your equation #1 will work if you are using a First-Choice rule simulator. (Each respondent fully gives his vote to just one product).

But, if you are using a method that "splits votes" such as Share of Preference (Logit Rule) or RFC, then you'll need to estimate the standard deviation of the shares across respondents, and divide by sqrt(n).

Answered: Are Choice Based Holdouts ordinal scaled?

Mon, 23 Apr 2012 15:34:53 +0000

A holdout choice task for a CBC-looking question involves showing respondents multiple profiles and asking them to pick their favorite.

Thus, it's the same thing as saying: if we show respondents items A, B, and C, which one is their favorite?

So, within each respondent, all we gather is a scaling where the chosen item is 1 and the not chosen items are 0. So, if "A" is chosen for a given respondent, we know order: A>B, and A>C.

And, we know Ratio-scaling within the respondent: (A was chosen 100% for this respondent, and the other two 0%).

And, if we tabulate the results across a sample of respondents, we'll get probabilities of choice for the same, for choice likelihood of A, B, and C. Those values have Ratio scaling.

Answered: Using jQuery inside SSI Web

Mon, 23 Apr 2012 11:27:23 +0000

Try the following steps:

1) Download the whole jQuery UI package from http://jqueryui.com/download
2) Unzip the package and copy the JS and CSS folder to your survey inside the folder /graphics/system/ .
3) Open the Survey Settings of your survey and go to 'Headers and Footers'.
4) Inside the 'HTML <head> Tag' box, copy and paste the following lines:

<link rel="stylesheet" type="text/css" href="[%GraphicsPath()%]system/css/ui-lightness/jquery-ui-1.8.19.custom.css">
<script type="text/javascript" src="[%GraphicsPath()%]system/js/jquery-1.7.1.min.js"></script>
<script type="text/javascript" src="[%GraphicsPath()%]system/js/jquery-ui-1.8.18.custom.min.js"></script>

<script type="text/javascript">
var j = jQuery.noConflict();
</script>

5) Click on OK and create a new Free Format question
6) Go the 'Question Text' tab and add the following lines to the 'Header 1' field:

<h3>jQuery Slider Example</h3>

<div id="slider"></div>

<script type="text/javascript">
j(document).ready(function(){
  j("#slider").slider();
});
</script>

7) Test and see if it works ;)

At this point, I think you can manage.

Storing hidden variables in a .xls file

Mon, 23 Apr 2012 09:55:20 +0000

Good morning all,

I have a survey where I'll use hidden variables under some free format questions. It is possible to automatically store the respondents answers in this free format questions into a .xls file? And is it possible to edit how the information will be displayed there?

Thanks

Market size estimation

Fri, 20 Apr 2012 10:46:07 +0000

Hello,

any one have sharable information on how to assess category growth/shrinkage via a conjoint study? ( i do know that SoP adds to 100%) :)

Cheers,

Answered: Custom verification freezing in IE

Thu, 19 Apr 2012 20:20:46 +0000

If you are in SSI Web 8 then this is an unfortunate bug that will be fixed in release 8.0.2. The bug causes SSI_GetValue() to not work in Internet Explorer. However, you can find a workaround in this post here: http://sawtoothsoftware.com/forum/979/problem-with-js-validation-with-ie9#a1073

Answered: RLH / Pct cert for constant sum conjoint

Thu, 19 Apr 2012 19:14:14 +0000

Indeed, as you've probably noticed, the RLH computation is different when chips are involved, and much lower than standard discrete choice.

The RLH norms will depend on how much respondents tend to push the chips all onto one concept versus dividing them more evenly.

Maybe some users out there who have constant sum datasets can give you some norms from their observations. I don't deal with chip allocation much at all, as my projects are all discrete choice.

Answered: "Hide" a checkbox in a grid question - Part 2

Wed, 18 Apr 2012 15:34:59 +0000

if Q1 is a column based grid, its items will be named
Q1_c1_r1 instead of Q1_r1_c1 (notice the r vs c order)

if this is not the reason, then try

if([%Q1_r1_c1%] <> "1")

instead of

if([%Q1_r1_c1%] <> 1)

and let us know whether or not it worked

Answered: Can i do a latent class segmentation for the below situation?

Wed, 18 Apr 2012 06:29:12 +0000

If you are thinking of a regression, or similar, to predict level of higher education reached, then yes you can, unless I am misunderstanding what you are asking [which is always possible :-) ]

But what tends to happen (in my experience, anyway, using the Q software - whether this also happens with Latent Gold I have no idea) is that the segments found using Latent Class sometimes differ more on the constant term in the regression, and less on the coefficients.

Answered: can i do a latent class segmentation in a study of student background factors who enter to the higher education

Tue, 17 Apr 2012 22:19:43 +0000

Or use the Q software . . its Latent Class capability is very good.
Www.q-researchsoftware.com

Scott

Answered: Recreating on-the-fly computed rank order of MaxDiff-items

Tue, 17 Apr 2012 15:16:50 +0000

Constructed lists data is always saved and can be exported. Constructed lists only save the final results of the list not each intermediary step. If you want the rank order of all the items but did not save them in the list then this requires you sending the data to us here at Sawtooth Software and doing some custom work for you. I have done this a couple of times for other people so its not difficult to do. Please contact me directly if you wish to have this done.

Email Validation

Mon, 16 Apr 2012 21:09:07 +0000

I had read some posts in the old forum regarding using the custom javascript verification area to check that an answer appears to be a valid email address and I've attempted to try some solutions but have not been successful. Does anybody have any script they currently use that works for them?

Answered: How to show code at the bottom of each concept on the task card

Mon, 16 Apr 2012 16:56:45 +0000

open your first screen in browser, right click and select view source, find the lines like

 <input type="text" name="CBC_Random1_1" size="3" id="CBC_Random1_1">

and replace with

the 1 here refers to the last digit after CBC_Random1_XX
when you replace all these in all the screens and save and open in browser, you'll end up with your requested look (I hope :))

Answered: Another question still regarding sliders

Mon, 16 Apr 2012 15:34:42 +0000

I do got to put a slider showing its current value, in HTML5 it is very, very easy. The only problems are the fact that I cannot define its width (it's very small), and I don't know what kind of browser compatibility issues may appear, since this is a kind of new feature. If anyone knows how to improve any of the two approaches, I would appreciate a lot to know it.

Code:

<input type="range" min="0" max="100" value="30" step="5" onchange="showValue(this.value)"/>

<span id="range">0</span>

<script type="text/javascript">

function showValue(newValue)
{
    document.getElementById("range").innerHTML=newValue;
}

</script>

Answered: Please describe the format of .labels file for import into online simulator.

Mon, 16 Apr 2012 15:26:37 +0000

If you are exporting your variables from SSI Web then the .labels file is created for you automatically.

If you are comfortable with XML, then you could also create / modify the labels file manually. Email me directly (dave@sawtoothsoftware.com) and I can send you a document that describes the file.

Answered: "Hide" a checkbox in a grid question

Thu, 12 Apr 2012 21:35:43 +0000

I've just tested both of these solutions and they both work.

For Jay's solution, I did need to refer to the graphical element of the checkbox:

#QB_r1_c1_graphical

Thank you both!

Export Latentclass membership per respondent into SPSS

Thu, 12 Apr 2012 15:14:42 +0000

Hi there,

I would like to describe the segments that have been created by Sawtooth latent class by some other variables (demographics, attitude questions etc.). As fas as I know, it is not possible with Sawtooth to describe the segments by some other variables. So my supervisor's solution was to export class membership per respondent to SPPS and then describe each class of respondents by the demographics etc. that are in SPSS. But how do I export data (including class membership per respondent) from Sawtooth Latent Class to SPSS?