Conditional function not extracting the trail data as intended

[FayDu716]

Hi all,

I really need help here! Because I have been stuck with this problem for a long time. I really need someone to help me figure where the problem is.
I am designing a dynamic test which will provide error specific feedback to the participant. As the code that I provided below, I created a trail called 'Item4' . And in this trial, if the participant provide an incorrect answer no matter what response it is in the first attempt, feedback "Incorrect, try again" will be provided, and then the participant have to try Item 4 again. In the second attempt, if the incorrect answer is "A" and "F", "Item4_feedback2a" will be provided. This feedback is included in a condition node called "Item4_feedback2a_node". There are other feedback nodes (like the incorrect response "B" and "C" will lead to "Item4_feedback2b_node; and "D" will lead to "Item4_feedback2ab_node) which are written in the same format.
My problem is that this conditional node is not extracting the data from last trial (the second attempt's response) but the last (2) trail (the first attempt's response).
Could someone help me find out why this is the case.Please, I will really appreciate any help! Thank you very much!

Here is the code:

 //define Item 4
     var Item4 = {
      type: jsPsychSurveyHtmlForm,
      preamble: `
       <img src="img/Item4.png" alt="Item4" style="float:center;width:1100px;"></img>
       `,
      html: '<span style = "font-size: 20px">I choose  <input name = "Item4_response" id = "test-resp-box" size:"50" oninput="this.value 
= this.value.toUpperCase()" />.  </span>',
      autofocus:'test-resp-box',
      button_label: ["Submit"],
      data: function(){
        return{
        Item_name: 'Training_Item4',
        correct_response: 'E',
        attempt:jsPsych.data.get().filter({ Item_name: 'Training_Item4' }).count() + 1
        };
      },
     on_finish: function(data){
      var responseKey = "Item4_response" + data.attempt;
      data.response[responseKey] = data.response.Item4_response;
      delete data.response.Item4_response;
      if (jsPsych.pluginAPI.compareKeys(data.response[responseKey], "E")) {
        data.correct = true;
      } else {
        data.correct = false;
      }
      console.log (data.attempt,data.response);
     }
     };

     //define Item 4 feedback 1
     var Item4_feedback1= {
      type: jsPsychHtmlButtonResponse,
      stimulus: function (){
       var last_trial_correct = jsPsych.data.get().last(1).values()[0].correct;
       if (last_trial_correct){
        return `
        <img src="img/Item4.png" alt="Item4" style="float:center;width:1100px;"></img>
        <p style="color:green;">Correct！</p>
        `;
        } else {
          return `
        <img src="img/Item4.png" alt="Item4" style="float:center;width:1100px;"></img>
        <p style="color:red;">Incorrect, please try again！</p>
        `;
        }
       },
      choices: ['OK'],
      };

     //define Item 4 feedback 2
      var Item4_feedback2a = {
      type: jsPsychHtmlButtonResponse,
      stimulus: function(){
       var last_trial_correct = jsPsych.data.get().last(1).values()[0].correct;
       if (last_trial_correct){
        return `
        <img src="img/Item4.png" alt="Item4" style="float:center;width:1000px;"></img>
        <p style="color:green;">Correct！</p>
        `;
        } else {
          return `
          <img src="img/Item4.png" alt="Item4" style="float:center;width:1000px;"></img>
          <p style = "color:red;">Incorrect，(some feedback)  Please try again.</p>
          `
          ;
        }
       },
      choices: ['OK'],
      };

      var Item4_feedback2a_node = {
          timeline: [Item4, Item4_feedback2a],
          conditional_function: function (){
            var data = jsPsych.data.get().filter({Item_name: 'Training_Item4'}).last(1).values()[0];
            var responseKey = "Item4_response" + data.attempt;
            if (data.attempt === 1) {
             if(jsPsych.pluginAPI.compareKeys(data.response[responseKey], "A") ||
                jsPsych.pluginAPI.compareKeys(data.response[responseKey], "F")){
             return true;
               } else {
             return false;
               }
            } else {
           return false;
            }
          }
      };

// some other feedback and feedback node

//define Item 4 node. This condition node occurs only if the test taker got Item 1 wrong in the previous trial.
      var Item4_node = {
          timeline: [Item4,Item4_feedback1,Item4_feedback2a_node, Item4_feedback2b_node, 
             Item4_feedback2ab_node,Item4_feedback3a_node, Item4_feedback3b_node, 
             Item4_feedback3ab_node,Item4_feedback4_node],
          conditional_function: function (){
            var data = jsPsych.data.get().filter({Item_name: 'Test_Item1'}).last(1).values()[0];
            var responseKey = "Item1_response" + data.attempt;
               if(jsPsych.pluginAPI.compareKeys(data.response[responseKey], "C")){
                   return false;
               } else {
                   return true;
               }
           }
       };

[ChristopheBossens]

Are you referring to the conditional function in Item4_feedback2a_node ? I think this will never reference to the last trial because it is executed only once before the start of the timeline, and it is only in that timeline that you specify item4 a second time.

Because I could not run your code without making to many changes, I tried to implement a proof of concept of what I think you need:

const jsPsych = initJsPsych();

    var item4 = {
      type: jsPsychHtmlButtonResponse,
      stimulus: "<p>Item4</p>",
      choices: ["Correct response", "Incorrect response A" ,"Incorrect response B", "Incorrect response C"],
    }

    // Provide general feedback the first time
    var item4_feedback1 = {
      type: jsPsychHtmlButtonResponse,
      stimulus: function(){
        var last_trial_correct = jsPsych.data.get().last(1).values()[0]['response'];
        if (last_trial_correct === 0){
          return "correct";
        } else {
          return "Incorrect";
        }
      },
      choices: ["Continue"],
    }

    // Provide response specific feedback the second time
    var item4_feedback2 = {
      type: jsPsychHtmlButtonResponse,
      stimulus: function(){
        var last_trial_correct = jsPsych.data.get().last(1).values()[0]['response'];
        if (last_trial_correct == 0){
          return "You responded correct";
        } else if (last_trial_correct == 1){
          return "You responded A";
        } else if (last_trial_correct == 2){
          return "You responded B";
        } else if (last_trial_correct == 3){
          return "You responded C";
        }
      },
      choices: ["Continue"],
    }

    // Look at the response to the first item4, if it was correct this timeline does not
    // need to be executed.
    var item4_second_attempt = {
      timeline : [item4, item4_feedback2],
      conditional_function : function(){
        var last_trial_correct = jsPsych.data.get().last(2).values()[0]['response'];
        if (last_trial_correct == 0){
          return false;
        } else {
          return true;
        }
      }
    }

    // Decide if we should run item 4
    var item4_node = {
        timeline : [item4, item4_feedback1, item4_second_attempt],
        conditional_function: function(){
          return true;
        }
    }

    jsPsych.run([item4_node]);

You can use buttons to simulate the responses. The important parts are item4_second_attempt , where it is decided if it should run item 4 again (only if the response to the first presentation of item 4 was correct I think?). The second important part is item4_feedback2 , where you set the stimulus specifically according to the response to the second presentation of item 4.

[FayDu716]

Thanks @ChristopheBossens for your generous reply. Your code make perfect sense by condensing the feedback in different "if" conditions. I think I should try it on to see whether I can condense my feedback too. Unfortunately, I can't use button response, because at my later test complexity level, the task will no longer be multiple choice, but rather sorting the different options. For the consistency throughout the test presentation, I set it to survey text to facilitate better data extraction later.
I think another problem I ran into was that I don't know how to capture the data from the survey text more accurately. For instance, if in the next complexity level, the correct response is "AF", if the answer is "AB", it would lead to "Feedback_a" and if the answer is "AC", it would lead to "Feedback_b", and so on. I tried to mark the data by adding "attempt count", like "Item4_response1" or "Item4_response2" to make the feedback specific to these responses. But the setting in my '"if" condition can't find these corresponding data. Do you have any good advice towards it? Many thanks again!!!

[FayDu716]

Update my reply. When I dense my feedback, it works perfectly fine. Thank you very much! @ChristopheBossens