You could use Array methods:

sort: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/sort

slice: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/slice

map: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/map

Here is snippet i tried for your problem:

function sortQuestions(results) {
  const questionsWithRatio = results.map((result) => ({
    ...result,
    ratio: result.likes[0] / Math.max(result.dislikes[0], 1),
  }));
  // console.log(questionsWithRatio);
  const top15 = questionsWithRatio
    .sort((a, b) => b.ratio - a.ratio)
    .slice(0, 15);
  console.log(top15);
}

The answer above is a superb pure Javascript solution. However, it might be the case that we do not want to drag 1000s or possibly more of items out of the database just to sort and pick 15 on the client side. Given a set of docs like this:

var r =   [
    {likes: ['A','A'], dislikes: ['A','A']},
    {likes: ['A','A']},
    {dislikes: ['A','A']},
    {likes: ['A','A','A','A','A'], dislikes: ['A','A']},
    {likes: ['A','A','A','A','A','A','A','A','A'], dislikes: ['A','A']},
    {likes: ['A','A','A','A','A','A','A','A'], dislikes: ['A','A']},
    {likes: ['A','A','A','A','A','A','A'], dislikes: ['A','A']},
    {likes: ['A'], dislikes: ['A','A']}
];

Then here is an aggregation that will drive the sort and "slice" logic into the DB:

db.foo.aggregate([
    // Protect against missing fields with $ifNull                                                         
    // Protect against division by zero dislikes using max[1,size]                                         
    {$addFields: {ratio: {$divide:[{$size:{$ifNull:["$likes",[]]}},
                                 {$max:[1,{$size:{$ifNull:["$dislikes",[]]}}]}
                                ]}
               }},
    {$sort: {ratio:-1}},
    {$limit: 3}  // change to 15 for the real dataset
]);

If we must accomodate the "run length encoding" nature of the arrays e.g. [2, 'user1', 'user2'] then the agg is very similar only using $arrayElemAt to grab array[0] instead of using $size:

var r =   [
    {likes: [2, 'A','A'], dislikes: [2, 'A','A']},
    {likes: [2, 'A','A']},
    {dislikes: [2, 'A','A']},
    {likes: [5, 'A','A','A','A','A'], dislikes: [2, 'A','A']},
    {likes: [9, 'A','A','A','A','A','A','A','A','A'], dislikes: [2, 'A','A']},
    {likes: [8, 'A','A','A','A','A','A','A','A'], dislikes: [2, 'A','A']},
    {likes: [7, 'A','A','A','A','A','A','A'], dislikes: [2, 'A','A']},
    {likes: [1, 'A'], dislikes: [2, 'A','A']}
];

db.foo.aggregate([
    // Protect against missing fields with $ifNull but this time it is not
    // an empty array [] but rather an array of one elem, the length, which
    // is 0:                                                        
    {$project: {ratio: {$divide:[
        {$arrayElemAt:[{$ifNull:["$likes",[0]]}, 0]},
        {$max:[1, {$arrayElemAt:[{$ifNull:["$dislikes",[0]]}, 0]}]}
        ]}
    }},
    {$sort: {ratio:-1}},
    {$limit: 3}
]);

{ "_id" : 4, "ratio" : 4.5 }
{ "_id" : 5, "ratio" : 4 }
{ "_id" : 6, "ratio" : 3.5 }

Before I try to help, please try to improve your code's readability. You can look up async/await and ES6 syntax to help you in this.

Also, if you're using an IDE like Visual Studio Code then you can install prettier which will autoformat your codes.

Here's a rough solution for you:

router.get('/questions/best', function(req,res,next){
  Question.find({}).exec(function(err,results){
    if (err) return next(err);

    let top15 = [];

    // note that the code is slightly more readable already
    results.forEach(result => {
      const dislikesToUse = result.dislikes[0]

      let ratio = 0;

      if(dislikesToUse > 0) {
        ratio = result.likes[0]/dislikesToUse
      } else {
        ratio = result.likes[0]/1
      }

      // push each result with ratio to top15
      top15.push({...result, ratio})
    });

    // sort top15 by ratio
    top15.sort((a,b) => b.ratio - a.ratio) // do a.ratio - b.ratio to change order

    // get only the top 15
    top15 = top15.slice(0, 15)

    res.render('content', { whichOne: 5, user: req.user, questions: top15 });
  })
});

NOTE

This is only one way to do it and might not be the most efficient way. You should also consider Mongo's aggregate functions.

You should also definitely reconsider your schema design.