Friday, 1 August 2014

The problem with using Search as a learning tool



The world of learning has changed significantly with the introduction of the Internet and search engines. Many children in their first years of education in the UK are aware of what search is and the basics of how to use it. The next major innovation that will change the classroom environment will be the introduction of tablets. Still in its infancy in being introduced effectively at school, it’s easy to see a future where each child uses their tablet just as an adult worker would use their laptop. Search should then replace the need for supporting paper-based knowledge right?

Well children are already doing this and it is a new age life skill that is (or should be) taught in the classroom. These skills are mostly taught like other research skills including understanding fact Vs fiction, the evaluation of evidence, quality, relevance and suitability. Although these skills are vitally important for a digital savvy future my argument is that we are teaching skills to children based on the limitations these platforms supply.

So what are the problems with search?

It’s easier to answer this question by re-emphasising the purpose of search. Most people would define successful Search as:

‘The returning of a list of results as quick as possible ordered by relevancy from inputting a string of text’.

This definition is limited by the preconception of what search engines such as Google actually do. If we strip the definition back to the core purpose of search we get something simpler like:

‘Returning the most relevant content as quick as possible’.

It may be that a user searching for a term gets the results they asked for in milliseconds but that isn’t the breadth of the whole user-experience. Most of the time we search for a term and then work down the list of results, opening each page, going back to the results, trying again, maybe refining our search again until we finally find something that is relevant (in our own opinions). This human behavioral element to finding what we want is also part of the search. There must be some holes, as the perfect search would give you what you wanted first time around.

Now before I proceed please don’t get me wrong, Search is incredibly useful and has its purpose in life but in the capacity of using it as a learning tool it could be better, couldn’t it? I’ll go into detail about some of the problems that exist and then go into our solution.

Problem 1 - Search Input
The input for all search engines is a string of text. In our lives there are things we know and things we know that we don’t know. In most cases we type into search things that we know that we don’t know; headlines we’ve heard in the news we want more detail on, industry buzz words that we want to understand, a film that we’ve heard is good and want to watch the trailer. The trouble is it’s the things we don’t know that we don’t know that forms the majority of the world we live in. The Internet is full of this type of content but how do you get to it?

When learning a subject from scratch the entire subject is full of things we don’t know that we don’t know. It’s only as we progress from the introduction we find things to develop our knowledge on. Within Search the typical user experience can be expressed as something I call sidestepping. It’s when a user carries out a search, uses the results to find things they didn’t know they didn’t know and then refines the search and repeats until they have learnt the things they didn’t know existed from the start of the process. This sidestepping takes time and is limited by the users willingness to step out of their comfort zone.

Problem 2 – Non-linear
Learning from search is mostly difficult because there is an open way to navigate the search results. We all know that teaching a subject matter needs to be done in some sort of linear order. We need to learn English language before we can learn Shakespeare and we need to understand basic chemistry before we can understand photosynthesis. In its most simple form we need to read chapter 1 before we can make sense of chapter 3.

Search does not help us navigate in this linear fashion and is dependent on us selecting the appropriate steps in the right order or finding content that does it for us. It cannot guide a student but can offer support when guidance is given.

Problem 2 - Relevancy of content
This is probably the part search does best but there are still some problems. Content relevancy is retrieved based on a complex set of algorithms, which is limited by a formula. This formula is different for different search engines, constantly evolves and is as much dependent on the optimization of the page as its real relevancy.

Problem 3 – Quality
The quality of search content is variable. The search engines have limited capacity to filter out bad quality content, no way of telling the difference between fact and fiction and no way of assessing the credibility of the contents author.

Problem 4 - Suitability
Suitability refers to the matching of content with the users profile. This could be for example the age of the user, something very important in primary and secondary education as content written for an 11 year old might be very different for a 14 year old.

Despite the big-brother controversy around collection of user data the search engines know little personal details about the user. Social networks on the other hand do hold a considerable amount of data about you and if the search engines linked to the API’s of social networks for example Google plus to Google there is huge potential to increase the content suitability.

The problem though, will still remain until the author has a way of tagging their content with the attributes of its target audience.

Problem 5 - Is it online and accessible?
Some of the best content that exists in the learning arena has been commissioned by private companies such as BBC learning, wishing to license or sell its content. Its obvious that content in these cases is hidden from the Internet and/or from search.

All of these problems with search contribute to the reasons why content is still digested in the classroom in a linear fashion and even why we still use textbooks. Sure you can digitize the textbooks but what’s the point? Wouldn’t it be intelligent to make use of the wealth of knowledge that already exists?

So consider this as a solution. What if we give access to the plethora of content on the Internet but solve the aforementioned problems by bolting on some software as an interface.

The search results are taken from combining Yahoo, Bing and Google results. We then remove duplicate content.

We start by solving the problems of Search input and its open navigation approach by creating what we call a narrative. Each subject mater will be packaged with a different narrative and can be as specific or as broad as we like. For example we could create a specific narrative of ‘blue whales’ which was a sub-narrative of ‘natural history’. The narrative would put content into a logical order under chapters much like a books contents page.

Each chapter of the narrative has pre-determined keywords assigned to it that act as the search term thus a student would never have to think about what string of text they should be searching for. Each chapter can also be cross-referenced to specific criteria for an educational curriculum. This solves order and input but what about the relevance, quality, and suitability?

There is only one way to do this and that is to use human expertise. An expert in the subject matter vets each piece of content on the learning platform. It goes through an approval process that checks out the quality of the content and the credibility of the author. If approved, it is tagged with extra meta-data that indicates which chapter in the narrative it should appear, a tiny overview of the content, which parts of the curriculum it satisfies, suitability of age-range and how long it will take to digest.

There’s also one other thing to add to the content…. premium content that does not exist on search! We make arrangements with the content owners and then add our enhanced meta-data to this content in the same way we do with the existing search content.

This software could also includes tools for the user such as note taking, testing and evaluation as well as tools for teachers such as student activity, plagiarism checking and the creation of custom narratives.

If anyone wants to make this with me let me know 😊

1 comment: