Using a Decision Tree

I spent about half an hour writing a response to a thread that was removed while I was working on it, so I thought I would post a version of it here, in case anyone finds it useful. The student was struggling to decide among four very different schools.

For college-bound students struggling to choose a school, particular if they are quantitative-style thinkers, using a decision tree can really help.

Decision trees allow a person to translate his/her preferences into data, and to evaluate various options in light of this.

Start by listing the things you care about and then rating (10=high, 1=low) how important each thing is to you. This is a rating, not a ranking.

Random example:

Size under 2,000 students (importance = 4)
Greek system yes (importance = 2)
Sunny weather (importance = 8)
Low tuition (not including room/board) (importance = 7)
Short distance to airport (importance = 2)
etc.

Then, do some research and note how much or how little each school aligns to each of your preferences (10=high, 1=low.) Examples:

UC Santa Cruz
Size (17,000+) = 1 x 4 = 4
Greek system (yes) = 10 x 2 = 20
Sunny weather (262 days per year) = 8 x 8 = 64
Tuition ($13,962 in-state …or… $28,992 out-of-state) = 9 x 7 = 56 …or… 5 x 7 = 35
Distance to airport (35 miles) = 5 x 2 = 10
etc.
From this example: Total Score for UC Santa Cruz = 154 (in-state) or 133 (out-of state.)

Denison University
Size (~2,300) = 9 x 4 = 36
Greek system (yes) = 10 x 2 = 20
Sunny weather (175 days per year) = 5 x 8 = 40
Tuition ($52,620) = 3 x 7 = 21
Distance to airport (30 miles) = 5 x 2 = 10
etc.
From this example: Total Score for Denison University = 127.

You can then use the decision tree to help distinguish the schools from one another. You may look at the final scores and be pleased. Or, you might say to yourself, “Actually, the weather isn’t THAT important to me,” change its value, and recalculate. You might narrow your choice to two schools and visit both, having confidently eliminated the others. Or, you may see a numerical result but realize that you feel disappointed in it, with your gut pulling you toward another school. That’s a also GREAT end result of the decision tree, because you will have gotten in touch with your own instincts. Follow your gut if it tells you something!

Hopefully, the author of the original (now deleted) post will see this and find it helpful. Decision trees are a lot of work, but do a great job of providing quantitative data in the middle of a mushy, subjective process… and/or stimulating users’ hidden emotions when they see numerical results. Either way, this process can help a confused person to be more confident in his or her final choice.

We scored my kid’s schools. But in the end she chose between the top few with highest scores based on how her accepted student visits went. Some things don’t show up in a scoring model.

This is, by the way, more of a scoring model than an actual decision tree. But it can be a useful tool as part of the decision process.

“In the end she chose between the top few with highest scores based on how her accepted student visits went.” What a perfect use of this technique! Congratulations on your daughter’s success.

Agree, I would not eliminate gut feel as part of the decision process.

An additive point system with limited ranges may not capture the decision well, since preferences may be nonlinear. You may have to allow values of -∞ (negative infinity) if the college characteristic eliminates it (e.g. if it is unaffordable when looking at cost).

When it comes to the purpose of college education, the most important factors for most students are affordability and offerings of suitable academic programs of interest (including their availability to the student, considering whether possible majors are competitive admission).

I really like the idea of including a negative point range, including negative infinity, @ucbalumnus. Thanks for that addition!

The author of the post to which I’d originally applied had already been admitted to the major of his/her choice at four different schools, so in his/her particular case, “availability of academic programs of interest” wasn’t relevant. However, it could easily be included for other students. Similarly, s/he expressed no interest in cost.

The most important factors will vary by student, naturally, as will the extent of their importance. I hope this system is helpful to some, especially the original student - a quantitative thinker who was very confused and overwhelmed trying to subjectively evaluate the options.

As I mentioned above, using a technique like this can be especially helpful in narrowing choices or giving rise to a “gut feeling” that a student might not have preciously felt or recognized.

My son was interested in whichever college was going to ultimately earn him the most money at the fairest price. After narrowing his final choices based on his preferred type of college and successful visits, we started comparing on practical matters. We used Payscale and focused on the figures that showed the best ROI, overall COA, and ranking in College Salary Reports. We also looked at first year retention rates and average early career and mid-career salaries. I typed it up in an organized way so he could easily compare.

He chose the least expensive option, a respected state school. Personally, I have reservations about his choice, as I am not sure it’s the best fit for him, but he was adamant. He will start college in August and is getting excited.

My daughter was all about academic rigor and fit. No charts for her. My husband and I were interested in career services, alumni networks, and reputation. She made her final decision based on her feelings. For her, it was the right way to do it. She’s thriving.