If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

## High school statistics

### Course: High school statistics>Unit 4

Lesson 1: Fitting trend lines to scatterplots

# Estimating the line of best fit exercise

The "line of best fit" is a line that shows the pattern of data points. If we can find a good line, it means there is a linear trend. If not, it means there is no linear trend. We can't ignore points that don't fit the trend. Created by Sal Khan.

## Want to join the conversation?

• I love Khan academy :)))!
• i am lost. when i do it on the practice, i am usually wrong by one or two points. is there a way to calculate this??halp!
• Here's a process you might try. Find the point that is the closest to one corner. Then, find the point that is closest to the opposite corner. Connect those two points.

Then, look at the line you draw and compare the rest of the points to it. If there are more points above the line than below it, then you might need to move the line up some. And if there are more points below the line than above, you might need to move the line down some.

Just experiment until you've got the line as close as possible to as many points as you can. It's OK if some points are far away, as long as your really close to most of the points.
• Khan Academy makes it a lot easier to understand some of the difficult math tasks.
• And it helps confuse you with the easier math tasks!
• Can you have a line of best fit when you have no correlation?
• If the correlation is exactly 0, then the best fit line using least squares regression is the horizontal line (slope 0) through y-bar (the mean y-value).
• so if i just do a line that is close to the other scatter points would it be right?
• A tip my teacher gave us was that a good line of best fit maintains a roughly equal amount of points on both sides.
• How do we make predictions in scatterplots
• Pretty sure it has something to do with the unit rate(maybe)
(1 vote)
• Honestly, I still don't get it. How would you know if it's accurate? And there's a method called Least Squares to find the equation, but how would you calculate if there are hundreds of points?
• If there are hundreds of points you not be using a line graph, you should be using a f/x table (frequency table)
(1 vote)
• What does best line fit mean