Analysis of words per sentence in differeent Newspapers.

* The Express has an inter-quartile range smaller than The Mirror but bigger than The Times. This means that the data is less in the middle and more spread out than The Mirror but more in the middle and less spread out than The Times. It has a median smaller than The Times but larger than The Mirror, which supports my initial prediction.

* The Mirror has the biggest inter-quartile range. This means that the data is more in the middle and less spread out compared to the other two papers. Also it has the smallest median, which supports my initial prediction.

Also all the cumulative frequency polygons have the typical pattern; a gentle slope at first (because there are a few very short sentences), then a steep slope (where most of the data is concentrated), and then another gentle slope (a few very long sentences).

BOX AND WHISKER PLOTS

My box and whisker plots show the spread of data (the upper quartile and the lower quartile) and the median. Both the whiskers represent 25% of the data and both the regions of the box represent 25%.

* The Times is slightly positively skewed

* The Express is very slightly negatively skewed

* The Mirror is positively skewed.

From my box and whisker plots (and my cumulative frequency polygons) I can see that my results match my hypothesis because each paper has a median in the correct increasing order:

In order of increasing median (and increasing difficulty in readability)

The Mirror The Express The Times

8.5 words per sentence 21.5 words per sentence 22.5 words per sentence

In my case, the median was the most appropriate average to use because my data was skewed, which meant that there were values for certain results which I wanted to ignore in my comparison, as they were likely to be outliers and distort the mean.

The median of the Times is 4 more than the median of the Mirror. This shows that a typical sentence in The Times is 4 words longer than a typical sentence in the Mirror. This supports my hypothesis. The lower quartile in The Times is also bigger than the median of The Mirror. This shows that 3/4 of the sentences in The Times are longer that 1/2 the sentences in The Mirror. This again, supports my hypothesis. The median of the Express is 3 more than the lower quartile of The Mirror. This shows that a typical sentence in The Express is 3 words longer than a typical sentence in The Mirror. This also supports my hypothesis. From this I can see that The Times and The Express are both more difficult to read than The Mirror. However, the median of The Times is only 1 more than the median of The Express. This shows that a typical sentence in The Times is only 1 word longer than a typical sentence in The Express. Even though this supports my hypothesis, this is a very surprisingly small difference and this suggests that the two papers are nearly equally difficult to read, which is not what I expected.

Although I have worked out the standard deviation, I used the inter-quartile range for the measure of spread of data, because my data was skewed, which meant that there are values for certain results which I wanted to ignore in my comparison, as they were likely to be outliers and distort the standard deviation.

In order of increasing inter-quartile range

The Times The Express The Mirror

7.5 8 9

Like the median, there is not much difference between the inter-quartile ranges of the papers. It is also a surprise that The Mirror has the largest inter-quartile range. But this data does show that the spread of all The Times sentences is much more pronounced than The Express and The Mirror.