I have been asked to choose some newspapers, analyse them for content and style and make comparisons between them.

* Words that are formed with a dash (e.g. fifth-minute) are to be recorded as one word

* Any names of cities, people, clubs, etc are allowed in the results as I believe every name has a second meaning or are still always apart of our every day life

* No tables, headlines, bullet points, authors, etc are to be included, only the actual text itself

* Any numbers that are spelt as their proper English (e.g. four, three, etc) are allowed to be included

* Sub-headings and picture descriptions are not to be included in the results

* All words included in the sample are to be highlighted by a brightly coloured pen

I think with these rules in place, the results are fair and should make a good overall analysis.

I believe that The Sun will have more words dedicated to the sports section and so also will have the smallest mean of letters per word as I believe the language used will be shorter due to it being a tabloid paper.

Collecting the data

After purchasing each of my wanted papers I then looked through each one to try and find the stories I needed and proposed. Once I found an interesting story I then checked to make sure each paper had the same story article in it. With all three papers containing the same story I then cut out the articles excluding as many pictures, tables, statistics as possible in relation to the actual story itself.

I then analysed each article from each paper and then decided that due to the amount of words included in each article, that every tenth word would be an appropriate sample to take for my results. This is called Systematic Sampling. At first, before I purchased the papers, I felt that every fifth word would be appropriate, or even a bit more than that. But after purchasing the papers I then changed my mind and found myself with a respectable amount of results from each article to go towards my overall results.

My method of sampling is going to be chosen by choosing every tenth word of the article in a continuous way until the end of the article. Therefore I cannot be bias in any way and my results are fair and simple.

Analysing the data

After going through each of the newspaper articles and highlighting every tenth word I then wrote down in rough the amount of letters in each one of the highlighted pieces of text, onto paper using a tally as I felt this was the simplest and easiest way to collect the data. Here is the data I have collected so far:

Key

= The Times

= The Sun

= Daily Mail

The Times:

Amount of letters per sample

Frequency

Cumulative Frequency

4

4

2

7

1

3

7

28

4

6

44

5

2

56

6

5

61

7

3

64

8

7

71

9

0

71

0

3

74

The Sun:

Amount of letters per sample

Frequency

Cumulative Frequency

4

4

2

22

26

3

5

41

4

1

52

5

9

61

6

8

69

7

5

74

8

2

76

9

3

79

0

4

83

1

0

83

2

0

83

3

84

Daily Mail:

Amount of letters per sample

Frequency

Cumulative Frequency

2

7

8

3

5

23

4

7

40

5

0

50

6

9

59

7

6

65

8

4

69

9

2

71

0

2

73

1

74

2

0

74

3

75

As you can see from the table of results, bar charts and pie charts, I have a nice good range of results. Now I decided to use the entire article of each newspaper because I wanted to investigate which newspaper seems to use Sport as one of their main assets to their newspaper. Instead of taking maybe a sample of the first 50 words of each newspaper I was interested to know how much each paper used. As you can see from the cumulative frequency column from the table of results, The Sun clearly uses more words in their article than any of the others. This could be because of the fact it is a tabloid and now a broadsheet. The Daily Mail is a bit of both, tabloid and broadsheet, but you can see that its sports section is very similar to The Time's sport section.

I decided to add a cumulative frequency column because to make box plots I must first find out the median, lower and upper quartiles. So with a cumulative frequency column it is easy to find out the 50% mark or half of the total. Once you have that you can then find out the upper and lower quartiles. To double check this, I made pie charts to show the percentage of letters per word and as you can see from the box plots, the results are the same.

The bar charts represent the result tables but in a graphical way to make it clearer how much are dedicated to each section.

Standard Deviation

Standard Deviation is basically another method of dispersion, but more advanced. We already know two measures of dispersion, the range and the interquartile range.

The range = highest value - lowest value

The interquartile range = value of upper quartile - value of lower quartile.

Standard deviation gives a measure of the of the 'average' deviation of each number from the mean, although by squaring the deviations, values further away from the mean affect the answer more than the values nearer to the mean.

Here is the formula to find the standard deviation:

?x ²ƒ - x ²

(n-1)

Then to find out the standard deviation of our results, we must make a different results table as follows:

Amount of letters per sample (x)

Frequency (ƒ)

x²

x²ƒ

4

4

2

7

4

28

3

7

9

53

4

6

6

256

5

2

25

300

6

5

36

80

7

3

49

47

8

7

64

448

9

0

81

0

0

3

00

300

Total:

74

385

816

Amount of letters per sample (x)

Frequency (ƒ)

x²

x²ƒ

4

4

2

22

4

88

3

5

9

35

4

1

6

76

5

9

25

225

6

8

36

288

7

5

49

245

8

2

64

28

9

3

81

243

0

4

00

400

1

0

21

0

2

0

44

0

3

69

69

Total:

84

819

2101

Amount of letters per sample (x)

Frequency (ƒ)

x²

x²ƒ

2

7

4

28

3

5

9

35

4

7

6

272

5

0

25

250

6

9

36

324

7

6

49

294

8

4

64

256

9

2

81

62

0

2

00

200

1

21

21

2

0

44

0

3

69

69

Total:

75

819

2212

Conclusion

As you can see from the results given, I have a lot of information to prove my prediction. As you can see from the results tables, The Sun dedicates a lot of words to its sports section, but then again that's not the main aspect of my investigation. You can also see from the box plots that the median for each newspaper I chose was four letters per word.

Now we can not be certain that this will always be true as it is only systematic sampling but if I was to perform this again I would change many things.

Firstly I believed I set the investigation out wrong. I believe the newspapers I chose were right, but the fact I used different amounts of words per article means I went about it the wrong way. Instead of focusing upon one main aspect, I focused upon two. One I focused upon the size each newspaper dedicates to its sports article, and also the amount of letters per every tenth word. If I preformed this again I would most defiantly focus upon the second point given instead of the first.

Also I might not used Systematic sampling again. I might choose random sampling, maybe by some how using a computer based program to do this. Or If I would go into this investigation into great detail I would have chosen more articles of the same story but in three different newspapers. Like 3 different stories in 3 different newspapers, therefore leaving you with nine lots of results to compare and discuss. This way I am sure you would be able to find out which newspaper actually uses longer words or not compared with others.

Bradley Wells - Maths Coursework - Anglo European School - Miss Filby

Choose some newspapers, and analyse them for content and style and make comparisons between them.