In doing my pre-test, it became clear that counting the number of letters per word was not only time consuming and wasteful, it was also very tedious. Therefore I have only shown the results for the much more effective word count, showing the page number against the number of words in all three papers.
I am now going to calculate the average number of words per sentence in each type of paper. Despite this being a fairly brief and inconclusive test, it will give a clear indication of which paper I expect to contain most words and therefore what the readability of the paper is.
The Daily Star- Tabloid
523 / 30 = 17 words per sentence
The Guardian- Broadsheet
1055 / 30 = 35 words per sentence
The Financial Times- Quality
978 / 30 =33 words per sentence
It is clear that the tabloid has very few words in each sentence compared to both the broadsheet and the quality paper. This leads me on to my hypothesis…
Hypothesis:
Based on my pre test and its results, it would be very sensible to predict that broadsheets and quality newspapers will contain many more words than tabloid papers, therefore making them easier to read. This lead me on to believe that the more short sentences there are in a paper, the more readable it will be.
Why?
I am expecting to find this because, generally, shorter sentences contain fewer words because they do not have to support long complicated ones, so the sentence therefore becomes easier to read. For example-
‘An apple a day keeps the doctor away’
This sentence has very simple wording so is therefore very short in length, eight words. This sentence would be much more likely to found in a tabloid like ‘The star’ because the audience that reads this paper are perhaps not as literate or do not read it for the detailed descriptions, they want a general outline of the news in a few words.
‘A fresh golden delicious per day prevents needing to visit a general practitioner’.
This sentence uses longer words and is more descriptive, and it is also longer in length, 13 words. This sentence would be much more likely to be found in a broadsheet like ‘The Guardian’ because the audience that reads this paper are possibly more intelligent or read it for more information.
I am basing my hypothesis on my own knowledge already but by investigating I will hopefully prove it to be right.
Before I can continue to find out how to take my sample I must first consider the different sections of the paper which we could study. These sections are… adverts, news, leisure, pictures, headlines, stocks and shares and sport.
We chose to select four categories, as this would give us a chance to see which percentage of the paper was dedicated o certain subjects. Here are my results for this section… (All results shown are as a percentage of the page and are only approximate)
Tabloid (the daily star):
Broadsheet: The Guardian
Quality: The Financial Times
Now I can work out the totals for each section in each paper as a percentage. Shown alongside each of these is a pie chart to accurately and clearly show the percentage of each page (on average) given to each category.
Tabloid:
ADVERTS: 29%
NEWS: 22%
PICTURES: 28%
LEISURE: 21%
Broadsheet:
ADVERTS: 8%
NEWS: 61%
PICTURES: 25%
LEISURE: 6%
Quality:
ADVERTS: 17%
NEWS: 57%
PICTURES: 14%
LEISURE: 12%
I chose to display this information in the form of a pie chart because it is very clear, and all four sets of data can be compared in one go.
Now that I have a fair idea that my hypothesis is going to prove true, I am going to have to use a stratified sample to get my final set of results. My stratified sample should be started with working out the number of sentences I will need to count from. I can do this by gaining an approximation of the number of sentences in the whole paper. Firstly I counted the number of sentences in one a randomly chosen page. This is done completely at random to avoid being biased. We simply flicked through the paper and stopped at a random page. The number of sentences was recorded and then multiplied by the number of pages in the whole paper, thus giving an estimation of the number of sentences…
STAR GUARDIAN FINANCIAL TIMES
Page No: 3 5 3
No of
Sentences: 30 68 56
Estimated
Number of
Sentences
In whole 1800 1768 1120
Paper:
It was now crucial that we only took data from sections of the paper that actually contained words, otherwise we would not be able to count the number of words in, say adverts!
Because we did not have the time (or the energy) to count over 1000 sentences from each paper, we decided to take a reasonable percentage of the estimate to actually sample, rather than just predict like we’ve done so far. Here are the results we got using simple trial and improvement:
We were looking to get a sample as close to 100 (a fairly good number to take a sample from because it represents a fairly broad scale of results). Therefore we decided to take 7.5% of the total estimate to sample from.
I have shown my raw data, collected into a tally chart, below:
