The editor of a local newspaper sent the following memo to the managing editor of the paper:
“Three years ago when we surveyed our subscribers, they complained about the quality of the reporting and writing in the paper. Since that time, we have made a concerted effort to hire older, more experienced journalists. According to our most recent survey, these changes have made a vast improvement. Not only has our subscriber base grown by 13%, but they also rated us higher on both reporting and writing. Therefore, it is evident that to continue to increase our readership, we should hire the most experienced journalists we can find, and gradually fire our younger, less experienced reporters."
Write a response in which you discuss what questions would need to be answered in order to decide whether the recommendation is likely to have the predicted result. Be sure to explain how the answers to these questions would help to evaluate the recommendation.
It is certainly understandable that the editor would evaluate this trend and conclude that since survey results and the subscriber base improved once older and more experienced journalists were hired, the newspaper's hiring strategy should be altered. However, the logic presented in the memo is flawed and more questions need to be answered before this conclusion can be confirmed.
The first assumption that the editor makes is that correlation implies causation. There are no other external factors that could explain this positive subscriber growth. However, anyone who has taken an introductory statistics class knows that correlation most definitely does not imply causation. Hence, the first question that needs to be answered to support the recommendation presented is "Are there any other factors that could have explained this subscriber growth?" One possible explanation that immediately comes to mind is a major event such as an election increasing interest in current events. Just look at this most recent United States presidential elections. Millions of people suddenly became interested in politics, watched extremely boring debates, and waited in line for hours to cast their votes. A local election is certainly not as headline grabbing as controversial presidential election featuring Donald Trump, but local elections can have a significant impact on the day to day lives of people. It should not be a surprise then that at least 13% more people would start investing in their local newspaper to follow the election.
Another assumption that the editor makes is that survey results are reliable. A few questions must be addressed regarding the subscriber surveys referenced in the memo. What is the survey's sample size? Was this sample representative of the entire population or clearly biased towards one demographic or age group? How were the questions presented? Those three questions are crucial to understand before any rigorous data analysis can commence. Many people tend to ignore surveys or prefer to only disclose their opinions when they are dissatisfied, potentially explaining the negative survey results from three years before the memo was written. People who work long hours may also not have time to respond to survey questions, which would again limit the sample size and result in a biased sample of unemployed or retired customers. Or perhaps the younger reporters were liberal politically and the subscribers who responded to the survey were more conservative and appreciated the articles of the older reporters, who at least on average tend to more conservative or less liberal. The survey may have a large and unbiased sample, but a statistically significant conclusion cannot be drawn until these questions are answered.
Before the editor makes any further personnel decisions, the questions about correlation versus causation and the validity of the survey results need to be evaluated. It is important to remember that firing anyone could destroy their livelihood, so these drastic decisions should not be justified with lazy and superficial analysis.
Comments
e-rater score report
Attribute Value Ideal
Final score: 4.0 out of 6
Category: Good Excellent
No. of Grammatical Errors: 4 2
No. of Spelling Errors: 1 2
No. of Sentences: 23 15
No. of Words: 481 350
No. of Characters: 2577 1500
No. of Different Words: 254 200
Fourth Root of Number of Words: 4.683 4.7
Average Word Length: 5.358 4.6
Word Length SD: 2.815 2.4
No. of Words greater than 5 chars: 200 100
No. of Words greater than 6 chars: 147 80
No. of Words greater than 7 chars: 115 40
No. of Words greater than 8 chars: 80 20
Use of Passive Voice (%): 0 0
Avg. Sentence Length: 20.913 21.0
Sentence Length SD: 9.269 7.5
Use of Discourse Markers (%): 0.565 0.12
Sentence-Text Coherence: 0.274 0.35
Sentence-Para Coherence: 0.43 0.50
Sentence-Sentence Coherence: 0.037 0.07
Number of Paragraphs: 4 5
Transition Words or Phrases used:
also, but, first, hence, however, if, look, may, regarding, so, then, at least, such as
Attributes: Values AverageValues Percentages(Values/AverageValues)% => Comments
Performance on Part of Speech:
To be verbs : 30.0 19.6327345309 153% => OK
Auxiliary verbs: 18.0 12.9520958084 139% => OK
Conjunction : 19.0 11.1786427146 170% => OK
Relative clauses : 19.0 13.6137724551 140% => OK
Pronoun: 30.0 28.8173652695 104% => OK
Preposition: 44.0 55.5748502994 79% => OK
Nominalization: 17.0 16.3942115768 104% => OK
Performance on vocabulary words:
No of characters: 2636.0 2260.96107784 117% => OK
No of words: 481.0 441.139720559 109% => OK
Chars per words: 5.48024948025 5.12650576532 107% => OK
Fourth root words length: 4.68313059816 4.56307096286 103% => OK
Word Length SD: 2.89428750628 2.78398813304 104% => OK
Unique words: 260.0 204.123752495 127% => OK
Unique words percentage: 0.540540540541 0.468620217663 115% => OK
syllable_count: 832.5 705.55239521 118% => OK
avg_syllables_per_word: 1.7 1.59920159681 106% => OK
A sentence (or a clause, phrase) starts by:
Pronoun: 4.0 4.96107784431 81% => OK
Article: 8.0 8.76447105788 91% => OK
Subordination: 1.0 2.70958083832 37% => OK
Conjunction: 4.0 1.67365269461 239% => Less conjunction wanted as sentence beginning.
Preposition: 1.0 4.22255489022 24% => More preposition wanted as sentence beginning.
Performance on sentences:
How many sentences: 22.0 19.7664670659 111% => OK
Sentence length: 21.0 22.8473053892 92% => OK
Sentence length SD: 70.7789744426 57.8364921388 122% => OK
Chars per sentence: 119.818181818 119.503703932 100% => OK
Words per sentence: 21.8636363636 23.324526521 94% => OK
Discourse Markers: 3.95454545455 5.70786347227 69% => OK
Paragraphs: 4.0 5.15768463074 78% => More paragraphs wanted.
Language errors: 0.0 5.25449101796 0% => OK
Sentences with positive sentiment : 10.0 8.20758483034 122% => OK
Sentences with negative sentiment : 5.0 6.88822355289 73% => OK
Sentences with neutral sentiment: 7.0 4.67664670659 150% => OK
What are sentences with positive/Negative/neutral sentiment?
Coherence and Cohesion:
Essay topic to essay body coherence: 0.155174158312 0.218282227539 71% => OK
Sentence topic coherence: 0.0450077997911 0.0743258471296 61% => OK
Sentence topic coherence SD: 0.0389895170497 0.0701772020484 56% => OK
Paragraph topic coherence: 0.0905546366123 0.128457276422 70% => OK
Paragraph topic coherence SD: 0.0325241764051 0.0628817314937 52% => OK
Essay readability:
automated_readability_index: 15.3 14.3799401198 106% => OK
flesch_reading_ease: 41.7 48.3550499002 86% => OK
smog_index: 8.8 7.1628742515 123% => OK
flesch_kincaid_grade: 12.7 12.197005988 104% => OK
coleman_liau_index: 14.5 12.5979740519 115% => OK
dale_chall_readability_score: 9.34 8.32208582834 112% => OK
difficult_words: 142.0 98.500998004 144% => OK
linsear_write_formula: 11.0 12.3882235529 89% => OK
gunning_fog: 10.4 11.1389221557 93% => OK
text_standard: 11.0 11.9071856287 92% => OK
What are above readability scores?
---------------------
Rates: 83.33 out of 100
Scores by essay e-grader: 5.0 Out of 6
---------------------
Note: the e-grader does NOT examine the meaning of words and ideas. VIP users will receive further evaluations by advanced module of e-grader and human graders.